CONFLICT THEORY
MAJOR REFERRAL SITES FOR SOCIOLOGY IVSA International Visual Sociology Association
RESOURCES [email protected]
MATHSOC Mathematical Sociology
Demography and Population Studies WWW [email protected]
Virtual Library RURSOC-L Rural Sociology Discussion List
http://coombs.anu.edu.au/ResFacilities/ [email protected]
DemographyPage.html SOCIOLOGY-L Sociology Faculty Information
[email protected]
WWW Virtual Library: Sociology—SOCNET— SOCNET Social Networks
Courses, Resources [email protected]fl.edu
http://www.mcmaster.ca/socscidocs/
w3virtsoclib/socnet.htm SOCIOLOGY USENET NEWSGROUPS
SOCIOWEB A Sociology Resource Center In contrast to Internet Mailing Lists, UseNet
http://www.socioweb.com/~markbl/socioweb/ Newsgroups do not broadcast a message to all
those included on the list. Internet users must
Research engines for the social sciences specifically access the postings (messages) for any
http://tile.net/’’>tile.net such newsgroup. While on any given day, ‘‘sociolo-
gy’’ will be mentioned in a variety of newsgroup
WeacTies (for Sociology of the Internet) discussions, the only newsgroup where academic
http://www.princeton.edu//~soccomp/ sociological topics are regularly discussed is: alt.sci.so-
weacties/ ciology. Generally a few messages are added to this
discussion area every day.
SELECTED NONELECTRONIC
PERIODICALS IN SOCIOLOGY RONALD E. ANDERSON
Sociological Methodology: An Annual Volume CONFIDENCE INTERVALS
http://weber.u.washington.edu/~socmeth2/
software.html See Statistical Inference.
Social Science Computer Review
http://sagepub.com
The Information Society
http://www.ics.uci.edu/~kling/tis.html
INTERNET MAILING LISTS (LIST SERVERS) CONFLICT THEORY
ON SOCIOLOGICAL TOPICS
NOTE: Although the following article has not been revised for
This list of List Servers, or Listserv discussion this edition of the Encyclopedia, the substantive coverage is
groups, indicate the popularity of such communi- currently appropriate. The editors have provided a list of
cation vehicles. The format for each entry is: List- recent works at the end of the article to facilitate research and
Name followed by a brief one-phrase description. exploration of the topic.
The second line is the Internet subscription ad-
dress. Unless otherwise noted, anyone can sub- Conflict theory explains social structure and
scribe by sending a one-line message to these changes in it by arguing that actors pursue their
addresses. The one-line message should say ‘‘sub- interests in conflict with others and according to
scribe’’ followed by the List-Name. their resources for social organization. Conflict
theory builds upon Marxist analysis of class con-
ASASCAN Section on Sociology and Comput- flicts, but it is detached from any ideological com-
ing (ASA) mitment to socialism. Max Weber generalized con-
[email protected] flict to the arenas of power and status as well as
economic class, and this multidimensional approach
FAMLYSCI Family studies has become widespread since the 1950s.
[email protected]
414
CONFLICT THEORY
WHAT CONSTITUTES A CONFLICT process is limited when there are cross-cutting
GROUP? memberships among groups, for instance, if class,
ethnic, and religious categories overlap. In these
For Marx and Engels, a society’s conflicting inter- cases, mobilization of one line of conflict (e.g.,
ests derive from the division between owners and class conflict) puts a strain on other dimensions of
nonowners of property. Dahrendorf (1959) pro- conflict (e.g., ethnic identity). Thus, cross-cutting
posed that conflicts are based on power, dividing conflicts tend to neutralize each other. Converse-
order-givers, who have an interest in maintaining ly, when multiple lines of group membership are
the status quo, from order-takers, who have an superimposed, conflicts are more extreme.
interest in changing it. Property is only one of the
bases of power conflict, and conflicts can be ex- Conflicts escalate as each group retaliates
pected inside any type of organization, including against offenses received from the other. How
socialist ones. In the Weberian model there are long this process of escalation continues depends
even more types of conflict, since every cultural on how much resources a group can draw upon: its
group (such as ethnic, religious, or intellectual numbers of supporters, its weapons, and its eco-
groups) can also struggle for advantage. In addi- nomic goods. If one group has many more re-
tion, economic conflict takes place in three differ- sources than the other, the conflict ends when the
ent types of market relations, pitting employers mobilizing capacity of the weaker side is exhaust-
against workers, producers against consumers, and ed. When both sides have further resources they
lenders against borrowers (Wiley 1967). Gender have not yet mobilized, escalation continues. This
stratification produces yet another dimension of is especially likely when one or both sides have
conflict. sustained enough damage to outrage and mobilize
their supporters but not great enough damage to
THE PROCESS OF CONFLICT destroy their organizational resources for struggle.
Conflicting interests remain latent until a group Deescalation of conflict occurs in two very
becomes mobilized for active struggle. This occurs different ways. If one side has overwhelming supe-
when its members are physically concentrated, riority over the other, it can destroy opposition by
have material resources for communicating among breaking the other group’s organizational capacity
themselves, and share a similar culture. The high- to fight. The result is not harmony but an uneasy
er social classes are typically more mobilized than peace, in which the defeated party has been turned
lower classes, and most struggles over power take back into an unmobilized latent interest. If neither
place among different factions of the higher class- side is able to break up the other’s organization,
es. Lower classes tend to be fragmented into local- conflict eventually deescalates when resources are
ized groups and are most easily mobilized when eaten up and the prospects of winning become
they are a homogeneous ethnic or religious group dimmer. Although wars usually arouse popular
concentrated in a particular place. The better solidarity at first, costs and casualties reduce en-
organized a conflict group is, the longer and more thusiasm and bring most wars to an end within a
intensely it can struggle; such struggles become few years. Civilian uprisings, strikes, and other
routinized, as in the case of entrenched labor small-scale conflicts typically have fewer resources
unions or political parties. Less organized conflict to sustain them; these conflicts deescalate more
groups that become temporarily mobilized are quickly. During a deescalation, the points of con-
more likely to be violent but unable to sustain the tention among the opponents modulate from ex-
conflict. treme demands toward compromises and piece-
meal negotiation of smaller issues (Kriesberg 1982).
Overt conflict increases the solidarity of groups Very destructive levels of conflict tend to end
on both sides. Coser (1956), elaborating the theo- more rapidly than moderate conflicts in which
ry of Georg Simmel, points out that conflict leads resources are continuously replenished.
to a centralization of power within each group and
motivates groups to seek allies. A conflict thus COERCIVE POWER AND REVOLUTION
tends to polarize a society into two factions, or a
world of warring states into two alliances. This In a highly coercive state, such as a traditional
aristocracy or a military dictatorship, power is
415
CONFLICT THEORY
organized as an enforcement coalition (Collins these conflicts; hence, another condition must
1988; Schelling 1962). Members of the ruling or- also be present, that the society is relatively pro-
ganization monitor each other to ensure loyalty. A ductive and organizationally well integrated. If the
change in power is possible only when a majority state has high resources relative to its enemies,
of the enforcers disobey orders simultaneously. conflict is the route by which it concentrates power
Revolts occur in a rapid ‘‘bandwagon effect,’’ dur- in its own hands.
ing which most members scramble to become part
of the winning coalition. The more coercive the Prestige is determined by the concentration of
state, the more extreme the swings between long power and wealth. Groups that have these resourc-
periods of tyrannical stability and brief moments es can invest them in material possessions that
of political upheaval. make them impressive in social encounters. In
addition, they can invest their resources in culture-
Since the state claims a monopoly on the producing organizations such as education, enter-
instruments of violence, revolutionary changes in tainment, and art, which give them cultural domi-
power occur through the reorganization of coer- nation. According to Pierre Bourdieu’s research
cive coalitions. Revolts from below are almost (1984), the realm of culture is stratified along the
always unsuccessful as long as the state’s military same lines as the stratification of the surrounding
organization stays intact. For this reason, revolu- society.
tions typically are preceded by a disintegration of
the military, due to defeat in war, depletion of EFFECTS OF CONFLICT GROUPS UPON
economic resources in previous conflicts, and splits INDIVIDUALS
within the ruling group (Skocpol 1979). These
breakdowns of military power in turn are deter- The latent lines of conflict in a society divide
mined by geopolitical processes affecting the ex- people into distinctive styles of belief and emo-
pansion or contraction of states in the surround- tion. Collins (1975) proposed that the differences
ing world (Stinchcombe 1968; Collins 1986). among stratified groups are due to the
microinteractions of daily experience, which can
WHO WINS WHAT? occur along the two dimensions of vertical power
and horizontal solidarity. Persons who give orders
Conflict shapes the distribution of power, wealth, take the initiative in the interaction rituals de-
and prestige in a society. The victorious side is scribed by Goffman (1959). These persons who
generally the group that is better mobilized to act enact the rituals of power identify with their front-
in its collective interest. In many cases, the domi- stage selves and with the official symbols of the
nant group is well organized, while the opposing organizations they control; whereas persons who
interest group remains latent. The result is a stable take orders are alienated from official rituals and
structure of stratification, in which overt conflict identify with their private, backstage selves. Indi-
rarely occurs. viduals who belong to tightly enclosed, localized
groups emphasize conformity to the group’s tradi-
Lenski (1966) showed that concentration of tions; persons in such positions are suspicious
wealth throughout world history is determined by of outsiders and react violently and emotionally
the interaction of two factors. The higher the against insiders who are disrespectful of the group’s
production of economic surplus (beyond what is symbols. Loosely organized networks have less
necessary to keep people alive), the greater the solidarity and exert less pressure for conformi-
potential for stratification. This surplus in turn is ty. Individuals build up emotional energy by
appropriated according to the distribution of power. microexperiences that give them power or soli-
darity, and they lose emotional energy when they
Turner (1984) theorizes that the concentra- are subordinated to power or lack experiences of
tion of power is unequal to the extent that there is solidarity (Collins 1988). Both emotions and be-
external military threat to the society or there is a liefs reproduce the stratification of society in eve-
high level of internal conflict among social groups. ryday life.
Both external and internal conflict tend to central-
ize power, providing that the government wins
416
CONTENT ANALYSIS
(SEE ALSO: Coalitions; Game Theory and Strategic Inter- Lenski, Gerhard E. 1966 Power and Privilege: A Theory of
action; Interpersonal Power) Stratification. New York: McGraw-Hill.
REFERENCES Schelling, Thomas C. 1962 The Strategy of Conflict.Cam-
bridge, Mass.: Harvard University Press.
Bailey, Kenneth D. 1997 ‘‘System and Conflict: Toward
a Symbiotic Reconciliation.’’ Quality and Quantity Skocpol, Theda 1979 States and Social Revolutions. New
31:425–442. York: Cambridge University Press.
Bourdieu, Pierre 1984 Distinction. Cambridge, Mass.: Stinchcombe, Arthur L. 1968 Constructing Social Theo-
Harvard University Press. ries. New York: Harcourt.
Briggs, E. Donald 1992 ‘‘New Directions in Conflict Turner, Jonathan H. 1984 Societal Stratification: A Theo-
Theory: Conflict Resolution and Conflict Transfor- retical Analysis. New York: Columbia University Press.
mation.’’ Canadian Journal of Political Science / Revue
canadienne de science politique 25:430–431. ——— 1993 Classical Sociological Theory: A Positivist’s
Perspective. Chicago, Ill.: Nelson-Hall Publishers.
Chapin, Mark 1994 ‘‘Functional Conflict Theory, the
Alcohol Beverage Industry, and the Alcoholism Treat- Van-Huyssteen, Elsa F. 1994 ‘‘Interpretation of the South
ment Industry.’’ Journal of Applied Social Sciences African Legal System in Terms of the Analytical
18:169–182. Conflict Perspective.’’ Suid Afrikaanse Tydskrif vir
Sosiologie / South African Journal of Sociology 25:87–94.
Collins, Randall 1975 Conflict Sociology. New York: Aca-
demic Press. Wiley, Norbert F. 1967 ‘‘America’s Unique Class Poli-
tics: The Interplay of the Labor, Credit, and Com-
——— 1986 Weberian Sociological Theory. Cambridge: modity Markets.’’ American Sociological Review
Cambridge University Press. 32:529–540.
——— 1988 Theoretical Sociology. San Diego: Harcourt RANDALL COLLINS
Brace Jovanovich.
CONSISTENCY THEORY
Collins, Randall 1993 ‘‘What Does Conflict Theory Pre-
dict about America’s Future?’’ Sociological Perspectives See Cognitive Consistency Theories.
36:289–313.
CONTENT ANALYSIS
Coser, Lewis A. 1956 The Functions of Social Conflict. New
York: Free Press. ‘‘Content analysis’’ has evolved into an umbrella
label that includes various procedures for making
Dahrendorf, Ralf 1959 Class and Class Conflict in Industri- reliable, valid inferences from qualitative data,
al Society. Stanford, Calif.: Stanford University Press. including text, speech, and images. These proce-
dures have improved and expanded due to numer-
Gagnon, V. P., Jr. 1995 ‘‘Ethnic Conflict as an Intra- ous developments in recent years since this ency-
Group Phenomenon: A Preliminary Framework.’’ clopedia’s first edition.
Revija za Sociologiju 26:1–2.
Traditionally, ‘‘content analysis’’ has referred
Glaser, James M. 1994 ‘‘Back to the Black Belt: Racial to systematic procedures for assigning prespecified
Environment and White Racial Attitudes in the codes to text, such as interviews, newspaper edito-
South.’’ Journal of Politics 56:21–41. rials, open-ended survey answers, or focus-group
transcripts, and then analyzing patterns in the
Goffman, Erving 1959 The Presentation of Self in Everyday codings. Some projects will count each specific
Life. New York: Doubleday. occurrence within a text, while others will have
coders tally the number of column inches assigned
Hanneman, Robert A., Randall Collins, and Gabriele a code. Either way, the procedure usually employs
Mordt 1995 ‘‘Discovering Theory Dynamics by Com-
puter Simulation: Experiments on State Legitimacy
and Imperialist Capitalism.’’ Sociological Methodolo-
gy 25:1–46.
Haugaard, Mark 1997 ‘‘The Consensual Basis of Conflictual
Power: A Critical Response to ‘Using Power, Fighting
Power’ by Jane Mansbridge.’’ Constellations 3:401–406.
Kriesberg, Louis 1982 Social Conflicts. Englewood Cliffs,
N.J.: Prentice-Hall.
417
CONTENT ANALYSIS
a ‘‘top down’’ strategy, beginning with a theory by different forms of media, and inasmuch as
and hypotheses to be tested, developing reliable different analytic procedures can complement one
coding categories, applying these to coding-speci- another in uncovering important insights, it makes
fied bodies of text, and finally testing the hypothe- sense to strive toward an integration, rather than
ses by statistically comparing code indexes across fragmentation.
documents.
Many advances in content-analysis procedures
With the increasing popularity of qualitative have been made possible by the convenience and
sociology, content analysis has also come to refer power of desktop and laptop computers. In addi-
to ‘‘grounded’’ inductive procedures for identify- tion, an overwhelming proportion of text docu-
ing patterns in various kinds of qualitative data ments are now generated on computers, making
including text, illustrations, and videos. For exam- their text files computer accessible for content
ple, the data might include observers’ detailed analysis. And a revolution in hand-held analogue
notes of children’s behaviors under different forms and digital video cameras, together with comput-
of supervision, possibly supplemented with video- er-based technology for editing and analyzing video-
tapes of those same behaviors. While traditional tapes, makes new research procedures feasible.
content analysis usually enlisted statistical analyses
to test hypotheses, many of these researchers do Consider, for example, new possibilities for
not start with hypotheses, but carefully search for analyzing responses to open-ended questions in
patterns in their data. survey research. For years, survey researchers have
been well aware that closed-ended questions re-
However, rather than just produce statistical quire respondents to frame how they think about
analyses or search for patterns, investigators should an issue in terms of a question’s multiple choices,
also situate the results of a content analysis in even when the choice options had little to do with
terms of the contexts in which the documents were how a respondent views an issue. But the costs and
produced. A content-analysis comparison of let- time involved in analyzing open-ended responses
ters to stockholders, for example, should take into resulted in such questions rarely being used. Even
consideration the particular business sectors cov- when they were included in a survey, the interview-
ered and the prevailing economic climates in which ers usually just recorded capsule summaries of the
they were written. An analysis of American presi- responses that omitted most nuances of what
dential nomination acceptance speeches should was said.
consider that they changed dramatically in form
once they started to be broadcast live on national Contrast this then with survey research using
radio. A content analysis may have reliable coding, today’s audio information-capturing technologies.
but the inferences drawn from that coding may Telephone survey interviewers are guided by in-
have little validity unless the researcher factors in structions appearing on a computer screen. When-
such shaping forces. ever an open-ended question appears, the inter-
viewer no longer needs to type short summaries of
Like any expanding domain, there has been a the responses. Instead, a computer digitally cap-
tendency for content analysis to segment into tures an audio recording of each open-ended re-
specialized topics. For example, Roberts (1997) sponse, labels it, and files it as a computer record.
focuses on drawing statistical inferences from text, Any audio response can later be easily fetched and
including Carley’s networking strategies and replayed, allowing a researcher, for example, to
Gottschalk’s clinical diagnostic tools. There also identify a ‘‘leaky voice,’’ that is, one indicative of
has been a stream of instructional books, includ- the respondent’s underlying emotion or attitude.
ing several series published by Sage, that focus on And the full audio responses are then available to
particular kinds of qualitative data such as focus- be transcribed to text, including, if desired, nota-
group transcripts. A technical literature also has tions indicating hesitations and voice inflections.
developed addressing specialized computer soft- Until computer voice recognition is completely
ware and video-analysis techniques. Nevertheless reliable, transcribing usually remains a manual
the common agenda is analyzing the content of task. But with spreadsheet software (such as Excel)
qualitative data. Inasmuch as our lives are shaped no longer restrictively limiting the amount of text
418
CONTENT ANALYSIS
in any one cell, the text of each individual’s entire possibly ‘‘authority-role.’’ Computer content-analy-
response to a question can be placed in a spread- sis software may search the contexts of words in
sheet cell, thus capturing both the closed-ended the text to ferret out and correctly code common
and open-ended data for a survey into a conven- word senses. For example, for a national study of
ient, single spreadsheet for researchers to analyze. people’s perceptions of African-American young
males on several open-ended questions, it was
With the survey data in this convenient form, particularly important for the computer to identi-
researchers can then code open-ended responses fy correctly each respondent’s usages of such mul-
manually, putting their assigned codes in addition- ti-meaning words as ‘‘race,’’ ‘‘white,’’ ‘‘black,’’ and
al spreadsheet columns. As a teaching exercise, it ‘‘color.’’
is instructive to assign students a task such as
identifying gender differences among a thousand In addition to developing their own coding
responses to a broad open-ended question, such as categories, researchers may enlist existing com-
a question asking respondents’ views about peace puter-scored categories that are relevant to the
or family values. Students first might sort the task at hand. For example, it might be hypothe-
spreadsheet by gender in order to read separately sized that one group being studied is more opti-
samples of male responses and female responses mistic and its responses will reflect more ‘‘positive
and obtain a sense of what possible gender differ- thinking’’ while another is more negative or pessi-
ences exist. They then develop coding instructions mistic. To code ‘‘positive thinking’’ a researcher
that capture these differences and apply the codes may want the computer to apply an existing con-
to the entire set of responses. This coding, of tent-analysis category that includes over 1200 words,
course, is better done without knowledge of the word roots, word senses, phrasal verbs, and idi-
respondents’ genders, with responses in a random oms, thus essentially covering most expressions of
order, and on different respondents than those ‘‘positive-thinking’’ that occur as infrequently as
used to develop the codes. After coding several three times per million words of ordinary English
hundred responses, however, students usually be- text. A similar category exists for negative-think-
gin to glaze over and soon the most ardent human- ing, allowing the investigator to check whether the
ist student is asking whether the computer could groups being studied differ in their coded positive
possibly be of help in assigning codes. thinking, negative thinking, or both. And by enlist-
ing such standard categories, the results obtained
For some kinds of coding, computer help is in one study can be readily compared with results
indeed available in the form of computer pro- found in other studies.
grams that assign codes. Such codings can be
treated as advisory and then manually confirmed, Once data has been captured in a convenient
augmented perhaps by also assigning a weight. Or format for computer use, they can be repeatedly
they may be used as is after being spot-checked for analyzed. For example, should our now glazed-
accuracy. Not only can a computer complete huge over students have any energy left after analyzing
amounts of tedious coding in minutes, possibly the responses by gender, they could be given an
assigning many different types of codings to each additional assignment of identifying and coding
text, but these codings may uncover statistically rural-urban differences in these same open-ended
significant frequency differences that human coders responses. Given so many analyses that can be
would not uncover, if only because computer analy- made, it makes sense to let the computer do what it
sis is so even-handed and untiring. The static can, saving manual labor for those types of codings
created by occasional miscodings may be more that would be hard to have a computer assign.
than offset by gains from a reliable consistency in Even multimedia qualitative-analysis software such
making many codings. as HyperResearch includes some rudimentary tools
for automatic assignments.
Computer coding assignments are usually based
on the occurrence of words, particular senses of Moreover, desktop computer software has al-
words, or multiword idioms appearing in the text. so become available that identifies patterns in text
For example, the word ‘‘father’’ in a text might be without having to develop coding categories. For
coded as ‘‘male,’’ ‘‘family member,’’ etc. as well as example, SPSS’s TextSmart uses an algorithm that
419
CONTENT ANALYSIS
groups respondents into clusters based on word- often complement those gained from another.
use co-occurrences within responses and then maps Instead of being limited by technology, the limits
these clusters into a two-dimensional grid that uses now may lie in the skills, proclivities, and comfort
colors to represent each cluster group. If such an zones of the researchers. Research teams, rather
automated inductive procedure can produce addi- than individual researchers, may prove the best
tional valid insights that other techniques are like- solution, for only in a team made up of people with
ly to overlook, then why not use it too? complementary strengths is one likely to find the
full range of statistical, conceptual, intuitive, expe-
Pioneering work in inductive automatic cate- riential, and perhaps clinical strengths needed to
gorizing, such as Iker’s (1969), usually enlisted carry out penetrating, comprehensive content-analy-
procedures based upon correlation matrixes, such sis projects. Moreover, some researchers will pre-
as factor analysis. These procedures tended not to fer to learn from the main trends while others will
be particularly suited to analyzing text both be- learn more from studying outlying cases. Some
cause of the shape of word-usage frequency distri- will learn from bottom-line numbers while others
butions as well as the limited number of words that will learn more from innovative graphics that high-
a correlation matrix could feasibly handle. TextSmart, light information patterns. Some will focus on
based upon a word-distance measure, provides current data while others will contextualize data
much more suitable solutions. Further automatic historically by comparing them with data in ar-
categorizing procedures may be expected from chives. Data that has been gathered and assembled
artificial intelligence, as well as from categorizing at considerable cost, especially data-gathering that
techniques being developed for Internet search imposed on many respondents, merit as thorough
engines. and comprehensive analyses as these various pro-
cedures collectively offer.
Content-analysis research strategies can thus
now easily be multipronged, spanning from com- Unfortunately, however, an ‘‘either-or’’ assump-
pletely automatic inductive procedures to manual tion about how to do content analysis has contin-
coding. But even manual coding these days is likely ued to be supported both by books and computer
to utilize computer software to help coders mana- software. Authors who do an excellent job of
ge information. Consider these changes in costs describing one approach to content analysis, such
and convenience: Unlike mainframe computing of as Boyatzis (1998), give an impression that an
the 1960s and 1970s, when the cost of an hour of either-or decision has to be made about which
computer time was about the same as a coder’s approach to use. Some software—especially that
wage for several weeks, the marginal cost of using ported from mainframe computers or developed
a desktop computer is essentially the electricity it for early desktop computers—still may steer or
uses. Today’s desktop computer is likely to be even limit researchers who use it to just one ap-
more than fivefold faster at content-analysis cod- proach. For example, some software packages cre-
ing than those mainframe computers ever were. ate specialized data formats such as ‘‘classification
They also can access much larger dictionaries and trees’’ that then in effect constrain the user to
other information in their RAM than was ever analyses that can be readily derived from that
feasible on a user partition of a mainframe com- format. Software reviews such as Lewis’s (1998)
puter, thus making their coding more accurate excellent comparison of ATLAS/ti and NUD-IST
and comprehensive. Moreover, a single CD-ROM software have been explicit about what assump-
full of text to be analyzed is easily popped into a tions a researcher buys into when utilizing each
desktop computer, whereas in the days of main- package.
frame computing, a comparable amount of text
would have to be keypunched on over 3,000 boxes Additional leverage in analyzing qualitative
of IBM cards. information has stemmed from computer-based
tools, such as newer versions of HyperResearch and
Given today’s convenience and low cost of ATLAS/ti, that integrate the handling of multiple
computer-based procedures, there is no reason to media (text, illustrations, and video). Especially as
limit an analysis to one approach, especially if more software comes from countries where there
insights gained from one approach will differ and are expert programming skills and programming
420
CONTENT ANALYSIS
labor is relatively inexpensive, we can expect ambi- to address such issues as how economic
tious content-analysis software, of which TextAnalyst cycles impact ideology, as examined by
from Russia (www.megaputer.com) may be a Namenwirth and Weber (1987), or to
forerunner. uncover cycles of creativity, as demonstrat-
ed by Martindale (1990)
Given continuing content-analysis software de-
velopments, those who would like to learn what is 4. Investigations, several of which are already
currently available are advised search Internet sites underway, of both intranet communi-
rather than rely upon even recently published cation patterns within organizations as
materials. One recommended starting point is the well as Internet communications, includ-
Georgia State University content-analysis site ing analyses of the content of communica-
(www.gsu.edu/~wwwcom/content.html), which gives tions over those networks.
links to software web sites (including software
mentioned in this article), indexes recent content- There is also, however, good reason for cau-
analysis publications, and has a mailing list of tion. Never before in history has so much qualita-
more than 700 members. Technical reviews of tive information been available electronically. High-
relevant language-analysis tools, such as Berleant’s volume image scanning will also further increase
(1995), occasionally appear in computational lin- the amount of information that can be electroni-
guistics journals and web sites. For training, the cally accessed and content-analyzed. Quite under-
University of Essex Summer School in Social Sci- standably, those agencies responsible for limiting
ence Data Analysis and Collection (www.essex.ac.uk), terrorist activities may look on content-analysis
as part of a program of a European consortium, procedures as possibly providing early warnings
has been offering a content-analysis module for that could save lives. But these procedures can also
years as part of its program. become tools for a ‘‘big-brother’’ monitoring so-
ciety. Sociologists have an important role in antici-
Given the developments described here, some pating these problems and helping resolve them.
of the contributions that content analysis should
be able to make to sociological research include: REFERENCES
1. A major shift from reliance upon closed- Berleant, Daniel 1995 ‘‘Engineering ‘Word Experts’ for
ended questions to an appropriate use of Word Disambiguation.’’ Natural Language Engineer-
open-ended questions that lets people be ing 1 (4):339–362.
heard in the ways they frame issues, as
well as the way they think and feel Boyatzis, Richard E. 1998 Transforming Qualitative Infor-
about them, as discussed in detail by mation: Thematic Analysis and Code Development. Thou-
Stone (1997) sand Oaks, Calif.: Sage.
2. A better understanding of both print and Iker, Howard, and Norman Harway 1969 ‘‘A Computer
television media and its impact on public Systems Approach Toward the Recognition and Analy-
opinion, both in setting agendas and in sis of Content.’’ In George Gerbner, Ole Hosti, Klaus
influencing opinion intensity, as laid out Krippendorff, William Paisley, and Philip Stone, eds.,
in Neuman (1989). This will involve The Analysis of Communication Content. New York: Wiley.
research that compares the content of
media with the content of opinions. Not Lewis, R. Berry 1998 ‘‘ATLAS/ti and NUD-IST: A Com-
only will survey research data be archived parative Review of Two Leading Qualitative Data
and accessible from Internet servers, but Analysis Packages.’’ Cultural Anthropology Methods 10
full-display media will be accessible from (3):41–47.
Lexis-Nexis and on-line editions supplied
by media providers, as well as television Martindale, Colin 1990 The Clockword Muse: The Predicta-
news archives such as those at Vanderbilt bility of Artistic Change. New York: Basic Books.
University
Namenwirth, J. Zvi, and Robert Philip Weber 1987
3. Better use of historical qualitative data, Dynamics of Culture. Winchester, Mass.: Allen and Unwin.
including both text and graphic materials,
Neuman, Richard 1989 ‘‘Parallel Content Analysis: Old
Paradigms and New Proposals.’’ Public Communica-
tion and Behavior 2:205–289.
421
CONVERGENCE THEORIES
Roberts, Carl W. (ed.) 1997 Text Analysis for the Social closely associated with modernization theories,
Sciences: Methods for Drawing Statistical Inferences from which generally held that developing societies will
Texts and Transcripts. Mahwah, N.J.: Lawrence Erlbaum. follow a path of economic development similar
to that followed by developed societies of the
Stone, Philip J. 1997 ‘‘Thematic Text Analysis: New West. Structural-functionalist theorists, such as
Agendas for Analyzing Text Content.’’ In Carl W. Parsons (1951) and Davis (1948), while not actually
Roberts, ed., Text Analysis for the Social Sciences: Meth- employing the terminology of convergence theo-
ods for Drawing Statistical Inferences from Texts and ry, paved the way for its development and use
Transcripts. Mahwah, N.J.: Lawrence Erlbaum. in modernization studies through their efforts
to develop a systematic statement of the func-
PHILIP STONE tional prerequisites and structural imperatives of
modern industrial society; these include an occu-
CONTINGENCY TABLES pational structure based on achievement rather
than ascription, and the common application of
See Tabular Analysis; Typologies. universalistic rather than particularistic evaluative
criteria. Also, beginning in the 1960s, convergence
CONVERGENCE THEORIES theory was invoked to account for apparent simi-
larities in industrial organization and patterns of
The idea that societies move toward a condition of stratification found in both capitalist and commu-
similarity—that they converge in one or more nist nations (Sorokin 1960; Goldthorpe 1964;
respects—is a common feature of various theories Galbraith 1967).
of social change. The notion that differences among
societies will decrease over time can be found in CONVERGENCE THEORY AND
many works of eighteenth and nineteenth cen- MODERNIZATION
tury social thinkers, from the prerevolutionary
French philosophes and the Scottish moral philoso- The conventional and most controversial applica-
phers through de Tocqueville, Toennies, Maine, tion of convergence theory has been in the study
Marx, Spencer, Weber, and Durkheim (Weinberg of modernization, where it is associated with the
1969; Baum 1974). More recently, the study of idea that the experience of developing nations will
‘‘postindustrial’’ society and the debate over follow the path charted by Western industrialized
‘‘postmodernist’’ aspects of contemporary society nations. Related to this idea is the notion of a
also reflect to some degree the idea that there is a relatively fixed pattern of development through
tendency for broadly similar conditions or attrib- which developing nations must pass as they mod-
utes to emerge among otherwise distinct and dis- ernize (Rostow 1960). Inkeles (1966), Inkeles and
similar societies. Smith (1974), and Kahl (1968) pursued the idea of
convergence at the level of individual attitudes,
In sociological discourse since the 1960s, the values, and beliefs, arguing that the emergence of
term convergence theory has carried a more specific a ‘‘modern’’ psychosocial orientation accompa-
connotation, referring to the hypothesized link nies national modernization (see Armer and
between economic development and concomitant Schnaiberg 1972 for a critique).
changes in social organization, particularly work
and industrial organization, class structure, demo- Kerr and colleagues’ Industrialism and Industri-
graphic patterns, characteristics of the family, edu- al Man (1960) offers the classic statement of the
cation, and the role of government in assuring ‘‘logic of industrialism’’ thesis, which the authors
basic social and economic security. The core no- proposed as a response to Marxian theory’s equa-
tion of convergence theory is that as nations achieve tion of industrial society with capitalism. More
similar levels of economic development they will specifically, Kerr et al. sought to identify the ‘‘in-
become more alike in terms of these (and other) herent tendencies and implications of industriali-
aspects of social life. In the 1950s and 1960s, zation for the work place,’’ hoping to construct
predictions of societal convergence were most from this a portrait of the ‘‘principal features of
the new society’’ (p. 33). The features common to
422
CONVERGENCE THEORIES
industrial society, they argued, include rapid character (Goldthorpe 1971), its Western ideo-
changes in science, technology, and methods of logical bias (Portes 1973), and for ignoring the
production; a high degree of occupational mobili- structurally dependent position of less-developed
ty, with continual training and retraining of the countries in the world economy (Wallerstein 1974).
work force; increasing emphasis on formal educa- Yet a careful review of the literature suggests that
tion, particularly in the natural sciences, engineer- many criticisms have often tended to caricature
ing, medicine, managerial training, and adminis- convergence theory rather than addressing its ap-
trative law; a workforce highly differentiated in plication in actual research studies. Since the 1960s
terms of occupational titles and job classifications; few if any researchers have explicitly claimed con-
the increasing importance of urban areas as cen- vergence theory, at least in its unreconstructed
ters of economic activity; and the increasing role form, as their own. For example, Moore (1979), an
of government in providing expanded public serv- exponent of the ‘‘conventional’’ view of moderni-
ices, orchestrating the varied activities of a large zation, subtitled his book, World Modernization,
and complex economy, and administering the ‘‘web ‘‘the limits of convergence,’’ and went to great
of rules’’ of industrial society. Importantly, Kerr et pains to distance himself from the ‘‘model mod-
al. envisioned these developments as cutting across ernized society’’ position associated with early ver-
categories of political ideology and political systems. sions of convergence theory (see Moore 1979, pp.
26–28, 150–153). And Parsons (1966), whose name
Although the ‘‘logic of industrialism’’ argu- is virtually synonymous with structural functional-
ment is often cited as a prime example of conver- ism, concluded one of his later writings on com-
gence theory (see Form 1979; Moore 1979; parative sociology with the statement that ‘‘any
Goldthorpe 1971), Kerr et al. never explicitly made linear theory of societal evolution’’ is ‘‘untenable’’
this claim for their study. While mentioning con- (p. 114). As Form (1979) observes, convergence
vergence at various points in their study, the au- theory passed through a cycle typical of social
thors pay equal attention to important countercur- science theories: a burst of initial interest and
rents leading toward diverse outcomes among enthusiasm, followed by intense criticism and con-
industrial societies. The concluding chapter of troversy, finally giving way to neglect. The major
Industrialism and Industrial Man is, in fact, entitled challenge to those wishing to revive convergence
‘‘Pluralistic Industrialism,’’ and addresses the sourc- theory and rescue it from its critics is to specify its
es of diversity as well as uniformity among indus- theoretical underpinnings more precisely, to de-
trial societies. Among sources of diversity identi- velop appropriate empirical studies, and finally
fied are the persistence of existing national account for variation as well as similarity among
institutions, enduring cultural differences, varia- observed cases.
tions in the timing of industrialization (late versus
early), the nature of a nation’s dominant industry, FORMS OF CONVERGENCE AND
and the size and density of population. Counterposed DIVERGENCE
against these factors are various sources of uni-
formity, such as technological change, exposure to In recent years Inkeles (1980, 1981; also Inkeles
the industrial world, and a worldwide trend to- and Sirowy 1983) has made the most systematic
ward increased access to education leading to an attempt to reformulate convergence theory and
attenuation of social and economic inequality. respecify its core hypotheses and propositions.
Inkeles (1981) argues that earlier versions of con-
The critique of convergence theory in the vergence theory failed to distinguish adequately
study of modernization recalls critiques of earlier between different elements of the social system,
theories of societal evolution advanced under the which is problematic because these elements not
rubric of social Darwinism in the nineteenth cen- only change at different speeds, but may move in
tury and structural functionalism in the mid-twen- opposite directions. He proposes dividing the so-
tieth century. The use of convergence theory to cial system into a minimum of five elements for
analyze modernization has been attacked for its purposes of assessing convergence: modes of pro-
alleged assumptions of unilinearity and determin- duction and patterns of resource utilization; insti-
ism (i.e., a single path of development that all tutional arrays and institutional forms; structures
societies must follow), its teleological or historicist
423
CONVERGENCE THEORIES
or patterns of social relationships; systems of popu- empirical research on specific institutional spheres
lar attitudes, values, and behavior; and systems of and social processes. As the following sections
political and economic control. Finally, he speci- demonstrate, a good deal of work along these lines
fies the different forms convergence and diver- is already being done across a wide range of sub-
gence may take: (1) simple convergence involving stantive questions and topical concerns that can
the movement from diversity to uniformity; (2) aptly be described in the plural as convergence
convergence from different directions involving theories, indicating their revisionist and more plu-
movement toward a common point by an increase ralistic approach.
for some cases and a decrease for others; (3)
convergence via the crossing of thresholds rather INDUSTRIAL SOCIOLOGY
than changes in absolute differences; (4) divergent
paths toward convergence, where short-term fluc- Despite criticisms of Kerr and colleagues’ (1960)
tuations eventually fall into line or a ‘‘deviant’’ case concept of the logic of industrialism, the question
that eventually defines the norm for other cases of convergent trends in industrial organization
(for example, France’s move toward small family has remained the focus of active debate and much
size in the late eighteenth century); and (5) conver- research. The large research literature related on
gence in the form of parallel change, where na- this question, reviewed by Form (1979), has pro-
tions all moving in the same direction along some duced mixed evidence with respect to conver-
dimension of change continue to remain separat- gence. Studies by Shiba (1973, cited in Form 1979),
ed by a gap. Although parallel change of this sort Form (1976), and Form and Kyu Han (1988),
does not represent true convergence, it is consis- covering a range of industrializing and advanced
tent with the key assumption of convergence theo- industrial societies found empirical support for
ry, namely, that ‘‘insofar as they face comparable convergence in workers’ adaptation to industrial
situations of action . . . nations and individuals will and related social systems, while Gallie’s (1977,
respond in broadly comparable ways’’ (p. 21). cited in Form 1979) study of oil refineries in Great
Britain and France found consistent differences in
Inkeles (1981) also describes various forms workers’ attitudes toward systems of authority. On
that divergence may take: (1) simple divergence, the question of sectoral and occupational shifts,
the mirror image of simple convergence, in which Gibbs and Browning’s (1966) twelve-nation study
movement occurs away from a common point of industrial and occupational division of labor
toward new points further apart than the original found both similarities—consistent with the con-
condition; (2) convergence with crossover, where vergence hypothesis—as well as differences. Stud-
lines intersect and then proceed to spread apart; ies across nations varying in levels of industrial
and (3) convergent trends masking underlying development revealed only ‘‘small and unsystematic
diversity (for example, although the United States, differences’’ in worker commitment (Form 1979,
Great Britain, and Sweden all experienced large p. 9), thus providing some support for the conver-
increases in public assistance programs from 1950 gence hypothesis. Japan has been regarded as an
to the early 1970s, the social groups receiving exceptional case among industrialized nations be-
benefits were quite different among the three cause of its strong cultural traditions based on
nations, as were the political dynamics associated mutual obligation between employers and em-
with the spending increases within each nation). ployees. These characteristics led Dore (1973), for
Finally, Inkeles (1981) notes the importance of example, to argue vigorously against the conver-
selecting appropriate units of analysis, levels of gence hypothesis for Japan. A more recent study
analysis, and the time span for which convergence, by Lincoln and Kalleberg (1990) ‘‘stands conver-
divergence, or parallel change can be assessed. gence on its head,’’ arguing that patterns of work
These comments echo earlier sentiments expressed organization in the United States are being im-
by Weinberg (1969) and Baum (1974) about how pelled in the direction of the Japanese model.
to salvage the useful elements of standard conver- Finally, with respect to women in the labor force,
gence theory while avoiding the pitfalls of a sim- the evidence of convergence is mixed. Some stud-
plistic functionalist-evolutionary approach. Com- ies found no relationship between female labor-
mon to these attempts to revive convergence theory force participation and level of industrialization
is the exhortation to develop more and better
424
CONVERGENCE THEORIES
(Ferber and Lowry 1977; Safilios-Rothchild 1971), societies, raising the question of whether conver-
though there is strong evidence of a trend toward gence theory or an explanation based on the func-
increasing female participation in nonagricultural tional imperatives of social structure of all com-
employment among advanced industrial societies plex societies, past or present, was most consistent
(Paydarfar 1967; Wilensky 1968) along with the with the empirical findings. The conclusion was
existence of dual labor markets stratified by sex, a that both explanations had some merit, since al-
pattern found in both communist and capitalist though all complex societies—whether developed,
nations in the 1970s (Cooney 1975; Bibb and Form undeveloped, or developing—showed similar oc-
1977; Lapidus 1976). cupational-prestige rankings, there was also evi-
dence that the more similar societies were in levels
STRATIFICATION of industrialization, the more similar their pat-
terns of occupational-prestige evaluation ap-
Closely related to the study of industrial organiza- peared to be.
tion is the question of converging patterns of
stratification and mobility. The attempt to discov- DEMOGRAPHIC PATTERNS
er common features of the class structure across
advanced industrial societies is a central concern The theory of demographic transition provides
for social theorists of many stripes. The question one of the most straightforward examples of con-
has inspired intense debate among both neo- vergence. The essence of the theory is that fertility
Weberian and Marxist sociologists, although the and mortality rates covary over time in a predict-
latter, for obvious ideological reasons, tend to able and highly uniform manner. Moreover, these
eschew the language of convergence theory. An changes are directly linked to broad developmen-
early statement of the class convergence thesis was tal patterns, such as the move from a rural, agricul-
made by Lipset and Zetterberg (1959), to the turally based economy to an urban-industrial one,
effect that observed rates of mobility between increases in per capita income, and adult literacy
social classes tend to be similar from one industrial (Berelson 1978). In the first stage of the demo-
society to another. Erikson et al. (1983) conducted graphic transition, both fertility rates and death
a detailed test of the class mobility convergence rates are high, with population remaining fairly
hypothesis in England, France, and Sweden, and constant. In the second stage, death rates drop (as
found little support for it. They conclude that the a result of improvements in living conditions and
‘‘process of industrialization is associated with medical care) while fertility rates remain high, and
very variable patterns . . . of the social division of population levels increase rapidly. In the third
labour’’ (p. 339). stage, fertility rates begin to decline, with total size
of the population leveling off or even decreasing.
A subcategory of comparative stratification This simple model works remarkably well in ac-
research concerns the evidence of convergence in counting for demographic patterns observed among
occupational prestige. A study published in 1956 all industrialized (and many industrializing) socie-
by Inkeles and Rossi, based on data from six ties during the post-World War II period. A large
industrialized societies, concluded that the pres- spread in fertility rates among nations at the begin-
tige hierarchy of occupations was ‘‘relatively in- ning of the 1950s gave way to declining rates of
variable’’ and tended to support the hypothesis fertility ending with a nearly uniform pattern of
that modern industrial systems are ‘‘highly coher- zero population growth in the 1970s.
ent. . . relatively impervious to the influence of
traditional culture patterns’’ (p. 329). Although The convergent tendencies predicted by the
the authors did not specifically mention conver- theory of demographic transition have not gone
gence, their conclusions were fully consistent with unchallenged, however. Freedman (1979), for ex-
the idea of emergent similarities. A subsequent ample, suggests that cultural factors mediate the
study by Treiman (1977) extended the compari- effects of social structural factors central to transi-
son of occupational prestige to some sixty nations, tion theory. Coale (1973) and Teitelbaum (1975)
ranging from the least developed to the most note that demographic transition theory has not
developed. The study found that occupational- provided much explanatory or predictive power
prestige rankings were markedly similar across all
425
CONVERGENCE THEORIES
with regard to the timing of population changes or that convergence is a response to pressures arising
the regional variations observed within nations from a complex, technologically advanced social
undergoing change. and economic system, they also identify diffusion
via integration of networks through which ideas,
FAMILY standards, and practices in education are shared.
These networks operate largely through interna-
Inkeles (1980) explored the effects of putative tional organizations, such as UNESCO and the
convergent tendencies discussed above for family OECD; their role as mediating structures in a
patterns. While he found evidence of convergence process leading toward cross-national similarities
in some aspects of family life, other patterns con- in education constitutes an important addition to
tinue ‘‘to be remarkably stable in the face of great convergence theory, with wide-ranging implica-
variation in their surrounding socio-economic con- tions for convergence in other institutions.
ditions’’ (p. 34). Aspects of family life that show
clear convergent patterns include the trend to- THE WELFARE STATE
ward falling fertility rates and a shift in relative
power and resource control in the direction of The development of the welfare state has inspired
increasing autonomy of women and declining au- active theoretical debate and empirical research
thority of parents. Other aspects of the family, on convergence theory, with researchers divided
such as age at first marriage, appear to present a over the nature and extent of convergence found
more complex picture, with short-term fluctua- across nations. On the one hand, there is indisput-
tions obscuring long-term changes, and great varia- able evidence that extensive social security, health
tion from one culture to another. Still other char- care, and related benefit programs is restricted to
acteristics of family life seem resistant to change; nations that have reached a level of economic
cited as examples in Inkeles (1980) are cultural development where a sufficient surplus exists to
patterns such as veneration of elders in many support such efforts. Moreover, the development
Asian societies, basic human needs for compan- of programs of the welfare state appears to be
ionship and psychological support, and the role of empirically correlated with distinct bureaucratic
husbands helping wives with housework. In all, and demographic patterns that are in turn grounded
Inkeles (1980) estimates that only about half the in economic development. For example, Wilensky
indicators of family life he examined showed any (1975) found that among sixty nations studied the
convergence, and even then not always of a line- proportion of the population sixty-five years of age
ar nature. and older and the age of social security programs
were the major determinants of levels of total
EDUCATION welfare-state spending as a percent of gross nation-
al product. Since levels of economic development
Following Inkeles’ (1981) reformulation of con- and growth of the elderly population both repre-
vergence theory, Inkeles and Sirowy (1983) stud- sent areas of convergence among advanced socie-
ied the educational systems of seventy-three rich ties, it is reasonable to expect that patterns of
and poor nations. Among thirty different ‘‘pat- welfare-state development will also tend to con-
terns of change’’ in educational systems examined, verge. Indeed, in such respects as the development
they found evidence of marked convergence in of large and expensive pension and health-care
fourteen, moderate convergence in four, consid- programs, of which the elderly are the major
erable variability in nine, mixed results in two, and clientele, this is the case (Coughlin and Armour
divergence in only one. Based on these findings, 1982; Hage et al. 1989). Other empirical studies
they conclude that the tendency toward conver- have found evidence of convergence in public
gence on common structures is ‘‘pervasive and attitudes toward constituent programs of the wel-
deep. It is manifested at all levels of the education- fare state (Coughlin 1980), in egalitarian political
al system, and affects virtually every major aspect movements affecting welfare effort across nations
of that system’’ (p. 326). Also worthy of note is that (Williamson and Weiss 1979), and in levels of
while the authors take the conventional position spending (Pryor 1968), normative patterns (Mishra
1976), and social control functions of welfare-state
426
CONVERGENCE THEORIES
programs across capitalist and communist nations Convergence theorists are surely on solid
(Armour and Coughlin 1985). ground when they assert that programs to
protect against the seven or eight basic risks of
Other researchers have challenged the idea of industrial life are primarily responses to
convergence in the welfare state. In a historical economic development . . . . However, showing
study of unemployment programs in thirteen West- that societies have adopted the same basic
ern European nations, Alber (1981) found no programs. . . is only a partial demonstration of
evidence that programs had become more alike in convergence insofar as it does not demonstrate
eligibility criteria, methods of financing, or gener- convergence in substantive features of the
osity of benefits, although he did find some evi- programs or in the amount of variation among
dence of convergence in duration of unemploy- affluent countries compared to poor countries.
ment benefits in nations with compulsory systems.
A study conducted by O’Connor (1988) testing the GLOBALIZATION
convergence hypothesis with respect to trends in
welfare spending from 1960 to 1980 concluded Growing attention to a variety of large-scale changes
that ‘‘despite the adoption of apparently similar in economic relations, technology, and cultural
welfare programmes in economically developed relations, broadly subsumed under the descrip-
countries there is not only diversity but divergence tion ‘‘globalization,’’ has inspired renewed inter-
in welfare effort. Further, the level of divergence is est in the ideas of convergence and modernity (see
increasing’’ (p. 295). A much broader challenge to Robertson 1992 for a critical account). The litera-
the convergence hypothesis comes from studies ture on globalization has several threads. One
focusing on variations in welfare-state develop- approach focuses on the economic and cultural
ment among western capitalist democracies. Hewitt impact of transnational capitalist enterprises that
(1977), Castles (1978, 1982), and Korpi (1983), to are judged to be responsible for the spread of a
cite a few leading examples, argue that variations pervasive ideology and culture of consumerism
across nations in the strength and reformist char- (Sklair 1995). Ritzer (1993) summarizes this phe-
acter of labor unions and social democratic par- nomenon as the ‘‘McDonaldization of society’’—a
ties account for large differences in the levels of broad reference to the ubiquity and influence of
spending for and redistributive impact of welfare- the consumer brand names (and the large corpo-
state programs. However, the disagreement among rate interests behind them) that are instantly rec-
these studies and scholars arguing for conver- ognizable in virtually every country in the world
gence may be simply a function of case selection. today. The main implication of this perspective is
For example, in a study of nineteen rich nations, that indigenous industries, habits, and culture are
Wilensky (1976, 1981) linked cross-national diver- rapidly being driven aside or even into extinction
sity in the welfare state to differences in ‘‘demo- by the ‘‘juggernaut’’ of the world capitalist econo-
cratic corporatism,’’ and secondarily to the pres- my dominated by a relatively few powerful interests.
ence of Catholic political parties, thus rejecting
the simplistic idea that the convergence observed Meyer et al. (1997) provide a different inter-
across many nations at widely different levels of pretation of globalization in their work on ‘‘world
economic development extends to the often diver- society.’’ Although they argue that ‘‘many features
gent policy developments in the relatively small of the contemporary nation-state derive from a
number of advanced capitalist societies. worldwide model constructed and propagated
through global cultural and associational process-
The debate over convergence in the welfare es’’ (pp. 144–145), the essence of their position is
state is certain to continue. A major obstacle in that nations are drawn toward a model that is
resolving the question involves disagreement on ‘‘surprisingly consensual. . .in virtually all the do-
the nations chosen for study, selection and con- mains of rationalized social life’’ (p. 145). Meyer et
struction of measures (see Uusitalo 1984), and al. contend that various core principles, such as
judgments about the time frame appropriate for a those legitimating human rights and favoring
definitive test of the convergence hypothesis. environmentalism, do not emerge spontaneously
Wilensky et al. (1985, pp. 11–12) sum up the mixed as an imperative of modernity, but rather diffuse
status of current research on convergence in wel-
fare-state development as follows:
427
CONVERGENCE THEORIES
rapidly among nations worldwide through the Union during the 1990s may cast serious doubt on
agency of international organizations, networks of the long-term prospects for convergence, develop-
scientists and professionals, and other forms of ments have clearly moved in that direction with
association. Although not referring specifically to astonishing speed.
convergence theory, this world society and culture
approach makes a strong case for the emergence The continuing movement toward unification
of widely shared structural and cultural similari- in Europe associated with the European Union
ties, many of which hold out the promise of im- (EU, formerly the European Community) repre-
provement, among otherwise diverse nation-states. sents another significant case of political and eco-
nomic convergence on a regional scale. The gradu-
The rapid growth of telecommunications and al abolition of restrictions on trade, the movement
computing technology, especially apparent in the of labor, and travel among EU nations (and not
emergence of the Internet as a major social and least of all the establishment of a single currency in
economic phenomenon of the 1990s, presents yet 1999), and the harmonization of social policies
another aspect of globalization that holds pro- throughout the EU, all signal profound changes
found implications for possible societal conver- toward growing convergence in the region that
gence. However important and wide-ranging, the promises to continue into the twenty-first century.
precise patterns that will ultimately emerge from
these technological innovations are not yet clear. CONCLUSION
While new computing and communication tech-
nologies compress the time and space dimensions The idea of convergence is both powerful and
of social interaction (Giddens 1990), and have the intuitively attractive to sociologists across a range
potential to undercut national identities and cul- of backgrounds and interests (Form 1979). It is
tural differences along the lines envisioned by difficult to conceive of an acceptable macro theory
McLuhan’s (1960) ‘‘global village,’’ the same for- of social change that does not refer to the idea of
ces of advanced technology that can level tradi- convergence in one way or another. Despite the
tional differences may ultimately reinforce the controversy over, and subsequent disillusionment
boundaries of nation, culture, and social class. For with, early versions of convergence theory in the
example, even as the computers and related com- study of modernization, and the often mixed re-
munication technologies become more ever more sults of empirical studies discussed above, it is
widely disseminated, access to and benefits from clear that the concept of societal convergence (and
the new technologies appear to be disproportion- convergence theories that allow for the possibility
ately concentrated among the ‘‘haves,’’ leaving the of divergence and invariance) provides a useful
‘‘have nots’’ more and more excluded from partici- and potentially powerful analytical framework with-
pation (Wresch 1996). Over time, such disparities in which to conduct cross-national studies across a
might well serve to widen differences both across broad range of social phenomena. Even where the
and within nations, thus leading toward diver- convergence hypothesis ultimately ends up being
gence rather than convergence. rejected, the perspective offered by convergence
theories can provide a useful point of departure
Finally, interest in convergence has also been for research. Appropriately reformulated, focused
given a boost by various political developments in on elements of the social system amenable to
the 1990s. In particular, the twin developments of empirical study, and stripped of the ideological
the collapse of communism in Eastern Europe and baggage associated with its earlier versions, con-
the Soviet Union and the progressive weakening vergence theories hold promise to advance the
of economic and political barriers in Europe are understanding of the fundamental processes and
notable in this regard. The demise of state social- regularities of social change.
ism has revived interest in the possibilities of
global economic and political convergence among REFERENCES
advanced industrial societies (see, for example,
Lenski et al. 1991, p. 261; Fukuyama 1992). Al- Alber, Jens 1981 ‘‘Government Responses to the Chal-
though ongoing economic and political turmoil in lenge of Unemployment: The Development of Un-
the ‘‘transition to capitalism’’ in the former Soviet employment Insurance in Western Europe. In Peter
428
CONVERGENCE THEORIES
Flora and Arnold J. Heidenheimer, eds., The Develop- Ferber, M., and H. Lowry 1977 ‘‘Women’s Place: Na-
ment of Welfare States in Europe and America. New tional Differences in the Occupational Mosaic.’’ Jour-
Brunswick, N.J.: Transaction Books. nal of Marketing 41:23–30.
Armer, Michael, and Allan Schnaiberg 1972 ‘‘Measur- Form, William 1979 ‘‘Comparative Industrial Sociology
ing Individual Modernity: A Near Myth.’’ American and the Convergence Hypothesis.’’ Annual Review of
Sociological Review 37:301–316. Sociology 4:1–25.
Armour, Philip K., and Richard M. Coughlin 1985 ——— 1976 Blue-Collar Stratification. Princeton, N.J.:
‘‘Social Control and Social Security: Theory and Princeton University Press.
Research on Capitalist and Communist Nations.’’
Social Science Quarterly 66:770–788. ———, and Bae Kyu Han 1988 ‘‘Convergence Theory
and the Korean Connection.’’ Social Forces 66:618–644.
Aron, Raymond 1967 The Industrial Society. New York:
Simon and Schuster. Fox, William S., and William W. Philliber 1980 ‘‘Class
Convergence: An Empirical Test.’’ Sociology and So-
Baum, Rainer C. 1974 ‘‘Beyond Convergence: Toward cial Research 64:236–248.
Theoretical Relevance in Quantitative Moderniza-
tion Research.’’ Sociological Inquiry 44:225–240. Freedman, Ronald 1979 ‘‘Theories of Fertility Decline:
A Reappraisal. Social Forces 58:1–17.
Bendix, Reinhard 1967 ‘‘Tradition and Modernity Re-
considered.’’ Comparative Studies in Society and History Fukuyama, Francis 1992 The End of History and the Last
9:292–346. Man. New York: Free Press.
Berelson, B. 1978 ‘‘Prospects and Programs for Fertility Galbraith, John Kenneth 1967 The New Industrial State.
Reduction: What? Where?’’ Population and Develop- Houghton-Mifflin.
ment Review 4:579–616.
Gibbs, J.P., and H.L. Browning 1966 ‘‘The Division of
Bibb, R. and W.H. Form 1977 ‘‘The Effects of Industrial, Labor, Technology, and the Organization of Produc-
Occupational, and Sex Stratification in Blue-Collar tion in Twelve Countries.’’ American Sociological Re-
Markets.’’ Social Forces 55:974–996. view 31:81–92.
Castles, Francis 1982 ‘‘The Impact of Parties on Public Giddens, Anthony 1990 The Consequences of Modernity.
Expenditure.’’ In Frances Castles, ed., The Impact of Stanford, Calif.: Stanford University Press.
Parties: Politics and Policies in Democratic Capitalist
States. Beverly Hills, Calif.: Sage. Goldthorpe, John H. 1971 ‘‘Theories of Industrial So-
ciety: Reflections on the Recrudescence of Histori-
——— 1978 The Social Democratic Image of Society. Lon- cism and the Future of Futurology.’’ Archives of Euro-
don: Routledge and Kegan Paul. pean Sociology 12:263–288.
Coale, Ansley J. 1973 ‘‘The Demographic Transition.’’ ——— 1964 ‘‘Social Stratification in Industrial Society.’’
In The International Population Conference, Liège. Liège, In P. Halmos, ed., The Development of Industrial Socie-
Belgium: International Union for the Scientific Study ties, 97–123. Keele, Eng.: University of Keele.
of Population.
Hage, Jerald, Robert Hanneman, and Edward T. Gargan
Cooney, R.S. 1975 ‘‘Female Professional Work Oppor- 1989 State Responsiveness and State Activism. London:
tunities: A Cross-National Study.’’ Demography Unwin Hyman.
12:107–120.
Hewitt, Christopher 1977 ‘‘The Effect of Political De-
Coughlin, Richard M. 1980 Ideology, Public Opinion and mocracy and Social Democracy on Equality in Indus-
Welfare Policy. Berkeley: University of California, In- trial Societies: A Cross-National Comparison.’’ Ameri-
stitute of International Studies. can Sociological Review 42:450–464.
———, and Philip K. Armour 1982 ‘‘Sectoral Differen- Inkeles, Alex 1981 ‘‘Convergence and Divergence in
tiation in Social Security Spending in the OECD Industrial Societies.’’ In M.O. Attir, B.H. Holzner,
Nations.’’ Comparative Social Research 6:175–199. and Z. Suda, eds., Directions of Change: Modernization
Theory, Research, and Realities, 3–38. Boulder, Colo.:
Davis, Kingsley 1948 Human Society. New York: Macmillan. Westview Press.
Dore, Ronald 1973 British Factory-Japanese Factory. Berke- ——— 1980 ‘‘Modernization and Family Patterns: A
ley: University of California Press. Test of Convergence Theory.’’ In D.W. Hoover and
J.T.A. Koumoulides, eds., Conspectus of History, 31–
Erikson, Robert, John H. Goldthorpe, and Lucienne 64. Cambridge, Eng.: Cambridge University Press.
Portocarero 1983 ‘‘Intergenerational Class Mobility
and the Convergence Thesis: England, France, and ——— 1966 ‘‘The Modernization of Man.’’ In M. Weiner,
Sweden.’’ British Journal of Sociology 34:303–343. ed., Modernization New York: Basic Books.
429
CONVERGENCE THEORIES
———, and Larry Sirowy 1983 ‘‘Convergent and Diver- Paydarfar, A.A. 1967 ‘‘Modernization Process and Demo-
gent Trends in National Educational Systems.’’ Social graphic Changes.’’ Sociological Review 15:141–153.
Forces 62:303–333.
Portes, Alejandro 1973 ‘‘Modernity and Development:
Inkeles, Alex, and David E. Smith 1974 Becoming Mod- A Critique.’’ Comparative International Development
ern. Cambridge, Mass.: Harvard University Press. 8:247–279.
Inkeles, Alex, and Peter H. Rossi 1956 ‘‘National Com- Pryor, Frederick 1968 Public Expenditure in Capitalist and
parisons of Occupational Prestige.’’ American Journal Communist Nations. Homewood, Ill.: Irwin.
of Sociology 61:329–339.
Ritzer, George 1993 The McDonaldization of Society.
Kahl, J.A. 1968 The Measurement of Modernization: A Newbury Park, Calif.: Sage.
Study of Values in Brazil and Mexico. Austin: University
of Texas Press. Robertson, Roland 1992 Globalization: Social Theory and
Global Culture. London and Newbury Park, Calif.: Sage.
Kerr, Clark, John T. Dunlap, Frederick H. Harbison,
and Charles A. Myers 1960 Industrialism and Industri- Rostow, Walt W. 1960 The Stages of Economic Growth.
al Man. Cambridge, Mass.: Harvard University Press. Cambridge, Eng.: Cambridge University Press.
Korpi, Walter 1983 The Democratic Class Struggle. Lon- Safilios-Rothchild, C. 1971 ‘‘A Cross-Cultural Examina-
don: Routledge and Kegan Paul. tion of Women’s Marital, Educational and Occupa-
tional Aspirations.’’ Acta Sociologica 14:96–113.
Lapidus, Gail W. 1976 ‘‘Occupational Segregation and
Public Policy: A Comparative Analysis of American Sklair, Leslie 1995 Sociology of the Global System. Balti-
and Soviet Patterns.’’ In M. Blaxell and B. Reagan, more, Md.: Johns Hopkins University Press.
eds., Women and the Workplace Chicago: University of
Chicago Press. Sorokin, Pitirim A. 1960 ‘‘Mutual Convergence of the
United States and the U.S.S.R. to the Mixed Sociocul-
Lenski, Gerhard, Jean Lenski, and Patrick Nolan 1991 tural Type.’’ International Journal of Comparative Soci-
Human Societies, 6th ed. New York: McGraw-Hill. ology. 1:143–176.
Lincoln, James R., and Arne L. Kalleberg 1990 Culture, Teitelbaum, M.S. 1975 ‘‘Relevance of Demographic Tran-
Control, and Commitment: A Study of Work Organization sition for Developing Countries.’’ Science 188:420–425.
and Work Attitudes in the United States and Japan.
Cambridge, Eng.: Cambridge University Press. Treiman, Donald J. 1977 Occupational Prestige in Com-
parative Perspective. New York: Academic Press.
Lipset, Seymour M., and Hans Zetterberg 1959 ‘‘Social
Mobility in Industrial Societies.’’ In S.M. Lipset and Uusitalo, Hannu 1984 ‘‘Comparative Research on the
R. Bendix, eds., Social Mobility in Industrial Socie- Determinants of the Welfare State: The State of the
ty.London: Heinemann. Art.’’ European Journal of Political Research 12:403–442.
McLuhan, Marshall 1960 Exploration in Communications. Wallerstein, Immanuel 1974 ‘‘The Rise and Future De-
Boston: Beacon Press. mise of the World Capitalist System: Concepts for
Comparative Analysis.’’ Comparative Studies in Social
Meyer, John W., John Boli, George M. Thomas, and History 16:287–415.
Francisco O. Ramirez 1997 ‘‘World Society and the
Nation-State.’’ American Journal of Sociology 103:144–181. Weed, Frank J. 1979 ‘‘Industrialization and Welfare
Systems: A Critical Evaluation of the Convergence
Mishra, Ramesh 1976 ‘‘Convergence Theory and Social Hypothesis.’’ International Journal of Comparative Soci-
Change: The Development of Welfare in Britain and ology 20:282–292.
the Soviet Union.’’ Comparative Studies in Society and
History 18:28–56. Weinberg, Ian 1969 ‘‘The Problem of the Convergence
of Industrial Societies: A Critical Look at the State of
Moore, Wilbert E. 1979 World Modernization: The Limits the Theory.’’ Comparative Studies in Society and Histo-
of Convergence. New York: Elsevier. ry 11:1–15.
O’Connor, Julia S. 1988 ‘‘Convergence or Divergence?: Wilensky, Harold L. 1981 ‘‘Leftism, Catholicism, and
Change in Welfare Effort in OECD Countries 1960– Democratic Corporatism: The Role of Political Par-
1980.’’ European Journal of Political Research 16:277–299. ties in Recent Welfare State Development.’’ In Peter
Flora and Arnold J. Heidenheimer, eds., The Develop-
Parsons, Talcott 1966 Societies: Evolutionary and Com- ment of Welfare States in Europe and America, 345–382.
parative Perspectives. Englewood Cliffs, N.J.: Pren- New Brunswick, N.J.: Transaction Books.
tice Hall.
——— 1976 The ‘‘New Corporatism’’: Centralization and
Parsons, Talcott 1951 The Social System. Glencoe, Ill.: the Welfare State. Beverly Hills, Calif.: Sage.
The Free Press.
430
CONVERSATION ANALYSIS
——— 1975 The Welfare State and Equality. Berkeley: order’’ is locally constituted (Garfinkel [1967] 1984).
University of California Press. In addressing this problem, he devised a remark-
ably innovative approach. Working with tapes and
——— 1968 ‘‘Women’s Work: Economic Growth, Ide- transcripts of telephone calls to a suicide preven-
ology, Structure.’’ Industrial Relations 7:235–248. tion center (and with recordings of other, some-
what more mundane sorts of conversations), Sacks
———, Gregory M. Luebbert, Susan Reed Hahn, and began examining the talk as an object in its own
Adrienne M. Jamieson 1985 Comparative Social Policy: right, as a fundamental type of social action, rather
Theory, Methods, Findings. Berkeley: University of Cali- than primarily as a resource for documenting
fornia, Institute of International Studies. other social processes. In short, Sacks came to
recognize that the talk itself was the action. It was
Williamson, John B., and Joseph W. Weiss 1979 ‘‘Egali- in the details of the talk that we could discover just
tarian Political Movements, Social Welfare Effort how what was getting done in the activity was
and Convergence Theory: A Cross-National Analy- accomplished, systematically and procedurally, then
sis.’’ Comparative Social Research 2:289–302. and there, by the coparticipants themselves. This
appeared to be an especially fruitful way of investi-
Williamson, John B., and Jeanne J. Fleming 1977 ‘‘Con- gating the local production of social order.
vergence Theory and the Social Welfare Sector: A
Cross-National Analysis.’’ International Journal of Com- As Schegloff (1989, p. 199) later wrote in a
parative Sociology 18:242–253. memoir of these first years, Sacks’s strategy in his
pioneering studies was to first take note of how
Wresch, William 1996 Disconnected: Haves and Have Nots members of society, in some actual occasion of
in the Information Age. New Brunswick, N.J.: Rutgers interaction, achieved some interactional effect—
University Press. for example, in the suicide center calls, how to
exhibit (and have others appreciate) that one has
RICHARD M. COUGHLIN reasonably, accountably, arrived at the finding ‘‘I
have no one to turn to’’—and then to ask: Was this
CONVERSATION ANALYSIS outcome accomplished methodically? Can we de-
scribe it as the product of a method of conduct,
Conversation analysis has evolved over several such that we can find other enactments of that
decades as a distinct variant of ethnomethodolo- method that will yield the same outcome, the same
gy. Its beginnings can be traced to the mid-1960s, recognizable effect? This approach provided, Sacks
to the doctoral research and the unpublished but suggested, an opportunity to develop formal ac-
widely circulated lectures of Harvey Sacks. Sacks counts of ‘‘members’ methods’’ for conducting
was a University of California sociologist who had social life.
studied with Harold Garfinkel, the founder of the
ethnomethodological movement, as well as with In this way, Sacks sought to address the basic
Erving Goffman. While not an ethnomethodologist, question of (as he put it in one of his early manu-
Goffman’s proposal that face-to-face interaction scripts) ‘‘what it is that sociology can aim to do,
could be an analytically independent domain of and . . . how it can proceed’’ (Sacks [1964–1968]
inquiry certainly helped inspire Sacks’s work. Two 1984, p. 21). Sociology, he argued, could be a
other key figures whose writings (separately and ‘‘natural observational science,’’ concerned with
together with Sacks) contributed to the emer- the methodic organization of naturally occurring
gence of conversation analysis were Gail Jefferson, events, rather than with behavior that was manipu-
one of Sacks’s first students, and Emanuel A. lated through experimental techniques or other
Schegloff, another sociologist trained in the Uni- interventions such as surveys, interviews, and the
versity of California system who was decisively like. And it could be committed to direct observa-
influenced by Garfinkel and, in much the same tion of this organization in situ, rather than de-
manner as Sacks, by Goffman (Schegloff 1988). pendent upon analytic theorizing and a concomi-
tant reliance on idealized models of action.
Sacks, like Garfinkel, was preoccupied with
discovering the methods or procedures by which Naturalistic observation also met the
humans coordinate and organize their activities, ethnomethodological mandate that all evidence
and thus with the procedures of practical, com-
mon-sense reasoning in and through which ‘‘social
431
CONVERSATION ANALYSIS
for the use of members’ methods, and for mem- and perhaps could not even be imagined, by con-
bers’ orientation to or tacit knowledge of them, ventional analytic sociology. When had sociolo-
was to be derived exclusively from the observed gists concerned themselves with the profoundly
behavior of the coparticipants in an interactional methodic character of things like how to avoid
event. As Schegloff and Sacks (1973, p. 290) subse- giving your name without refusing to give it, he
quently summarized the logic of this stance, if the argued, or with how to get help for suicidalness
event, the recorded conversational encounter, ex- without requesting it? Or with members’ methods
hibited a methodically achieved orderliness, it ‘‘did for things like ‘‘doing describing’’ and ‘‘recogniz-
so not only to us [the observing analysts], indeed ing a description,’’ methods that provide for hear-
not in the first place for us, but for the co-partici- ing the first two sentences from a story told by a
pants who had produced’’ it. After all, the task was young child—’’The baby cried. The mommy picked
to discover members’ methods for coordinating it up.’’—as saying: The mommy who picked up the
and ordering conversational events, and these could baby is the baby’s mommy, and she picked it up
not in any way be determined by analysts’ concep- because it was crying. Though apparently mun-
tual stipulations or deduced from inventive theories. dane, this observation provided the basis for a
series of investigations regarding members’ cate-
The contrast with other methods and approach- gorization methods and eventually came to pro-
es for studying interactional processes could not vide for an entirely different approach to studies
be sharper, particularly with those methods adopt- of social institutions and phenomena like race and
ed by Bales (1950) and Homans (1961) and their gender. Sacks noted that since every person can be
many followers, with their commitment to theo- categorized at any time in various ways (for exam-
retically derived and precise operational defini- ple, in terms of age, gender, or stage of life), a
tions of social phenomena as a prerequisite to any person’s use of, or reliance on, one category rather
scientific investigation (see Sacks 1992, vol. 1, p. 28 than another to guide his or her actions with
and p. 105, for his thinking on Bales and Homans). others must be grounded in one of a multitude of
But Sacks’s methodological stance contrasts even discoverable systems of relevance, some of which
with those traditions usually regarded as neigh- are potentially applicable in any situation, such as
bors of conversation analysis, such as symbolic age, race, and gender, and others that are more
interactionism and Goffman’s ‘‘micro-Durkheimian’’ limited in their use, such as occupationally defined
approach. These approaches assume that without categories or those made relevant by the organiza-
an analytically stipulated conceptual scheme, there tion of conversation itself, such as speaker/hearer,
is no orderliness (or the orderliness cannot be caller/called, and the like.
seen) in what Garfinkel (1991) terms ‘‘the ple-
num’’—the plentitude of members’ lived experience. As these last examples of categories suggest,
‘‘membership analysis’’—making sense of who
Sacks was also making a well-reasoned argu- someone is for the purpose of appropriately de-
ment for the importance of studying mundane signing some next action—is an unremitting prob-
conversation, directly confronting the belief that lem for members of a society, one that has to be
sociology’s overriding concern should be the study solved in real time, and for which there is no single
of ‘‘big issues’’; that is, the belief that the search for solution. Furthermore, who a person relevantly
social order should center on the analysis of large- ‘‘is’’ for the purposes of some next action can
scale, massive institutions. Social order, he insist- change from moment to moment. As a conse-
ed, can be found ‘‘at all points,’’ and the close quence, Sacks pointed out that sociologists can no
study of what from conventional sociology’s point longer innocently categorize populations whatev-
of view seemed like small (and trivial) phenome- er way they (and their theories) see fit. Instead,
na—the details of conversation’s organization— analysts must similarly demonstrate the relevance
might actually give us an enormous understanding of a category to the participants in any scene, as well
of the way humans do things and the kinds of as its consequentiality for them in terms of how the
methods they use to order their affairs (Sacks action proceeds, in order to ground its use in any
[1964–1968] 1984, p. 24). sociological investigation (see Silverman 1998 for
a useful summary of this argument).
This last proposition bears special emphasis,
for Sacks felt that these details went unnoticed,
432
CONVERSATION ANALYSIS
Such grounding would stand in contrast to the Psathas (1995). Sacks’s lectures have now been
epidemiological uses of categorization that under- edited and published in complete form (Sacks
pin standard social scientific research. Sacks’s col- 1992). Special issues of Sociological Inquiry, Social
league, anthropologist Michael Moerman (1974, Psychology Quarterly, Human Studies, Social Prob-
pp. 67–68), once noted that social scientists ‘‘have lems, Research on Language and Social Interaction,
an apparent inability to distinguish between warm Text, and the Western Journal of Speech Communica-
. . . human bodies and one kind of identification tion have also been devoted to ethnomethodological
device which some of those bodies sometimes and conversation-analytic topics.
use.’’ Further, since Sacks was primarily concerned
with categorization as a thoroughly practical, pro- Three major domains in conversation’s or-
cedural activity for members he was not much ganization identified in this literature are the or-
interested in the content of categories, with draw- ganization of sequences, of turn taking, and of
ing the cultural grids that preoccupied cognitive repair. These organizations can be described as
anthropologists and many social psychologists. In- systems of naturally organized activity, systems
stead, Sacks believed that by starting with the close known and used by members as courses of practi-
study of actual events—such as members’ observ- cal action and practical reasoning, and designed to
able use of categories in situ—and showing that resolve generic problems of coordination that con-
they happened in an endogenously, socially organ- front any conversationalist (and perhaps members
ized manner, a much sounder basis for studying of all social species). A sketch of some research
and understanding social life could be established. findings with respect to these organizations should
serve to illustrate how they function in this fash-
It should be evident, then, that the appellation ion, as well as the interlocking nature of their
conversation analysis does not really capture the domains.
enterprise’s commitment to addressing the most
basic problem for the social sciences: the underly- Consider first the organization of sequences.
ing character and structure of social action. As the Begin with the fact that even the most cursory
title of one of Sacks’s first publications, ‘‘An Initial inspection of conversational materials reveals that
Investigation of the Usability of Conversational talk-in-interaction has a serial arrangement to it.
Data for Doing Sociology,’’ makes clear, the use of For example, in a conversation between two par-
recorded conversational materials was more of an ties, party A will talk first, then party B, then A,
opportunistic research strategy than a commit- then B, and so forth. Accordingly, in two-party
ment to studying talk per se (Sacks 1972). Tape conversations, turns at talk constitute a series of
recordings of conversations constituted a record alternately produced utterances: ABABAB. But
of the details of actual, singular events that could overlaying this serial arrangement of utterances
be replayed and studied extensively, and would are distinctly characterizable conversational se-
permit other researchers direct access to exactly quences, where turns at talk do not simply happen
these same details. Still, for these identical rea- to occur one after the other but rather ‘‘belong
sons, it is conversation’s organization—its detect- together’’ as a socio-organizational unit, and where
able, orderly properties—that has remained the there is thus a methodic relationship between the
concrete object of study for the enterprise. various turns or parts.
During years since the ‘‘initial investigations,’’ This methodic, structurally linked relation-
conversation analysis has given rise to a substantial ship between sequence parts is central to how
research literature. Pursuing the lines of analysis sequences work in resolving coordination prob-
first identified in the early studies while simultane- lems in conversation. This point can be demon-
ously opening up many new avenues of inquiry, strated by briefly focusing on one of the earliest
researchers working in this tradition have pro- studies of sequence organization by Schegloff
duced findings that are, in the words of one con- (1968), an investigation into how the initiation of
temporary practitioner, ‘‘strikingly cumulative and conversational interactions is coordinated. Schegloff
interlocking’’ (Heritage 1987, p. 256). Important directed attention to a frequently occurring initial
collections of papers include those of Sudnow exchange, which was called a ‘‘summons–answer
(1972), Schenkein (1978), Atkinson and Heritage sequence.’’ This sequence is composed, he discov-
(1984), Button and Lee (1987), and ten Have and ered, of closely linked parts. The production of the
433
CONVERSATION ANALYSIS
first turn in the sequence, the summons, projected example, that for some firsts, there was not a single
a relevant next action, an answer, to be accom- appropriate second but rather a range of alterna-
plished by the recipient of the summons in the tive seconds. Note that in the examples of adjacen-
very next turn. Moreover, the occurrence of the cy pair structures listed just above, invitations
expected answer cannot properly be the final turn project either an acceptance or a declination as a
in the exchange. The summons–answer exchange course of action available to the recipient. In this
is therefore nonterminal: Upon production of the case, and in others like ‘‘request–granting/denial’’
answer, the summoner is then expected to speak and ‘‘compliment–acceptance/rejection,’’ it was
again, to provide the reason for the summons. found that the alternative second parts are not
This provides for a coordinated entry into conver- generally of equal status; rather, some second
sation, and for the possibility of an extended parts are preferred and others dispreferred, these
spate of talk. properties being distinct from the desires or moti-
vations of the coparticipants. ‘‘Preference’’ thus
Observe that a set of mutual obligations is refers to a structural rather than dispositional
established by the structural relationships between relationship between alternative but nonequiva-
these sequence parts, with each current action lent courses of action. Evidence for this includes
projecting some ‘‘next.’’ In the strongest form of distributional data across a wide range of speakers
these obligations (sequence classes vary in this and settings, and, more important, the fact that
regard), the property of ‘‘conditional relevance’’ preferred and dispreferred alternatives are regu-
holds between the parts of a sequence unit. A larly performed in distinctively different ways. The
‘‘summons–answer’’ sequence is but one type of a preference status of an action is therefore exhibit-
large class of utterance units, known as ‘‘adjacency ed in how it is done.
pairs,’’ that are characterized by this property.
Examples here include ‘‘greeting–greeting,’’ ‘‘ques- Related to this, conversation analytic research-
tion–answer,’’ and ‘‘invitation–acceptance/decli- ers observed that the producers of a first action
nation.’’ In adjacency pairs, when one utterance or often dealt in systematic, methodic ways with these
action is conditionally relevant on another, the properties of preference organization. To take
production of the first provides for the occurrence one example, the producer of a request can and
of the second. It could be said, then, using the often does analyze the recipient silence that fol-
example above, that the issuance of a summons is lows as displaying or implicating a denial—a deni-
an action that selects a particular next action, an al as-yet-unstated, but nevertheless projected—
answer, for its recipient. If this action does not and seeks to preempt the occurrence of this
occur, its nonoccurrence will be a noticeable event. dispreferred action by issuing a subsequent ver-
That is to say, it is not only nonoccurring, it is sion of the request, before the recipient starts to
notably, ‘‘officially’’ absent; accordingly, this would speak. Subsequent versions attempt to make the
warrant various inferences and actions. For in- request more acceptable and provide another op-
stance, the summoner might infer that a recipient portunity for a favorable response (Davidson 1984).
‘‘didn’t hear me,’’ which would provide for the
relevance and grounds of a repetition of the Moreover, members were observed to orient
summons. to the properties of preference organization through
their performance of actions plainly meant to be
The discovery that human activities like con- understood as specifically preliminary to some
versation were coordinated and organized in a adjacency pair first action. Such ‘‘pre"-type actions
very fundamental way by such methodic relation- are designed to explore the likelihood that pro-
ships between actions, with some current or ‘‘first’’ ducing that first part of some pair will not be
action projecting and providing for some appro- responded to in a dispreferred way. For instance,
priate ‘‘second,’’ led to investigations into the an utterance like ‘‘Are you doing anything to-
various methods by which the recipient of a first night?’’ provides, in a methodical way, an opportu-
may accomplish a second, or recognizably hold its nity for its producer to determine, without yet
accomplishment in abeyance until issues relevant having to actually issue the invitation, whether it
to its performance are clarified or resolved, or would most likely be declined. Similarly, this pro-
avoid its accomplishment altogether by undertak- vides an opportunity for the recipient of the ‘‘pre’’
ing some other activity. Researchers learned, for
434
CONVERSATION ANALYSIS
to indicate that a dispreferred action would be turn, is not predetermined or allocated in ad-
forthcoming without ever having to perform that vance. Instead, it is locally determined, moment-
action. Additionally, because ‘‘pre’’ actions them- by-moment, by the coparticipants in the talk. In
selves engender sequences by making some re- fact, this completely local determination of who
sponse to them a relevant next action, they consti- speaks when, how long they speak, and what they
tute the first part of a ‘‘pre-sequence.’’ It follows might say or do in their turn, is what provides for
that since these and other features of preference talk being hearable as a ‘‘conversation,’’ rather
organization together maximize the likelihood of than as, say, a debate or a ceremony of some kind.
preferred actions and minimize the likelihood of But this does not tell us just how—methodically—
dispreferred ones, they serve as important struc- speaker change is achieved such that, ordinarily,
tural resources for maintaining social solidarity one party talks at a time and there is little or no
and ‘‘preserving face.’’ silence (or ‘‘gap’’) between turns. Clearly, this
requires close coordination among coparticipants
These interrelated observations on the organi- in any conversational encounter. The systematic
zation of sequences were generalized outward in practices by which this is accomplished are ana-
conversation analytic research from the relatively lyzed by Sacks, Schegloff, and Jefferson in a 1974
simple adjacency pair organization by the recogni- paper that remains one of the most important in
tion that virtually every utterance occurs at some the conversation-analysis literature.
sequentially relevant, structurally defined place in
talk (see especially Atkinson and Heritage 1984, Basic to the accomplishment of turn taking is
pp. 5–9). Moreover, it is this placement that pro- the practice of changing speakers at possible utter-
vides the primary context for an utterance’s intelli- ance completion places, what Sacks, Schegloff,
gibility and understanding. Put another way, utter- and Jefferson term transition relevance places. How
ances are in the first place contextually understood are such places, where speaker change may relevantly
by reference to their placement and participation occur but is in no way guaranteed or required,
within sequences of action, and it is therefore discernable by members? A key feature of the
sequences of action, rather than single utterances units by and through which turns are constructed
or actions, that have become the primary units of offers one resource here: For an utterance to be
analysis for the conversation-analytic enterprise. usable as a turn constructional unit, it must have a
Accordingly, researchers in this tradition have not recognizable completion, and that completion must
restricted themselves to studying only especially be recognizable prior to its occurrence (Sacks,
‘‘tight’’ sequence units, but have instead broad- Schegloff, and Jefferson [1974] 1978, p. 12). That
ened their investigations to (mentioning just a is to say, its completion is projectable, and a
few) the sequencing of laughter, disputes, story coparticipant in the conversation who wishes to
and joke telling, political oratory, and the initia- speak next can therefore begin his or her turn just
tion and closing of topics. In addition, the sequen- at the place where the current speaker projects
tial organization of gaze and body movement in completion.
relation to turns at talk has been the focus of some
truly pathbreaking research using video record- Of course, this does not preclude this
ings (see, for example, Goodwin 1981, 1994; coparticipant, or any other, from starting to speak
Heath 1986). elsewhere in the course of a current speaker’s
turn. (Indeed, what actually constitutes a ‘‘turn at
Now let us consider the organization of turn talk’’ is as locally and mutually determined as any
taking, surely a central feature of virtually all talk- other aspect of conversation’s organization, even
in-interaction. Recall that the prior discussion on as the resources for doing so are general ones.)
the organization of sequences frequently made There are various interactional moves that could
reference to sequence parts as ‘‘turns,’’ implicitly involve, as one way they might be accomplished,
trading on the understanding that talk in conversa- this sort of action. At the same time, however,
tion is produced in and built for turns, with recur- research on turn-taking has revealed that turns
ring speaker change and a consequent serial or- beginning elsewhere may well be met with proce-
dering of utterances. In conversation, this turn dures systematically designed to enforce the prac-
ordering, as well as the size and content of each tice of starting at possible completion places. Fur-
ther, features of the turn taking system such as that
435
CONVERSATION ANALYSIS
described just above account for a great deal of the coordination of social activity, any such troubles
overlapping speech that can occasionally be ob- must be located and dealt with as quickly as possi-
served. For instance, a speaker might append a tag ble to avoid whole stretches of talk developing on a
question like ‘‘you know?’’ to his or her turn, while problematic basis. Finally, this set of methods
a coparticipant, having no resources available to must provide the opportunity to discover and
project such an action, starts to speak just prior to display trouble in speaking, hearing, or under-
or at the beginning of that appended tag, at the standing by any of the ratified coparticipants to
place that was projectably the ‘‘first possible com- the interaction, while simultaneously managing
pletion’’ of the turn. This would result in overlap- such trouble from the variety of quarters from
ping speech, with both parties talking simultane- which it might arise, whether the trouble is noticed
ously. This was just one example; studies of ‘‘more or produced by the current speaker or her recipi-
than one party at a time’’ speech have uncovered ent, and whether its source is endogenous to the
massive evidence that its occurrence and its resolu- interaction or impinges on it from outside.
tion (the restoration of one party at a time), as well
as the solution to the problem of which overlap- When Schegloff, Sacks, and Jefferson (1977)
ping action should then be consequential for next began examining the related set of practices
action, is methodically organized. through which speakers managed such troubles
they discovered two important features. First, they
Having described the function of turn con- noticed that participants in interaction treat the
structional practices in turn taking, Sacks, Schegloff, initiation of repair as a separate matter from the
and Jefferson still faced the issue of how actual accomplishment of a solution. That is, they
coparticipants, at possible completion places, de- distinguish between the various practices for locat-
termine just who will be the ‘‘next speaker’’ (note ing a trouble source and making it the focus of the
in this regard that conversation can involve more interaction and the set of practices for implement-
than two parties) or even if there will be a next ing a solution. Second, they observed that these
speaker, given that a current speaker might want two activities were not distributed evenly among
to continue talking. They discovered that to deal the parties: The organization of repair exhibited a
with this problem, members have available a ‘‘turn preference for self-repair and a preference for self-
allocational component’’ for the system. This com- initiation of repair. And they went on to show that
ponent consists of a set of ordered rules that come this latter feature is primarily a product of the way
into play at transition relevance places and which that the organization of repair relies on, and is
provide for the methodic allocation of the right to fitted to, the system for distributing turns.
produce a next turn, or more accurately, a turn
constructional unit. In related research, methods The organization of repair initiation operates
for securing the temporary suspension of turn- in a restricted ‘‘repair initiation opportunity space’’
taking procedures (to tell an extended story, for that is organized around the trouble source or
example) and for coordinating exit from the sys- ‘‘repairable.’’ Within this repair initiation oppor-
tem (to end the conversation) have been documented. tunity space each party to an interaction moves
through a series of discrete opportunities to locate
Finally there is the entire set of procedures by and indicate potential and actual troubles. In turn,
which any troubles in speaking, hearing, and un- these discrete opportunities to initiate repair shape
derstanding talk are systematically handled and where (relative to the trouble source) a repair is
‘‘repaired.’’ As Schegloff (1979, p. 269) points out, effected, and by whom. The current speaker has
insofar as ‘‘any of the systems and contingencies the first opportunity to initiate repair on any
implicated in the production and reception of trouble source within his or her own turn while
talk—articulatory, memory, sequential, syntactic, still in the midst of it, or just after it is complete but
auditory, ambient noise, etc.—can fail,’’ any piece before a next speaker starts. If they do initiate
of talk is susceptible to, or can reveal, troubles in repair during (or immediately following the possi-
speaking, hearing, or understanding. As a conse- ble completion of) their own turn, such speakers
quence, members of society must have some sys- also have the first opportunity to effect repair as well.
tematically organized set of methods for manag-
ing such trouble when it arises. Further, in order Of course, as we noted above, conversation is
for interaction to serve as a primary site for the characterized by the alternation between current
436
CONVERSATION ANALYSIS
and next, thus once a current speaker completes others. As Sacks, Schegloff, and Jefferson ([1974]
her turn a next speaker begins, typically by ad- 1978, p. 40) put it, the organization of turn taking
dressing herself to that just-prior talk. According- and the organization of repair ‘‘are thus ‘made for
ly, if a next speaker has some trouble with the prior each other’ in a double sense.’’ It is worth noting in
speaker’s turn, the next turn is the place where she this regard that insofar as interaction provides the
can initiate repair (using a variety of forms, includ- primary site for the achievement of intersubjectivity,
ing ‘‘what?’’ and ‘‘huh?’’ and other designs that for what makes sociality possible, the organization
vary in the degree to which they specify the exact of repair constitutes its last line of defense
source of trouble). By initiating repair using one of (Schegloff 1994).
these methods, that speaker selects the speaker of
the trouble source to speak next, and to offer a Taken together, the operation of the turn-
solution to the trouble indicated. If the next speak- taking system and the practices involved in the
er has no trouble with the prior turn, and she uses organization of sequences and repair account for
it to move the action forward (instead of stopping many of the detectable, orderly features of conver-
it to initiate repair), her turn will display a variety sation. This orderliness was shown to be locally
of understandings regarding the talk it follows. In organized and managed, the product of members’
doing so, her turn may also reveal some type of methods. It will be useful to make note once again
misunderstanding (from the point of view of the of the research strategy that enabled such findings.
speaker of the prior turn). If that occurs the speak- Because the data consisted of recordings of natu-
er of the prior turn can then initiate repair in ‘‘turn rally occurring activity, a scientific account of the
after next’’ (or ‘‘third position’’) and offer a solu- phenomenon under investigation could be em-
tion immediately. Perhaps the recurrent and rec- pirically grounded in the details of actual occur-
ognizable format for this is ‘‘I don’t mean x, I mean rences. The investigation began with a set of ob-
y.’’ servable outcomes of these occurrences—in the
case of turn taking, for example, speaker change
Thus, the movement of talk through these overwhelmingly recurred; overwhelmingly, one par-
three positions—current, next turn, and turn after ty talked at a time; turn order, size, and content
next—systematically provides the various parties were not fixed, but varied; and so on. It was then
to the interaction the opportunity to detect any asked: Could these outcomes be described as prod-
trouble in speaking, hearing, and understanding, ucts of certain social organized practices, of meth-
whatever its source, and initiate repair on it. As a ods of conduct? At the same time, if members of
consequence almost all instances of repair are society did in fact use such formal methods, how
initiated in one of these adjacent locations. The were they systematically employed to produce just
localization of repair initiation opportunities, and those outcomes, in just those occurrences, in all
the distribution of them over three turns, has their specificity? In addressing the problem in this
several consequences for the organization of so- way, then, conversation analysis was able to discov-
cial life. First, the localization of repair within a er how cardinal forms of social order were locally
finite, and relatively restricted, space ensures that constituted.
trouble is dealt with swiftly. Second, and related to
this, given the systematic relevance of repair, if The research on turn taking in conversation
speakers move through these three positions with- has provided one starting point for more recent
out any party initiating repair, a shared under- studies of interaction in ‘‘institutional’’ settings,
standing of the talk is thereby confirmed en passant. such as news interviews, doctor-patient and other
clinical consultations, courtrooms, plea bargain-
Finally, as with sequence organization, the ing sessions, job interviews, and citizen calls to
issue of preference is best grasped as a structural emergency services. In many of these studies, re-
property of the organization of talk-in-interaction searchers pursued Sacks, Schegloff, and Jeffer-
(rather than being a product of concerns regard- son’s ([1974] 1978, pp. 45–47) suggestion that the
ing the private desires of the parties). The two practices underlying the management of ordinary
preferences observed by Schegloff, Sacks, and Jef- conversation are the ‘‘base’’ or primary ones (for
ferson are a product of the distribution of oppor- an example, see Heritage and Maynard in press).
tunities to initiate and effect repair that systemati-
cally favors the speaker of the trouble source over
437
CONVERSATION ANALYSIS
Other forms of interaction—in this case, so-called This mode of research, with its commitment
‘‘institutional’’ forms—are in part constituted and to understanding precisely how any activity be-
recognizable through systematic variations from comes what it recognizably and accountably is—
conversational turn taking, or through the narrow- that is to say, how it acquires its social facticity-has
ing and respecification of particular conversation- tended to focus in the 1990s on work activities and
al practices involved in the organization of se- settings, under the rubric of ‘‘workplace studies.’’
quences, repair, and other activities. The scope of investigation has expanded to en-
compass all forms of ‘‘embodied action’’ (that is,
Take the case of courtroom interaction. The not only the talk), with extensive use of video
turn-taking system operative in these encounters recordings and, influenced by Suchman’s (1987)
places restrictions on turn construction and alloca- pioneering study of human-machine interaction,
tion: Coparticipants ordinarily restrict themselves with careful attention to how the machines, tech-
to producing turns that are at least minimally nologies, and other artifacts that saturate the mod-
recognizable as ‘‘questions’’ and ‘‘answers,’’ and ern work site are taken up and enter into the
these turn types are pre-allocated to different par- endogenous organization of work tasks (see, for
ties rather than locally determined. The relatively example, Whalen (1995), and the papers collected
restricted patterns of conduct observable in these in Luff, Hindmarsh, and Heath in press).
settings is, in large part, the product of this form of
turn taking. Accordingly, variation in turn taking Research into conversation’s organization al-
in such settings has been shown to have a ‘‘perva- so continues to evolve. While there has been rela-
sive influence both on the range and design of the tively little work that attempts to fundamentally
interactional activities which the different parties deepen the original account of the turn-taking
routinely undertake and on the detailed manage- system developed by Sacks, Schegloff, and Jeffer-
ment of such encounters’’ (Heritage 1987, p. 261; son (for a notable exception see Lerner 1996),
Atkinson and Drew 1979). there has been important research at the intersec-
tion of grammar and interaction—recognizing that
Note that throughout the above discussion, talk-in-interaction is in fact the natural home of
the term ‘‘institutional’’ has been presented with human language. This work demonstrates that the
quotation marks around it. This was done to em- approach to language taken Chomsky’s Transfor-
phasize ethnomethodology’s preoccupation with mational Grammar could be supplanted by one
the local production of social order. From this based on naturalistic study of the ‘‘grammar for
view, that some activity or encounter is recogniz- conversation.’’ Given that Chomsky’s approach
ably either an ‘‘ordinary conversation’’ or more has decisively shaped, both directly and indirectly,
‘‘institutional’’ in nature—for example, is recog- the understanding of language in cognitive sci-
nizably a ‘‘cross-examination,’’ a ‘‘call to the po- ence, psychology, computational linguistics, and
lice,’’ a ‘‘clinical consultation,’’ or whatever—is the other disciplines that rely on a model of gram-
something that the coparticipants can and do matical organization for their own research, these
realize, procedurally, at each and every moment of findings are plainly significant.
the encounter. The task for the analyst is to dem-
onstrate how they actually do this; how, for exam- Conversation analytic work on graamar and
ple, they construct their conduct, turn by turn, so interaction was launched by Schegloff’s (1979)
as to progressively constitute and thus jointly and paper on the ‘‘relevance of repair for a syntax for
collaboratively realize the occasion of their en- conversation.’’ This line of work has underscored
counter, together with their own social roles in it, the need for studies of language to draw on natu-
as having some distinctively institutional sense rally occurring spates of talk. As Schegloff ob-
(Heritage and Greatbatch 1991). Conversation ana- served, while nearly every episode of ordinary talk
lytic research on ‘‘institutional’’ interaction has contains instances of repair within the ‘‘sentenc-
therefore undertaken, through its investigations es’’ (or sentential turn constructional units) out of
into the methodic practices by which this gets which it is built, the entire view of language devel-
done, a systematic study of a wide range of human oped by linguists is based on imagined (or what
activities. might as well be imagined) instances of language
438
CONVERSATION ANALYSIS
that are free of such repair. Schegloff went on to analysis, interaction and sequences of turns ap-
show that most instances of talk-in-interaction, at pear to be that within which grammar is most
least in English, are organized by reference to the proximally organized. This appears to be true even
systematic relevance of repair, whether an instance at levels beneath the turn whether a sentence,
of it actually occurs in the sentence or not. clause, or phrase. As Schegloff (1996) shows, turn
beginnings and turn endings, as well as what hap-
Of course repair is not the only organization pens in between, are sites of strategic manipula-
relevant for grammar, and so more recently schol- tion. Through this manipulation, both grammati-
ars have begun to examine what more might be cal and prosodic, speakers fit their utterances to
learned about language by studying it as produced prior talk, launch new actions, and shape when
in naturally occurring interaction. With respect to they will be heard as possibly complete. Any scien-
this problem, conversation analysts have argued tific analysis of language, then, must take into
that insofar as language most likely evolved in face- account this central function.
to-face encounters by members of our species, its
structure and organization must have evolved, at Thus, as the collection of papers assembled in
least in part, to manage the basic exigencies con- Ochs, Schegloff, and Thompson (1996) suggest,
fronted by speakers and hearers. Thus, in addition rather than viewing grammar as an independent,
to the systematic relevance of repair, the structure clearly delineated, and internally coherent struc-
and organization of grammar most likely evolved ture, it is best approached as one more of the
as resources that shape, and are shaped by, how interrelated set of resources through which inter-
opportunities to speak are distributed, what action, and social life more broadly, is organized.
constraints are introduced by a current turn on
subsequent ones, and how speakers’ formulations REFERENCES
of the events, persons, and objects are organized.
Perhaps most developed are a series of findings Atkinson, J. Maxwell, and Paul Drew 1979 Order in
that link the organization of grammar and the Court: The Organization of Verbal Interaction in Judicial
system for distributing turns at talk briefly de- Settings. London: Macmillan.
scribed above.
Atkinson, J. Maxwell, and John Heritage (eds.) 1984
As we stated earlier regarding turn-construc- Structures of Social Action: Studies in Conversation Analy-
tional units, one of their key features is that each sis. Cambridge, Eng.: Cambridge University Press.
sentence, or utterance, projects from its beginning
roughly what it will take for it to be possibly ——— 1984 ‘‘Introduction.’’ In J. Maxwell Atkinson
complete. And over its course, each utterance and John Heritage, eds., Structures of Social Action:
projects in finer and finer detail the exact moment Studies in Conversation Analysis. Cambridge, Eng.:
that a speaker may end her utterance. Thus, in- Cambridge University Press.
stead of expressing logical predicates or cognitive
states, grammar may be best understood, in the Bales, Robert F. 1950 Interaction Process Analysis. Cam-
first instance, as a sequentially sensitive resource bridge, Mass.: Addison-Wesley.
that progressively projects the course and dura-
tion of turns at talk (Ford and Thompson 1996). Button, Graham, and John R. E. Lee (eds.) 1987 Talk
and Social Organization. Clevedon, Eng.: Multilingual
One of the most striking consequences of Matters.
such a view of grammar is that the locus of its
organization is transformed. While most approach- Davidson, Judy 1984 ‘‘Subsequent Versions of Invita-
es to grammar rely on the sentence as the basic tions, Offers, Requests, and Proposals Dealing with
unit of organization (with occasional nods to the Potential or Actual Rejection.’’ In J. Maxwell Atkinson
organization of ‘‘discourse’’), the grammatical units and John Heritage, eds., Structures of Social Action:
produced in interaction are fundamentally organ- Studies in Conversation Analysis. Cambridge, Eng.:
ized relative to their sequential environment, most Cambridge University Press.
proximally the just prior, current, and next turns.
Thus, rather than the sentence, or even discourse, Ford, Cecilia E., and Sandra A. Thompson 1996
being the fundamental unit or environment of ‘‘Interactional Units in Conversation: Syntactic,
Intonational, and Pragmatic Resources for the Man-
agement of Turns.’’ In Elinor Ochs, Emanuel A.
Schegloff and Sandra A. Thompson, eds., Interaction
439
CONVERSATION ANALYSIS
and Grammar. Cambridge, Eng.: Cambridge Univer- ——— (1964–1968) 1984 ‘‘Notes on Methodology.’’ In
sity Press. J. Maxwell Atkinson and John Heritage, eds., Struc-
tures of Social Action: Studies in Conversation Analysis.
Garfinkel, Harold 1991 ‘‘Respecification: Evidence for Cambridge, Eng.: Cambridge University Press.
Locally Produced, Naturally Accountable Phenome-
na of Order, Logic, Reason, Meaning, Method, etc. ——— 1972 ‘‘An Initial Investigation into the Usability
in and as of the Essential Haecceity of Immortal of Conversational Data for Doing Sociology.’’ In
Ordinary Society (I)—An Announcement of Stud- David Sudnow, ed., Studies in Social Interaction. New
ies.’’ In Graham Button, ed., Ethnomethodology and the York: Free Press.
Human Sciences. Cambridge, Eng.: Cambridge Uni-
versity Press. ———, Emanuel A. Schegloff, and Gail Jefferson (1974)
1978 ‘‘A Simplest Systematics for the Organization
——— (1967) 1984 Studies in Ethnomethodology. Cam- of Turn-Taking for Conversation.’’ In Jim Schenkein,
bridge: Polity Press. ed., Studies in the Organization of Conversational Inter-
action. New York: Academic Press.
Goodwin, Charles 1994 ‘‘Professional Vision.’’ American
Anthropologist 96:606–633. Schegloff, Emanuel A. 1996 ‘‘Turn Organization: One
Intersection of Grammar and Interaction.’’ In Elinor
——— 1981 Conversational Organization: Interaction Be- Ochs, Emanuel A. Schegloff, and Sandra A. Thomp-
tween Speakers and Hearers. New York: Academic Press. son, eds., Interaction and Grammar. Cambridge, Eng.:
Cambridge University Press.
Heath, Christian C. 1986 Body Movement and Speech in
Medical Interaction. Cambridge, Eng.: Cambridge Uni- ——— 1989 ‘‘Harvey Sacks—Lectures 1964–1965. An
versity Press. Introduction/Memoir.’’ Human Studies 12:185–209.
Heritage, John 1987 ‘‘Ethnomethodology.’’ In Anthony ——— 1988 ‘‘Goffman and the Analysis of Conversa-
Giddens and Jonathan H. Turner, eds., Social Theory tion.’’ In Paul Drew and Anthony Wootton, eds.,
Today. Stanford, Calif.: Stanford University Press. Erving Goffman: Exploring the Interaction Order. Bos-
ton: Northeastern University Press.
———, and David Greatbatch 1991 ‘‘On the Institution-
al Character of Institutional Talk: The Case of News ——— 1979 ‘‘The Relevance of Repair for Syntax-for-
Interviews.’’ In Deirdre Boden and Don H. Zimmerman, Conversation.’’ In T. Givon, ed., Syntax and Semantics
eds., Talk and Social Structure. Cambridge: Polity Press. 12: Discourse and Syntax. New York: Academic Press.
Heritage, John, and Douglas W. Maynard (eds.) In press ——— 1968 ‘‘Sequencing in Conversational Openings.’’
Practicing Medicine: Talk and Action in Primary Care American Anthropologist 70:1075–1095.
Encounters. Cambridge, Eng.: Cambridge Universi-
ty Press. ———, Gail Jefferson, and Harvey Sacks 1977 ‘‘The
Preference for Self-Correction in the Organization
Homans, George C. 1961 Social Behavior: Its Elementary of Repair in Conversation.’’ Language 53:361–382.
Forms. New York: Harcourt Brace.
Schegloff, Emanuel A., and Harvey Sacks 1973 ‘‘Open-
Lerner, Gene H. 1996 ‘‘On the ‘Semi-Permeable’ Char- ing Up Closings.’’ Semiotica 7:289–327.
acter of Grammatical Units in Conversation: Condi-
tional Entry into the Turn Space of Another Speak- Schenkein, Jim (ed.) 1978 Studies in the Organization of
er.’’ In Elinor Ochs, Emanuel A. Schegloff, and Sandra Conversational Interaction. New York: Academic Press.
A. Thompson, eds., Interaction and Grammar. Cam-
bridge, Eng.: Cambridge University Press. Silverman, David 1998 Harvey Sacks: Social Science and
Conversation Analysis. New York: Oxford Universi-
Luff, Paul, John Hindmarsh, and Christian Heath (eds.) ty Press.
In press Workplace Studies: Recovering Work Practice
and Informing Systems Design. Cambridge, Eng.: Cam- Suchman, Lucy A. 1987 Plans and Situated Actions: The
bridge University Press. Problem of Human-Machine Communication. Cambridge,
Eng.: Cambridge University Press.
Maynard, Douglas W. 1997 ‘‘The News Delivery Se-
quence: Bad News and Good News in Conversational Sudnow, David (ed.) 1972 Studies in Social Interaction.
Interaction.’’ Research on Language and Social Interac- New York: Free Press.
tion 30:93–130.
ten Have, Paul, and George Psathas 1995 Situated Order:
Ochs, Elinor, Emanuel A. Schegloff, and Sandra A. Studies in the Social Organization of Talk and Embodied
Thompson (eds.) 1996 Interaction and Grammar. Cam- Activities. Washington, D.C.: University Press of
bridge, Eng.: Cambridge University Press. America.
Sacks, Harvey 1992. Lectures on Conversation. 2 vols. Whalen, Jack 1995 ‘‘A Technology of Order Produc-
Oxford: Blackwell. tion: Computer-Aided Dispatch in 9-1-1 Communica-
tions.’’ In Paul ten Have and George Psathas, eds.,
440
CORPORATE ORGANIZATIONS
Situated Order: Studies in the Social Organization of Talk proprietorship and the partnership. In a proprietor-
and Embodied Activities. Washington, D.C.: University ship a particular person owns the property of the
Press of America. organization; in a partnership, two or more per-
sons share it. The right to handle the property and
JACK WHALEN affairs of the organization rests with a designated
GEOFF RAYMOND proprietor or set of partners. Significantly, pro-
prietors and partners bear personal responsibility
COOPERATION AND for the debts of the organization.
COMPETITION
The corporation constitutes a social inven-
See Small Groups. tion. The form evolved to handle problems that
arose within religious, political, and other kinds of
CORPORATE communities. It holds a place of importance in
ORGANIZATIONS contemporary Western societies. Because it is the
product of social conditions and an influence on
NOTE: Although the following article has not been revised for them, the corporation represents a topic of sub-
this edition of the Encyclopedia, the substantive coverage is stantial interest in sociology.
currently appropriate. The editors have provided a list of
recent works at the end of the article to facilitate research and At present, the corporation appears common-
exploration of the topic. ly within the world of business. But when the
corporation began to take shape during the Mid-
Societies carry out many of their activities dle Ages, the questions to be resolved lay outside
through formal organizations. Organizations are that realm. One of these questions had to do with
units in which offices, or positions, have distinct church ownership. In medieval Germany, land-
but interdependent duties. Organizations—hospi- owners often set up churches on their estates and
tals, schools, governments, business firms—share placed a priest in charge of them. As priests gained
certain features. Usually, at least one of the offices authority over their charges, they argued that the
serves as the linchpin: It coordinates the separate church and the land surrounding it no longer
duties within the organization. The key office has belonged to the donor. Deciding the true owner
ultimate authority in that the orders it issues con- proved to be difficult. A given priest could die or
strain the actions of lower-level offices. be replaced; hence, any particular priest seemed
to have no claim to ownership. One practice re-
But organizations also differ from one anoth- garded the owner to be the saint whose name the
er. In some, the assets belong to particular indi- church bore. Eventually, the idea developed that
viduals. In others, ownership resides in a collectivi- ownership inhered in the church, and that the
ty. The latter represents a corporate organization church constituted a body independent of its cur-
or corporation. Three features describe the modern rent leaders or members (Coleman 1974; Stone 1975).
corporation. First, it has certain legal rights and
privileges. By law, a corporation can sue and be Thorny problems also arose as medieval settle-
sued in the courts, make contracts, and purchase ments formed into towns. A town required some-
and receive property. Second, it usually exists in one to manage its affairs such as collecting tolls
perpetuity: It outlasts the individuals who set it up. and transacting other business. But the laws that
Ownership rests with stockholders, whose num- prevailed at the time applied only to individuals.
bers and makeup can change from one time to Any actions individuals took obligated them per-
another. Third, the owners have only a limited sonally. By this principle, managers would have to
responsibility for the obligations the corpora- meet any commitments they made on behalf of the
tion makes. town. To eliminate the dilemmas that the principle
situation posed, new laws made the town a corpo-
These features distinguish the corporate or- rate person. The corporate person would have all
ganization from two other forms of ownership: the the rights and privileges of any human being. This
action reduced the risks that public service might
otherwise entail. For many of the same reasons
441
CORPORATE ORGANIZATIONS
that the church and town became corporate per- A bureaucracy constitutes a particular mode
sons, the university of the Middle Ages moved that organizations can take. Consistent with all
towards the corporate form. organizations, bureaucratic ones divide up duties.
Two features separate a bureaucracy from other
The early corporations played rather passive modes, however. First, a system of ranks or levels
roles. Essentially, they held property for a collec- operates. Second, fixed rules and procedures gov-
tive, whose members might change from time to ern actions. The rules define the tasks, responsi-
time. Contrastingly, the corporations of the twen- bilities, and authority for each office and each level.
tieth century constitute spirited forces. They hire
multitudes of employees. They produce goods Few of the factories in nineteenth-century
and services and mold ideals and tastes. The deci- America operated as bureaucracies. Instead, the
sions their leaders make about where to locate individuals who made the products decided how
often determine which locales will prosper and the work would be done. A minimum number of
which will languish. levels existed. Supervisors or foremen hired and
fired workers, but workers made the rules on the
The influence that corporations have produc- work itself. The workers were craftsmen or arti-
es concerns about the control of them. Much of sans, and they contended that only those who
the work on corporations that sociologists have possessed the skills that the work demands should
undertaken highlights these concerns. The work decide how or if it should be divided. Gradually,
on control and corporations covers three topics: machinery took over the skilled work. Machines
the means through which corporations control and not workers controlled the pace. By the end of
their employees; the allocation of control between the 1920s, neither the laborers nor the machinery
owners and managers; and the extent to which shaped the work. Professional managers did. These
societies control corporations. For all three topics, managers enforced rules and oversaw an organiza-
control implies command over the affairs of and tion where specialized tasks and graded authority
operations within the corporate organization. prevailed (Nelson 1975; Clawson 1980; and
Jacoby 1985).
CONTROL OVER EMPLOYEES
The corporation of the late twentieth century
The corporate form has a long history, yet it did continues to operate as a bureaucracy. Some sources
not typify the early factories that manufacturers argue that efficiency explains the adoption of the
established in the United States. Before the early bureaucratic model (see especially Chandler 1980,
1900s, most factories operated as small operations 1984). Others challenge the emphasis on efficien-
under the control of a single entrepreneur. The cy, charging it with being overly rational or too
entrepreneur hired an overseer who might in turn apolitical. The first challenge appears most nota-
choose a foreman to hire, discipline, and fire bly in the work on organizations as institutions.
workers. Through consolidation and merger, the This literature regards survival as the premier goal
economic landscape of the 1920s revealed far for any organization. The closer an organization
more large organizations than had the tableau of a approximates an institution—an element taken
half-century earlier. for granted in the society—the greater its chances
for survival.
More changed over the years of the late nine-
teenth and early twentieth centuries than just the According to the institutional perspective, or-
size of organizations. The corporate form spread; ganizations adopt practices that appear to be rea-
the faceless corporation replaced the corporeal sonable. Myths develop about which patterns prove
entrepreneur. Corporations moved towards pro- most useful and efficient, and any organization
fessional management. Factories that businessmen that does not adopt a pattern that the myth favors
once controlled personally now operated through courts failure (Meyer and Rowan 1977; DiMaggio
abstract rules and procedures. The people whom 1988; DiMaggio and Powell 1983; Tolbert and
the workers now contacted on a regular basis Zucker 1983; also see Scott 1987 for a review of the
consisted of staff for the corporation and not the different branches of institutional theory).
corporate owners themselves. Bureaucratic tenets
took root. A different argument maintains that the em-
phasis on efficiency fails to capture the politics of
442
CORPORATE ORGANIZATIONS
corporations. This perspective treats corporations heeded the call from Zeitlin for research on the
as systems in which the interests of owners clash networks that link shareholders. Useem conclud-
with those of workers. Owners, it asserts, seek to ed from his study on contacts and networks among
reduce uncertainties and to eliminate the vagaries large shareholders that a corporate community
that can plague organizations. From this angle, operated, held together by an inner circle whose
bureaucracy serves the interests of owners prima- interests transcended company, region, and indus-
rily because it reduces the influence that workers try lines. Beth Mintz and Michael Schwartz (1985)
exercise and thereby removes a source of uncer- examined the connections between financial insti-
tainty (Braverman 1974; Edwards 1979). tutions and other corporations and decided that
control over corporate directions rested dispro-
Workers need not have formal authority in portionately in the world of finance. The work
order to affect outcomes within organizations. from the critics cautions us against the assumption
Studies document the creative ways in which em- that a multiplicity of owners implies control by
ployees enliven monotonous jobs and pursue their managers.
own ends (Roy 1952; Mechanic 1962; Burawoy
1979, 1985). Yet, officially, the higher levels have SOCIAL CONTROL OVER CORPORATIONS
greater power than have the lower levels. This is
the consequence of the bureaucratic nature of The corporate form constitutes a remarkable in-
corporations, not of their pattern of ownership. novation. But as the corporation has become ever
The bureaucratic mode is not unique to corpora- more active and entrenched, it has generated prob-
tions. Proprietorships and partnerships can dis- lems for society. Corporations have at times en-
play the traits of bureaucracy. The diffuseness of gaged in criminal behavior (Sutherland 1949;
ownership that one finds in the corporation possi- Clinard and Yeager 1980). At other times, their
bly makes formal control less obvious than obtains actions have violated no law but have put the well-
when ownership resides in identifiable persons. being of the public at risk. Both situations often
show the inadequacy of the mechanisms through
OWNER VERSUS MANAGERIAL CONTROL which society attempts to control corporations.
Managers occupy important places in the contem- Corporations are creatures of the state. Osten-
porary organization. One argument regards man- sibly, then, they operate only at the indulgence of
agers as more powerful than stockholders. Adolph the state. But myriad corporations now have great-
Berle and Gardiner Means offered this argument er resources than do the states that chartered
in the 1930s. As Berle and Means saw the situation, them. Moreover, the laws that states have at their
stockholding had become too widely dispersed for disposal often fit individuals better than they do
any individual holder or even group of holders to corporations. Corporations can be sued for wrong-
command corporations. Managers, they contend- doing; but a fine that would bankrupt an individu-
ed, filled the void (Berle and Means 1932). Later al might be a mere pittance for a large corpora-
discussions echoed the thesis that the expansion of tion. Both James Coleman (1974) and Christopher
the corporate form had raised the power of corpo- Stone (1975) have argued that the law can never be
rate managers (Berg and Zald 1978; also see Chan- the sole means for controlling corporations; a
dler 1962, 1977). sense of responsibility to the public must prevail
within corporations.
Critics contend that the thesis overstates the
role and power of managers. They base their Even if the law were shown to be effective in
criticism on studies of the influence that corporate constraining corporations within a state, it might
leaders wield. Maurice Zeitlin (1974) helped launch prove rather impotent in the case of multinational
this line of research when he argued that few organizations. A multinational or transnational
scholars had tested the Berle and Means thesis and corporation holds a charter from one nation-state
that the handful of extant studies showed owners but transacts business in at least one other. The
to be less fractious and fractionated than the thesis governmental entity that issues the charter cannot
supposed. Michael Useem (1984), among others, alter the policies the corporation pursues in its
443
CORPORATE ORGANIZATIONS
other locales. In addition, the very size of many The perspective known as population ecology
multinationals restricts the pressure that either likewise connects the destiny of organizations to
the home or the host country can impose. conditions in their surroundings. Population ecolo-
gists think of organizations as members of a popu-
Through various actions corporations demon- lation. Changing social conditions can enrich or
strate that they are attentive to the societies they impoverish a population. Individual units within it
inhabit. Corporate leaders serve on the boards of can do little to offset the tide of events that threat-
social service agencies; corporate foundations pro- ens to envelop the entire population. (Hannan
vide funds for community programs; employees and Freeman 1988; Wholey and Brittain 1989; for
donate their time to local causes. The agenda of a critique of the approach see Young 1988).
corporations long have included these and similar
activities. Increasingly, the agenda organize such Neither resource dependency theory nor popu-
actions around the idea of corporate social respon- lation ecology theory focuses explicity on the cor-
sibility. Acting responsibly means taking steps to porate form. But just as analyses of corporations
promote the commonweal (Steckmest 1982). inform the discussions sociologists have undertak-
en on formal organizations, models drawn from
Some corporations strive more consistently to studies of organizations have proved useful as
advance social ends than do others. Differences in scholars have tracked the progress of corporations.
norms and values apparently explain the contrast.
Norms, or maxims for behavior, indicate the cul- The corporation clearly constitutes a power to
ture of the organization (Deal and Kennedy 1982). be reckoned with. As with its precursors, the mod-
The culture of some settings gives the highest ern corporation serves needs that collectivities
priority to actions that protect the health, safety, develop. In fact, the corporation rests on an as-
and welfare of citizens and their heirs. Elsewhere, sumption that is fundamental in sociology: A col-
those are not what the culture emphasizes (Clinard lectivity has an identity of its own. But the corpora-
1983; Victor and Cullen 1988). tion of the twentieth century touches more than
those persons who own its assets or produce its
The large corporation had become such a goods. This social instrument of the Middle Ages is
dominant force by the 1980s that no one envi- now a social fixture.
sioned a return to an era of small, diffuse organiza-
tions. Yet, during that decade some sectors had (SEE ALSO: Capitalism; Organizational Effectiveness;
started to move from growth to contraction. At Organizational Structure; Transnational Corporations)
times, the shift resulted from legislative action.
When the Bell Telephone System divided in 1984, REFERENCES
by order of the courts, the change marked a sharp
reversal. For more than a century the system had Antai, Ariane-Berthoin 1992 Corporate Social Perform-
glided toward integration and standardization ance: Rediscovering Actors in Their Organizational Con-
(Barnett and Carroll 1987; Barnett 1990). texts. Boulder, Colo.: Westview Press.
Whether through fiat or choice, corporations Barnett, William P. 1990 ‘‘The Organizational Ecology
contract (Whetten 1987; Hambrick and D’Aveni of a Technological System.’’ Administrative Science
1988). Two perspectives associate the rise and fall Quarterly 35:31–60.
in the fortunes of corporations to changes in the
social context. The first perspective, resource de- Barnett, William P., and Glenn Carroll 1987 ‘‘Competi-
pendence, centers on the idea that organizations tion and Mutualism Among Early Telephone Com-
must secure their resources from their environs panies.’’ Administrative Science Quarterly 32:400–421.
(McCarthy and Zald 1977; Jenkins 1983). When
those environs contain a wealth of resources— Berg, Ivar, and Mayer N. Zald 1978 ‘‘Business and
personnel in the numbers and with the qualifica- Society.’’ Annual Review of Sociology 4:115–143.
tions the organization requires, funds to finance
operations—the corporation can thrive. When hard Berle, Adolph A., and Gardiner C. Means 1932 The
times plague the environs, the corporation es- Modern Corporation and Private Property. New York:
capes that fate only with great difficulty. Macmillan.
Braverman, Harry 1974 Labor and Monopoly Capital: The
Degradation of Work in the Twentieth Century. New
York: Monthly Review.
444
CORPORATE ORGANIZATIONS
Breedlove, William L., and J. Michael Armer 1996 ‘‘Eco- Institutional Patterns: Environment and Culture. Cam-
nomic Disarticulation and Social Development in bridge, Mass.: Ballinger.
Less-Developed Nations: A Cross-National Study of
Intervening Structures.’’ Sociological Focus 29:359–378. ———, and Walter Powell 1983 ‘‘The Iron Cage Revisit-
ed: Institutional Isomorphism and Collective Ration-
Burawoy, Michael 1979 Manufacturing Consent: Changes ality in Organizational Fields.’’ American Sociological
in the Labor Process under Monopoly Capitalism. Chica- Review 48:147–160.
go: University of Chicago Press.
Donaldson, Thomas 1992 ‘‘Individual Rights and Multi-
———1985 The Politics of Production: Factory Regimes national Corporate Responsibilities.’’ National Fo-
under Capitalism and Socialism. London: Verso. rum 72: 7–9.
Chandler, Alfred D., Jr. 1962 Strategy and Structure. Dunford, Louise, and Ann Ridley 1996 ‘‘No Soul to Be
Cambridge, Mass.: MIT Press. Damned, No Body to Be Kicked: Responsibility,
Blame and Corporate Punishment.’’ International Jour-
———1977 The Visible Hand. Cambridge, Mass.: Har- nal of the Sociology of Law 24:1–19.
vard University Press.
Dunworth, Terence, and Joel Rogers 1996 ‘‘Corpora-
———1980 ‘‘The United States: Seedbed of Managerial tions in Court: Big Business Litigation in U.S. Feder-
Capitalism.’’ In A. D. Chandler, Jr., and Herman al Courts, 1971–1991.’’ Law and Social Inquiry
Daems, eds., Managerial Hierarchies. Cambridge, Mass.: 21: 497–592.
Harvard University Press.
Edwards, Richard 1979 Contested Terrain: The Transfor-
———1984 ‘‘The Emergence of Managerial Capital- mation of the Workplace in the Twentieth Century. New
ism.’’ Business History Review 58:484–504. York: Basic Books.
Clawson, Dan 1980 Bureaucracy and the Labor Process. Etzioni, Amitai 1993 ‘‘The U.S. Sentencing Commission
New York: Monthly Labor Review. on Corporate Crime: A Critique.’’ Annals of the Ameri-
can Academy of Political and Social Science 525: 147–156.
Clinard, Marshall B. 1983 Corporate Ethics and Crime.
Beverly Hills, Calif.: Sage. Hambrick, Donald C., and Richard A. D’Aveni 1988
‘‘Large Corporate Failures as Downward Spirals.’’
———, and Peter C. Yeager 1980 Corporate Crime. New Administrative Science Quarterly 33:1–23.
York: Free Press.
Hannan, Michael, and John Freeman 1988 Organization-
Coleman, James S. 1974 Power and the Structure of Society. al Ecology. Cambridge, Mass.: Harvard University Press.
New York: Norton.
Hawley, James P. 1995 ‘‘Political Voice, Fiduciary Activ-
Coleman, James S. 1996 ‘‘The Asymmetric Society: Or- ism, and the Institutional Ownership of U.S. Corpo-
ganizational Actors, Corporate Power, and the Ir- rations: The Role of Public and Noncorporate Pen-
relevance of Persons.’’ In M. David Ermann, and sion Funds.’’ Sociological Perspectives 38:415–435.
Richard J.Lundman, eds., Corporate and Governmental
Deviance: Problems of Organizational Behavior in Con- Ireland, Paddy 1996 ‘‘Corporate Governance, Stakeholding,
temporary Society. New York: Oxford University Press. and the Company: Towards a Less Degenerate Capi-
talism?’’ Journal of Law and Society 23:287–320.
Cuthbert, Alexander R., and Keith G. McKinnell 1997
‘‘Ambiguous Space, Ambiguous Rights-Corporate Jacoby, Sanford 1985 Employing Bureaucracy: Managers,
Power and Social Control in Hong Kong.’’ Cities Unions and the Transformation of Work in American
14: 295–311. Industry, 1900–1945. New York: Columbia Universi-
ty Press.
Davis, Gerald F. 1994 ‘‘The Corporate Elite and the
Politics of Corporate Control.’’ Current Perspectives in Jenkins, J. Craig 1983 ‘‘Resource Mobilization Theory
Social Theory 1: 215–238. and the Study of Social Movements.’’ Annual Review
of Sociology 42:249–268.
Davis, Gerald F., Kristina A. Diekmann, and Catherine
H. Tinsley. 1994. ‘‘The Decline and Fall of the Con- Jones, Marc T. 1996 ‘‘The Poverty of Corporate Social
glomerate Firm in the 1980s: The Deinstitutionalization Responsibility.’’ Quarterly Journal of Ideology 19:1–2.
of an Organizational Form.’’ American Sociological
Review 59: 547–570. Lofquist, William S. 1993 ‘‘Legislating Organizational
Probation: State Capacity, Business Power, and Cor-
Deal, Terrence, and Allen Kennedy 1982 Corporate Cul- porate Crime Control.’’ Law and Society Review
ture. Reading, Mass.: Addison-Wesley. 27: 741–783.
DiMaggio, Paul 1988 ‘‘Interest and Agency in Institu- Luchansky, Bill, and Gregory Hooks 1996 ‘‘Corporate
tional Theory.’’ In Lynne G. Zucker, ed., Research on Beneficiaries of the Mid-Century Wars: Respecifying
445
CORRELATION AND REGRESSION ANALYSIS
Models of Corporate Growth, 1939–1959.’’ Social Sutherland, Edwin H. 1949 White Collar Crime. New
Science Quarterly 77:301–313. York: Holt, Rinehart, and Winston.
McCarthy, John D., and Mayer N. Zald 1977 ‘‘Resource Tolbert, Pamela S., and Lynne G. Zucker 1983 ‘‘Institu-
Mobilization and Social Movements: A Partial Theo- tional Sources of Change in the Formal Structure of
ry.’’ American Journal of Sociology 82:1212–1241. Organizations: The Diffusion of Civil Service Re-
form, 1880–1935.’’ Administrative Science Quarterly
McKinlay, Alan, and Ken Starkey 1998 Foucalt, Manage- 28:22–39.
ment and Organizations Theory: From Panopticion to
Technologies of Self. London: Sage Publications. Useem, Michael 1984 The Inner Circle: Large Corporations
and the Rise of Political Activity in the U.S. and U.K. New
Mechanic, David 1962 ‘‘Sources of Power of Lower York: Oxford University Press.
Participants in Complex Organizations.’’Administra-
tive Science Quarterly 7:349–364. Victor, Bart, and John B. Cullen 1988 ‘‘The Organiza-
tional Bases of Ethical Work Climates.’’ Administra-
Meyer, John, and Brian Rowan 1977 ‘‘Institutionalized tive Science Quarterly 33:101–125.
Organizations: Formal Structure as Myth and Cere-
mony.’’ American Sociological Review 83:340–363. Westphal, James D., and Edward J. Zajac 1997 ‘‘Defec-
tions from the Inner Circle: Social Exchange, Reci-
Mintz, Beth, and Michael Schwartz 1985 The Power procity, and the Diffusion of Board Independence in
Structure of American Business. Chicago: University of U.S. Corporations.’’ Administrative Science Quarterly
Chicago Press. 42:161–183.
Nelson, Daniel 1975 Managers and Workers. Madison: Whetten, David A. 1987 ‘‘Organizational Growth and
University of Wisconsin Press. Decline Processes.’’ Annual Review of Sociology
13:335–358.
Parker-Gwin, Rachel, and William G. Roy 1996 ‘‘Corpo-
rate Law and the Organization of Property in the Wholey, Douglas, and Jack Brittain 1989 ‘‘Characteriz-
United States: The Origin and Institutionalization of ing Environmental Variation.’’ Academy of Manage-
New Jersey Corporation Law, 1888–1903.’’ Politics ment Journal 32:867–882.
and Society 24:111–135.
Wilks, Stephen 1992 ‘‘Science, Technology and the
Prechel, Harland 1994 ‘‘Economic Crisis and the Cen- Large Corporation.’’ Government and Opposition
tralization of Control over the Managerial Process: 27:190–212.
Corporate Restructuring and Neo-Fordist Decision-
Making.’’ American Sociological Review 59:723–745. Young, Ruth 1988 ‘‘Is Population Ecology a Useful
Paradigm for the Study of Organizations?’’ American
Ridley, Ann, and Louise Dunford 1994 ‘‘Corporate Journal of Sociology 94:1–24.
Liability for Manslaughter: Reform and the Art of
the Possible.’’ International Journal of the Sociology of Zeitlin, Maurice 1974 ‘‘Corporate Ownership and Con-
Law 22:309–328. trol: The Large Corporation and the Capitalist Class.’’
American Journal of Sociology 79:1073–1119.
Rodman, Kenneth A. 1995 ‘‘Sanctions at Bay? Hegemonic
Decline, Multinational Corporations, and U.S. Eco- Zey, Mary, and Brande Camp 1996 ‘‘The Transforma-
nomic Sanctions since the Pipeline Case.’’ Interna- tion from Multidimensional Form to Corporate
tional Organization 49:105–137. Groups of Subsidies in the 1980s: Capital Crisis
Theory.’’ Sociological Quarterly 37:327–351.
Roy, Donald 1952 ‘‘Restriction of Output in a Piecework
Machine Shop.’’ Ph.D. diss., University of Chicago, CORA B. MARRETT
Chicago.
CORRECTIONS SYSTEMS
Sanders, Joseph, and V. Lee Hamilton 1996 ‘‘Distribut-
ing Responsibility for Wrongdoing inside Corporate See Penology; Criminal Sanctions; Criminology.
Hierarchies: Public Judgments in Three Societies.’’
Law and Social Inquiry 21:815–855. CORRELATION AND
REGRESSION ANALYSIS
Scott, W. Richard 1987 ‘‘The Adolescence of Institu-
tional Theory.’’ Administrative Science Quarterly In 1885, Francis Galton, a British biologist, pub-
32:493–511. lished a paper in which he demonstrated with
graphs and tables that the children of very tall
Steckmest, Francis W. 1982 Corporate Performance: The
Key to Public Trust. New York: McGraw-Hill.
Stone, Christopher D. 1975 Where the Law Ends: The
Social Control of Corporate Behavior. New York: Harper.
446
CORRELATION AND REGRESSION ANALYSIS
parents were, on average, shorter than their par- Assume that each of the |X values (i.e., the mean
ents, while the children of very short parents income for each level of education) falls on a
tended to exceed their parents in height (cited in straight line, as in Figure 1. This straight line is the
Walker 1929). Galton referred to this as ‘‘rever- regression line of Y on X. Thus the regression line of
sion’’ or the ‘‘law of regression’’ (i.e., regression to Y on X is the line that passes through the mean Y
the average height of the species). Galton also saw for each value of X—for example, the mean in-
in his graphs and tables a feature that he named come for each educational level.
the ‘‘co-relation’’ between variables. The stature of
kinsmen are ‘‘co-related’’ variables, Galton stated, If this regression line is a straight line, as
meaning, for example, that when the father was shown in Figure 1, then the income associated
taller than average, his son was likely also to be with each additional year of school completed is
taller than average. Although Galton devised a way the same whether that additional year of school
of summarizing in a single figure the degree of represents an increase, for example, from six to
‘‘co-relation’’ between two variables, it was Galton’s seven years of school completed or from twelve to
associate Karl Pearson who developed the ‘‘coeffi- thirteen years. While one can analyze curvilinear
cient of correlation,’’ as it is now applied. Galton’s regression, a straight regression line greatly simpli-
original interest, the phenomenon of regression fies the analysis. Some (but not all) curvilinear
toward the mean, is no longer germane to contem- regressions can be made into straight-line regressions
porary correlation and regression analysis, but the by a relatively simple transformation of one of the
term ‘‘regression’’ has been retained with a modi- variables (e.g., taking a logarithm). The common
fied meaning. assumption that the regression line is a straight
line is known as the assumption of rectilinearity, or
Although Galton and Pearson originally fo- more commonly (even if less precisely) as the
cused their attention on bivariate (two variables) assumption of linearity.
correlation and regression, in current applications
more than two variables are typically incorporated The slope of the regression line reflects one
into the analysis to yield partial correlation coeffi- feature of the relationship between two variables.
cients, multiple regression analysis, and several If the regression line slopes ‘‘uphill,’’ as in Figure
related techniques that facilitate the informed in- 1, then Y increases as X increases, and the steeper
terpretation of the linkages between pairs of vari- the slope, the more Y increases for each unit
ables. This summary begins with two variables and increase in X. In contrast, if the regression line
then moves to the consideration of more than two slopes ‘‘downhill’’ as one moves from left to right,
variables. Y decreases as X increases, and the steeper the
slope, the more Y decreases for each unit increase
Consider a very large sample of cases, with a in X. If the regression line doesn’t slope at all but is
measure of some variable, X, and another variable, perfectly horizontal, then there is no relationship
Y, for each case. To make the illustration more between the variables. But the slope does not tell
concrete, consider a large number of adults and, how closely the two variables are ‘‘co-related’’ (i.e.,
for each, a measure of their education (years of how closely the values of Y cluster around the
school completed = X) and their income (dollars regression line).
earned over the past twelve months = Y). Subdi-
vide these adults by years of school completed, and A regression line may be represented by a
for each such subset compute a mean income for a simple mathematical formula for a straight line. Thus:
given level of education. Each such mean is called
a conditional mean and is represented by |X, that Y|X= ayx + byxX (1)
is, the mean of Y for a given value of X.
where |X = the mean Y for a given value of X, or
Imagine now an ordered arrangement of the the regression line values of Y given X; ayx = the Y
subsets from left to right according to the years of intercept (i.e., the predicted value of |X when X
school completed, with zero years of school on the = 0); and byx = the slope of the regression of Y on X
left, followed by one year of school, and so on (i.e., the amount by which |X increases or de-
through the maximum number of years of school creases—depending on whether b is positive or
completed in this set of cases, as shown in Figure 1. negative—for each one-unit increase in X).
447
CORRELATION AND REGRESSION ANALYSIS
Annual Income in Thousands of Dollars
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Education
(years of school completed)
*=the mean income for each education level
Figure 1. Hypothetical Regression of Income on Education
Equation 1 is commonly written in a slightly Equations 1 and 2 are theoretically equivalent.
different form: Equation 1 highlights the fact that the points on
the regression line are assumed to represent con-
Y = ayx + byxX (2) ditional means (i.e., the mean Y for a given X).
Equation 2 highlights the fact that points on the
where Ŷ = the regression prediction for Y for a regression line are not ordinarily found by com-
given value of X, and ayx and byx are as defined puting a series of conditional means, but are found
above, with Ŷ substituted for |X. by alternative computational procedures.
448
CORRELATION AND REGRESSION ANALYSIS
Typically the number of cases is insufficient to variance of the Y values around the regression
yield a stable estimate of each of a series of condi- predictions is defined as the mean of the squared
tional means, one for each level of X. Means based deviations between them. The variances around
on a relatively small number of cases are inaccu- each of the values along the regression line are
rate because of sampling variation, and a line assumed to be equal. This is known as the assump-
connecting such unstable conditional means may tion of homoscedasticity (homogeneous scatter or
not be straight even though the true regression variance). When the variances of the Y values
line is. Hence, one assumes that the regression line around the regression predictions are larger for
is a straight line unless there are compelling rea- some values of X than for others (i.e., when
sons for assuming otherwise; one can then use the homoscedasticity is not present), then X serves as a
X and Y values for all cases together to estimate the better predictor of Y in one part of its range than
Y intercept, ayx, and the slope, byx, of the regres- in another. The homoscedasticity assumption is
sion line that is best fit by the criterion of least usually at least approximately true.
squares. This criterion requires predicted values
for Y that will minimize the sum of squared devia- The variance around the regression line is a
tions between the predicted values and the ob- measure of the accuracy of the regression predic-
served values. Hence, a ‘‘least squares’’ regression tions. But it is not an easily interpreted measure of
line is the straight line that yields a lower sum of the degree of correlation because it has not been
squared deviations between the predicted (regres- ‘‘normed’’ to vary within a limited range. Two
sion line) values and the observed values than does other measures, closely related to each other, pro-
any other straight line. One can find the parame- vide such a normed measure. These measures,
ters of the ‘‘least squares’’ regression line for a which are always between zero and one in absolute
given set of X and Y values by computing value (i.e., sign disregarded) are: (a) the correla-
tion coefficient, r, which is the measure devised by
byx = –∑––(X–∑–––(–XX–)–––(XY––)–2––Y–) (3) Karl Pearson; and (b) the square of that coeffi-
cient, r2, which, unlike r, can be interpreted as a
ayx = Y – byxX (4) percentage.
These parameters (substituted in equation 2) Pearson’s correlation coefficient, r, can be
describe the straight regression line that best fits computed using the following formula:
by the criterion of least squares. By substituting
the X value for a given case into equation 2, one ∑ (X – X) (Y – Y)
can then find Ŷ for that case. Otherwise stated,
once ayx and byx have been computed, equation 2 N
will yield a precise predicted income level (Ŷ) for
each education level. ∑ (X – X)2 ∑ (Y – Y)2
NN
These predicted values may be relatively good ( )( )ryx = rxy = (5)
or relatively poor predictions, depending on wheth-
er the actual values of Y cluster closely around the The numerator in equation 5 is known as the
predicted values on the regression line or spread covariance of X and Y. The denominator is the
themselves widely around that line. The variance square root of the product of the variances of X
of the Y values (income levels in this illustration) and Y. Hence, equation 5 may be rewritten:
around the regression line will be relatively small if
the Y values cluster closely around the predicted Covariance (X, Y)
values (i.e., when the regression line provides rela- ryx = rxy = [Variance (X)] [Variance (Y)] (6)
tively good predictions). On the other hand, the
variance of Y values around the regression line will While equation 5 may serve as a computing
be relatively large if the Y values are spread widely guide, neither equation 5 nor equation 6 tells why
around the predicted values (i.e., when the regres- it describes the degree to which two variables
sion line provides relatively poor predictions). The covary. Such understanding may be enhanced by
stating that r is the slope of the least squares
regression line when both X and Y have been
449
CORRELATION AND REGRESSION ANALYSIS
transformed into ‘‘standard deviates’’ or ‘‘z meas- slope will be near 1.0, and closer to 0 when they
ures.’’ Each value in a distribution may be trans- scatter widely around the regression line.
formed into a ‘‘z measure’’ by finding its deviation
from the mean of the distribution and dividing by But while r has a precise meaning—it is the
the standard deviation (the square root of the slope of the regression line for standardized meas-
variance) of that distribution. Thus ures—that meaning is not intuitively understand-
able as a measure of the degree to which one
Zx = X–X (7) variable can be accurately predicted from the oth-
er. The square of the correlation coefficient, r2,
∑ (X – X)2 does have such an intuitive meaning. Briefly stat-
N ed, r2 indicates the percent of the possible reduc-
tion in prediction error (measured by the variance
When both the X and Y measures have been thus of actual values around predicted values) that is
standardized, ryx = rxy is the slope of the regression achieved by shifting from (a) as the prediction,
of Y on X, and of X on Y. For standard deviates, to (b) the regression line values as the prediction.
the Y intercept is necessarily 0, and the following Otherwise stated,
equation holds:
Zy = ryxZx (8) Variance of Y values Variance of Y values
r2 = around Y around Y
where Ẑy = the regression prediction for the ‘‘Z Variance of Y values
measure’’ of Y, given X; Zx = the standard deviates
of X; and ryx = rxy = the Pearsonian correlation Around Y (9)
between X and Y.
The denominator of Equation 9 is called the
Like the slope byx, for unstandardized meas- total variance of Y. It is the sum of two components:
ures, the slope for standardized measures, r, may (1) the variance of the Y values around Ŷ, and (2)
be positive or negative. But unlike byx, r is always the variance of the Ŷ around . Hence the numera-
between 0 and 1.0 in absolute value. The correla- tor of equation 9 is equal to the variance of the Ŷ
tion coefficient, r, will be 0 when the standardized values (regression values) around . Therefore
regression line is horizontal so that the two vari-
ables do not covary at all—and, incidentally, when Variance of Y values (10)
the regression toward the mean, which was Galton’s r2 = around Y
original interest, is complete. On the other hand, r
will be 1.0 or − 1.0 when all values of Zy fall Variance of Y values
precisely on the regression line rZx. This means around Y
that when r = + 1.0, for every case Zx = Zy—that is,
each case deviates from the mean on X by exactly = proportion of variance explained
as much and in the same direction as it deviates
from the mean on Y, when those deviations are Even though it has become common to refer to r2
measured in their respective standard deviation as the proportion of variance ‘‘explained,’’ such
units. And when r = − 1.0, the deviations from the terminology should be used with caution. There
mean measured in standard deviation units are are several possible reasons for two variables to be
exactly equal, but they are in opposite directions. correlated, and some of these reasons are inconsis-
(It is also true that when r = 1.0, there is no tent with the connotations ordinarily attached to
regression toward the mean, although this is very terms such as ‘‘explanation’’ or ‘‘explained.’’ One
rarely of any interest in contemporary applica- possible reason for the correlation between two
tions.) More commonly, r will be neither 0 nor 1.0 variables is that X influences Y. This is presumably
in absolute value but will fall between these ex- the reason for the positive correlation between
tremes, closer to 1.0 in absolute value when the Zy education and income; higher education facili-
values cluster closely around the regression line, tates earning a higher income, and it is appropri-
which, in this standardized form, implies that the ate to refer to a part of the variation in income as
being ‘‘explained’’ by variation in education. But
there is also the possibility that two variables are
correlated because both are measures of the same
450
CORRELATION AND REGRESSION ANALYSIS
dimension. For example, among twentieth-centu- Lacking such a large sample, one may hold age
ry nation-states, there is a high correlation be- constant by ‘‘statistical adjustment.’’
tween the energy consumption per capita and the
gross national product per capita. These two vari- To understand the underlying logic of partial
ables are presumably correlated because both are correlation, one considers the regression residuals
indicators of the degree of industrial develop- (i.e., for each case, the discrepancy between the
ment. Hence, one variable does not ‘‘explain’’ regression line value and the observed value of the
variation in the other, if ‘‘explain’’ has any of its predicted variable). For example, the regression
usual meanings. And two variables may be corre- residual of reading ability on age for a given case is
lated because both are influenced by a common the discrepancy between the actual reading ability
cause, in which case the two variables are ‘‘spuri- and the predicted reading ability based on age.
ously correlated.’’ For example, among elementa- Each residual will be either positive or negative
ry-school children, reading ability is positively cor- (depending on whether the observed reading abili-
related with shoe size. This correlation appears ty is higher or lower than the regression predic-
not because large feet facilitate learning, and not tion). Each residual will also have a specific value,
because both are measures of the same underlying indicating how much higher or lower than the age-
dimension, but because both are influenced by specific mean (i.e., regression line values) the read-
age. As they grow older, schoolchildren learn to ing ability is for each person. The complete set of
read better and their feet grow larger. Hence, shoe these regression residuals, each being a deviation
size and reading ability are ‘‘spuriously correlat- from the age-specific mean, describes the pattern
ed’’ because of the dependence of both on age. It of variation in reading abilities that would obtain if
would therefore be misleading to conclude from all of these schoolchildren were identical in age.
the correlation between shoe size and reading Similarly, the regression residuals for shoe size on
ability that part of the variation in reading ability is age describe the pattern of variation that would
‘‘explained’’ by variation in shoe size, or vice versa. obtain if all of these schoolchildren were identical
in age. Hence, the correlation between the two sets
In the attempt to discover the reasons for the of residuals—(1) the regression residuals of shoe
correlation between two variables, it is often useful size on age and (2) the regression residuals of
to include additional variables in the analysis. Sev- reading ability on age—is the correlation between
eral techniques are available for doing so. shoe size and reading ability, with age ‘‘held con-
stant.’’ In practice, it is not necessary to find each
PARTIAL CORRELATION regression residual to compute the partial correla-
tion, because shorter computational procedures
One may wish to explore the correlation between have been developed. Hence,
two variables with a third variable ‘‘held constant.’’
The partial correlation coefficient may be used for rxy•x = rxy – rxzryz (11)
this purpose. If the only reason for the correlation (1 – r2xz) (1 – r2yz)
between shoe size and reading ability is because
both are influenced by variation in age, then the where rxy⋅z; = the partial coefficient between X and
correlation should disappear when the influence Y, holding Z constant; rxy = the bivariate correla-
of variation in age is made nil—that is, when age is tion coefficient between X and Y; rxz = the bivari-
held constant. Given a sufficiently large number of ate correlation coefficient between X and Z; and
cases, age could be held constant by considering ryz = the bivariate correlation coefficient be-
each age grouping separately—that is, one could tween Y and Z.
examine the correlation between shoe size and
reading ability among children who are six years It should be evident from equation 11 that if Z
old, among children who are seven years old, eight is unrelated to both X and Y, controlling for Z will
years old, etc. (And one presumes that there would yield a partial correlation that does not differ from
be no correlation between reading ability and shoe the bivariate correlation. If all correlations are
size among children who are homogeneous in positive, each increase in the correlation between
age.) But such a procedure requires a relatively the control variable, Z, and each of the focal
large number of children in each age grouping.
451
CORRELATION AND REGRESSION ANALYSIS
variables, X and Y, will move the partial, rxy.z, X1 and X2 (assuming both influence Y) than with
closer to 0, and in some circumstances a positive either of these alone.
bivariate correlation may become negative after
controlling for a third variable. When rxy is posi- A regression equation including more than a
tive and the algebraic sign of ryz differs from the single predictor of Y is called a multiple regression
sign of rxz (so that their product is negative), the equation. For two predictors, the multiple regres-
partial will be larger than the bivariate correlation, sion equation is:
indicating that Z is a suppressor variable—that is, a
variable that diminishes the correlation between X Y = ay.12 + by1.2X1 + by2.1X2 (12)
and Y unless it is controlled. Further discussion of
partial correlation and its interpretation will be where Ŷ = the least squares prediction of Y based
found in Simon 1954; Mueller, Schuessler, and on X1 and X2; ay.12 = the Y intercept (i.e., the
Costner 1977; and Blalock 1979. predicted value of Y when both X1 and X2 are 0);
by1.2 = the (unstandardized) regression slope of Y
Any correlation between two sets of regres- on X1, holding X2 constant; and by2.1 = the
sion residuals is called a partial correlation coeffi- (unstandardized) regression slope of Y on X2,
cient. The illustration immediately above is called a holding X1 constant. In multiple regression analy-
first-order partial, meaning that one and only one sis, the predicted variable (Y in equation 12) is
variable has been held constant. A second-order commonly known as the criterion variable, and the
partial means that two variables have been held X’s are called predictors. As in a bivariate regres-
constant. More generally, an nth-order partial is one sion equation (equation 2), one assumes both
in which precisely n variables have been ‘‘con- rectilinearity and homoscedasticity, and one finds
trolled’’ or held constant by statistical adjustment. the Y intercept (ay.12 in equation 12) and the
regression slopes (one for each predictor; they are
When only one of the variables being correlat- by1.2 and by2.1 in equation 12) that best fit by the
ed is a regression residual (e.g., X is correlated criterion of least squares. The b’s or regression
with the residuals of Y on Z), the correlation is slopes are partial regression coefficients. The correla-
called a part correlation. Although part correlations tion between the resulting regression predictions
are rarely used, they are appropriate when it seems (Ŷ) and the observed values of Y is called the
implausible to residualize one variable. Generally, multiple correlation coefficient, symbolized by R.
part correlations are smaller in absolute value than
the corresponding partial correlation. In contemporary applications of multiple re-
gression, the partial regression coefficients are
MULTIPLE REGRESSION typically the primary focus of attention. These
coefficients describe the regression of the criteri-
Earned income level is influenced not simply by on variable on each predictor, holding constant all
one’s education but also by work experience, skills other predictors in the equation. The b’s in equa-
developed outside of school and work, the prevail- tion 12 are unstandardized coefficients. The analo-
ing compensation for the occupation or profes- gous multiple regression equation for all variables
sion in which one works, the nature of the regional expressed in standardized form is
economy where one is employed, and numerous
other factors. Hence it should not be surprising Z y= b*y1.2 Z1 + b*y2.1Z2 (13)
that education alone does not predict income with
high accuracy. The deviations between actual in- where Ẑ = the regression prediction for the ‘‘z
come and income predicted on the basis of educa-
tion are presumably due to the influence of all the measure’’ of Y, given X1 and X2 Z1 = the standard
other factors that have an effect, great or small, on
one’s income level. By including some of these deviate of X1 Z2= the standard deviate of X2 b*y1.2
other variables as additional predictors, the accu-
racy of prediction should be increased. Otherwise = the standardized slope of Y on X1, holding X2
stated, one expects to predict Y better using both
constant; and b* = the standardized slope of Y
y2.1
on X2, holding X1 constant.
The standardized regression coefficients in an
equation with two predictors may be calculated
from the bivariate correlations as follows:
452
CORRELATION AND REGRESSION ANALYSIS
b*y1.2 = ry1 – ry2ry12 (14) residuals themselves having been expressed in the
1 – r212 form of standard deviates.
b*y2.1 = ry2 – ry1r12 (15) A standardized partial regression coefficient
1 – r212 can be transformed into an unstandardized partial
regression coefficient by
where b* = the standardized partial regression by1.2 = b*y1.2 sy (17)
y1.2 s1
coefficient of Y on X1, controlling for X2; and
b* = the standardized partial regression coeffi- sy
y2.1 s2
cient of Y on X2, controlling for X1. by2.1 = b*y2.1 (18)
Standardized partial regression coefficients, where by1.2 = the unstandardized partial regres-
here symbolized by b* (read ‘‘b star’’), are fre- sion coefficient of Y on X1, controlling for X2; by2.1
quently symbolized by the Greek letter beta, and = the unstandardized partial regression coefficient
they are commonly referred to as ‘‘betas,’’ ‘‘beta of Y on X2, controlling for X1; b*y2.1 and b*y2.1 are
coefficients,’’ or ‘‘beta weights.’’ While this is com- standardized partial regression coefficients, as de-
mon usage, it violates the established practice of fined above; sy = the standard deviation of Y; s1 =
using Greek letters to refer to population parame- the standard deviation of X1; and s2 = the standard
ters instead of sample statistics. deviation of X2.
A comparison of equation 14, describing the Under all but exceptional circumstances, stand-
standardized partial regression coefficient, b*y1.2, ardized partial regression coefficients fall between
with equation 11, describing the partial correla- −1.0 and +1.0. The relative magnitude of the stand-
tion coefficient, ry1.2, will make it evident that ardized coefficients in a given regression equation
these two coefficients are closely related. They indicates the relative magnitude of the relation-
have identical numerators but different denomi- ship between the criterion variable and the predic-
nators. The similarity can be succinctly expressed by tor in question, after holding constant all the other
predictors in that regression equation. Hence, the
r2y1.2 = b*y1.2 b*1y.2 (16) standardized partial regression coefficients in a
given equation can be compared to infer which
If any one of the quantities in equation 16 is 0, all predictor has the strongest relationship to the
are 0, and if the partial correlation is 1.0 in abso- criterion, after holding all other variables in that
lute value, both of the standardized partial regres- equation constant. The comparison of unstandardized
sion coefficients in equation 16 must also be 1.0 in partial regression coefficients for different predic-
absolute value. For absolute values between 0 and tors in the same equation does not ordinarily yield
1.0, the partial correlation coefficient and the useful information because these coefficients are
standardized partial regression coefficient will have affected by the units of measure. On the other
somewhat different values, although the general hand, it is frequently useful to compare unstandardized
interpretation of two corresponding coefficients is partial regression coefficients across equations.
the same in the sense that both coefficients repre- For example, in separate regression equations pre-
sent the relationship between two variables, with dicting income from education and work experi-
one or more other variables held constant. The ence for the United States and Great Britain, if the
difference between them is rather subtle and rare- unstandardized regression coefficient for educa-
ly of major substantive import. Briefly stated, the tion in the equation for Great Britain is greater
partial correlation coefficient—e.g., ry1.2—is the than the unstandardized regression coefficient for
regression of one standardized residual on anoth- education in the equation for the United States,
er standardized residual. The corresponding stand- the implication is that education has a greater
ardized partial regression coefficient, b*y1.2, is the influence on income in Great Britain than in the
regression of one residual on another, but the United States. It would be hazardous to draw any
residuals are standard measure discrepancies from such conclusion from the comparison of standard-
standard measure predictions, rather than the ized coefficients for Great Britain and the United
453
CORRELATION AND REGRESSION ANALYSIS
States because such coefficients are affected by the inclusion of dichotomies as predictors in multiple
variances in the two populations. regression analysis, and then to the creation of a
set of dichotomies from a categorical variable with
The multiple correlation coefficient, R, is de- more than two subdivisions—that is, to dummy
fined as the correlation between the observed variable analysis (Cohen 1968; Bohrnstedt and
values of Y and the values of Y predicted by the Knoke 1988; Hardy 1993).
multiple regression equation. It would be unneces-
sarily tedious to calculate the multiple correlation Religious denomination—e.g., Protestant, Catho-
coefficient in that way. The more convenient com- lic, and Jewish—serves as an illustration. From
putational procedure is to compute R2 (for two these three categories, one forms two dichotomies,
predictors, and analogously for more than two called ‘‘dummy variables.’’ In the first of these, for
predictors) by the following: example, cases are classified as ‘‘1’’ if they are
Catholic, and ‘‘0’’ otherwise (i.e., if Protestant or
R2 = b*y1.2ry1 + b*y2.1ry2 (19) Jewish). In the second of the dichotomies, cases
are classified as ‘‘1’’ if they are Jewish, and ‘‘0’’
Like r2, R2 varies from 0 to 1.0 and indicates the otherwise (i.e., if Protestant or Catholic). In this
proportion of variance in the criterion that is illustration, Protestant is the ‘‘omitted’’ or ‘‘refer-
‘‘explained’’ by the predictors. Alternatively stat- ence’’ category (but Protestants can be identified
ed, R2 is the percent of the possible reduction in as those who are classified ‘‘0’’ on both of the other
prediction error (measured by the variance of dichotomies). The resulting two dichotomized
actual values around predicted values) that is ‘‘dummy variables’’ can serve as the only predic-
achieved by shifting from (a) as the prediction to tors in a multiple regression equation, or they may
(b) the multiple regression values, Ŷ, as the prediction. be combined with other predictors. When the
dummy variables mentioned are the only predic-
VARIETIES OF MULTIPLE REGRESSION tors, the unstandardized regression coefficient for
the predictor in which Catholics are classified ‘‘1’’
The basic concept of multiple regression has been is the difference between the mean Y for Catholics
adapted to a variety of purposes other than those and Protestants (the ‘‘omitted’’ or ‘‘reference’’
for which the technique was originally developed. category). Similarly, the unstandardized regres-
The following paragraphs provide a brief summa- sion coefficient for the predictor in which Jews are
ry of some of these adaptations. classified ‘‘1’’ is the difference between the mean Y
for Jews and Protestants. When the dummy vari-
Dummy Variable Analysis. As originally con- ables are included with other predictors, the
ceived, the correlation coefficient was designed to unstandardized regression coefficients are the same
describe the relationship between continuous, nor- except that the difference of each mean from the
mally distributed variables. Dichotomized predic- mean of the ‘‘reference’’ category has been statisti-
tors such as gender (male and female) were intro- cally adjusted to control for each of the other
duced early in bivariate regression and correlation, predictors in the regression equation.
which led to the ‘‘point biserial correlation coeffi-
cient’’ (Walker and Lev 1953). For example, if one The development of ‘‘dummy variable analy-
wishes to examine the correlation between gender sis’’ allowed multiple regression analysis to be
and income, one may assign a ‘‘0’’ to each instance linked to the experimental statistics developed by
of male and a ‘‘1’’ to each instance of female to R. A. Fisher, including the analysis of variance and
have numbers representing the two categories of covariance. (See Cohen 1968.)
the dichotomy. The unstandardized regression
coefficient, computed as specified above in equa- Logistic Regression. Early students of corre-
tion 3, is then the difference between the mean lation anticipated the need for a measure of corre-
income for the two categories of the dichotomous lation when the predicted or dependent variable
predictor, and the computational formula for r was dichotomous. Out of this came (a) the phi
(equation 5), will yield the point biserial correla- coefficient, which can be computed by applying
tion coefficient, which can be interpreted much the computational formula for r (equation 5) to
like any other r. It was then only a small step to the two dichotomies, each coded ‘‘0’’ or ‘‘1,’’ and (b)
the tetrachoric correlation coefficient, which uses
454
CORRELATION AND REGRESSION ANALYSIS
information in the form of two dichotomies to example, the Z variables are appropriate predic-
estimate the Pearsonian correlation for the corre- tors in the equation predicting X1 because they are
sponding continuous variables, assuming the assumed causes of X1. But X2 is not an appropriate
dichotomies result from dividing two continuous predictor of X1 because it is assumed to be a
and normally distributed variables by arbitrary spurious correlate of X1 (i.e., X1 and X2 are pre-
cutting points (Kelley 1947; Walker and Lev 1953; sumed to be correlated only because they are both
Carroll 1961). influenced by the Z variables, not because one
influences the other). And Y is not an appropriate
These early developments readily suggested predictor of X1 because Y is assumed to be an
use of a dichotomous predicted variable, coded effect of X1, not one of its causes. When the
‘‘0’’ or ‘‘1,’’ as the predicted variable in a multiple assumptions about the causal structure linking a
regression analysis. The predicted value is then the set of variables have been made explicit, the appro-
conditional proportion, which is the conditional priate predictors for each variable have been iden-
mean for a dichotomized predicted variable. But tified from this assumed causal structure, and the
this was not completely satisfactory in some cir- resulting equations have been estimated by the
cumstances because the regression predictions are, techniques of regression analysis, the result is a
under some conditions, proportions greater than path analysis, and each of the resulting coefficients
1 or less than 0. Logistic regression (Retherford is said to be a ‘‘path coefficient’’ (if expressed in
1993; Kleinman 1994; Menard 1995) is responsive standardized form) or a ‘‘path regression coeffi-
to this problem. After coding the predicted vari- cient’’ (if expressed in unstandardized form).
able ‘‘0’’ or ‘‘1,’’ the predicted variable is trans-
formed to a logistic—that is, the logarithm of the If the assumed causal structure is correct, a
‘‘odds,’’ which is to say the logarithm of the ratio of path analysis allows one to ‘‘decompose’’ a correla-
the number of 1’s to the number of 0’s. With tion between two variables into ‘‘direct effects’’;
the logistic as the predicted variable, impossible ‘‘indirect effects’’; and, potentially, a ‘‘spurious
regression predictions do not result, but the component’’ as well (Land 1969; Bohrnstedt and
unstandardized logistic regression coefficients, de- Knoke 1988; McClendon 1994).
scribing changes in the logarithm of the ‘‘odds,’’
lack the intuitive meaning of ordinary regression For example, we may consider the correlation
coefficients. An additional computation is required between the occupational achievement of a set of
to be able to describe the change in the predicted fathers and the occupational achievement of their
proportion for a given one-unit change in a predic- sons. Some of this correlation may occur because
tor, with all other predictors in the equation held the father’s occupational achievement influences
constant. the educational attainment of the son, and the
son’s educational attainment, in turn, influences
Path Analysis. The interpretation of multiple his occupational achievement. This is an ‘‘indirect
regression coefficients can be difficult or impossi- effect’’ of the father’s occupational achievement
ble when the predictors include an undifferenti- on the son’s occupational achievement ‘‘through’’
ated set of causes, consequences, or spurious cor- (or ‘‘mediated by’’) the son’s education. A ‘‘direct
relates of the predicted variable. Path analysis was effect,’’ on the other hand, is an effect that is not
developed by Sewell Wright (1934) to facilitate the mediated by any variable included in the analysis.
interpretation of multiple regression coefficients Such mediating variables could probably be found,
by making explicit assumptions about causal struc- but if they have not been identified and included
ture and including as predictors of a given variable in this particular analysis, then the effects mediat-
only those variables that precede that given vari- ed through them are grouped together as the
able in the assumed causal structure. For example, ‘‘direct effect’’—that is, an effect not mediated by
if one assumes that Y is influenced by X1 and X2, variables included in the analysis. If the father’s
and X1 and X2 are, in turn, both influenced by Z1, occupational achievement and the son’s occupa-
Z2, and Z3, this specifies the assumed causal struc- tional achievement are also correlated, in part,
ture. One may then proceed to write multiple because both are influenced by a common cause
regression equations to predict X1, X2, and Y, (e.g., a common hereditary variable), then that
including in each equation only those variables
that come prior in the assumed causal order. For
455
CORRELATION AND REGRESSION ANALYSIS
part of the correlation that is attributable to that Y = ay.12 + by1.2X1 + by2.1X2 + by.12X1X2 (20)
common cause constitutes the ‘‘spurious compo-
nent’’ of the correlation. If the variables responsi- In this equation, Y is said to be predicted, not
ble for the ‘‘spurious component’’ of a correlation simply by an additive combination of X1 and X2
have been included in the path analysis, the ‘‘spuri- but also by their product, X1X2. Although it may
ous component’’ can be estimated; otherwise such not be intuitively evident from the equation itself,
a ‘‘spurious component’’ is merged into the ‘‘di- the presence of a multiplicative effect (i.e., the
rect effect,’’ which, despite the connotations of the regression coefficient for the multiplicative term,
name, absorbs all omitted indirect effects and all by.12, is not 0) implies that the effect of X1 on Y
omitted spurious components. depends on the level of X2, and vice versa. This is
commonly called an interaction effect (Allison 1977;
‘‘Stepwise’’ Regression Analysis. The inter- Blalock 1979; Jaccard, Turrisi and Wan 1990; Ai-
pretation of regression results can sometimes be ken and West 1991; McClendon 1994). The inclu-
facilitated without specifying completely the pre- sion of multiplicative terms in a regression equa-
sumed causal structure among a set of predictors. tion is especially appropriate when there are sound
If the purpose of the analysis is to enhance under- reasons for assuming that the effect of one vari-
standing of the variation in a single dependent able differs for different levels of another variable.
variable, and if the various predictors presumed to For example, if one assumes that the ‘‘return to
contribute to that variation can be grouped, for education’’ (i.e., the annual income added by each
example, into proximate causes and distant causes, a additional year of schooling) will be greater for
stepwise regression analysis may be useful. Depend- men than for women, this assumption can be
ing on one’s primary interest, one may proceed in explored by including all three predictors: educa-
two different ways. For example, one may begin by tion, gender, and the product of gender and
regressing the criterion variable on the distant education.
causes, and then, in a second step, introduce the
proximate causes into the regression equation. When product terms have been included in a
Comparison of the coefficients at each step will regression equation, the interpretation of the re-
reveal the degree to which the effects of the distant sulting partial regression coefficients may become
causes are mediated by the proximate causes in- complex. For example, unless all predictors are
cluded in the analysis. Alternatively, one may be- ‘‘ratio variables’’ (i.e., variables measured in uni-
gin by regressing the criterion variable on the form units from an absolute 0), the inclusion of a
proximate causes, and then introduce the distant product term in a regression equation renders the
causes into the regression equation in a second coefficients for the additive terms uninterpretable
step. Comparing the coefficients at each step, one (see Allison 1977).
can infer the degree to which the first-step regres-
sion of the criterion variable on the proximate SAMPLING VARIATION AND TESTS
causes is spurious because of the dependence of AGAINST THE NULL HYPOTHESIS
both on the distant causes. A stepwise regression
analysis may proceed with more than two stages if Descriptions based on incomplete information
one wishes to distinguish more than two sets of will be inaccurate because of ‘‘sampling varia-
predictors. One may think of a stepwise regression tion.’’ Otherwise stated, different samplings of
analysis of this kind as analogous to a path analysis information will yield different results. This is true
but without a complete specification of the causal of sample regression and correlation coefficients,
structure. as it is for other descriptors. Assuming a random
selection of observed information, the ‘‘shape’’ of
Nonadditive Effects in Multiple Regression. the distribution of such sampling variation is often
In the illustrative regression equations preceding known by mathematical reasoning, and the magni-
this section, each predictor has appeared only tude of such variation can be estimated. For exam-
once, and never in a ‘‘multiplicative’’ term. We ple, if the true correlation between X and Y is 0, a
now consider the following regression equation, series of randomly selected observations will rare-
which includes such a multiplicative term: ly yield a correlation that is precisely 0. Instead, the
observed correlation will fluctuate around 0 in the
456
CORRELATION AND REGRESSION ANALYSIS
‘‘shape’’ of a normal distribution, and the stand- assumptions are met only approximately (Bohrnstedt
ard deviation of that normal sampling distribu- and Carter 1971). Even so, restricted or biased
tion—called the standard error of r—will be samples may lead to conclusions that are mislead-
ing if they are inappropriately generalized. Fur-
σr = 1 (21) thermore, regression results may be misinterpret-
N–1 ed if interpretation rests on an implicit causal
model that is misspecified. For this reason it is
where σr = the standard error of r (i.e. the standard advisable to make the causal model explicit, as in
deviation of the sampling distribution of r, given path analysis or structural equation modeling and
that the true correlation is 0); and N = the sample to use regression equations that are appropriate
size (i.e., the number of randomly selected cases for the model as specified. ‘‘Outliers’’ and ‘‘devi-
used in the calculation of r). For example, if the ant cases’’ (i.e., cases extremely divergent from
true correlation were 0, a correlation coefficient most) may have an excessive impact on regression
based on a random selection of 400 cases will have coefficients, and hence may lead to erroneous
a standard error of approximately 1/20 or .05. An conclusions. (See Berry and Feldman 1985; Fox
observed correlation of .15 or greater in absolute 1991; Hamilton 1992.) A ubiquitous but still not
value would thus be at least three standard errors widely recognized source of misleading results in
away from 0 and hence very unlikely to have regression analysis is measurement error (both
appeared simply because of random fluctuations random and non-random) in the variables (Stouffer
around a true value of 0. This kind of conclusion is 1936; Kahneman 1965; Gordon 1968; Bohrnstedt
commonly expressed by saying that the observed and Carter 1971; Fuller and Hidiroglou 1978;
correlation is ‘‘significantly different from 0’’ at a Berry and Feldman 1985). In bivariate correlation
given level of significance (in this instance, the and regression, the effect of measurement error
level of significance cited could appropriately be can be readily anticipated: on the average, random
.01). Or the same conclusion may be more simply measurement error in the predicted variable attenu-
(but less precisely) stated by saying that the ob- ates (moves toward zero) the correlation coeffi-
served correlation is ‘‘significant.’’ cient (i.e., the standardized regression coefficient)
but not the unstandardized regression coefficient,
The standard error for an unstandardized while random measurement error in the predictor
bivariate regression coefficient, and for an variable will, on the average, attenuate both the
unstandardized partial regression coefficient, may standardized and the unstandardized coefficients.
also be estimated (Cohen and Cohen 1983; Kleinman, In multiple regression analysis, the effect of ran-
Kupper and Muller 1988; Hamilton 1992; McClendon dom measurement error is more complex. The
1994; Fox 1997). Other things being equal, the unstandardized partial regression coefficient for a
standard error for the regression of the criterion given predictor will be biased, not simply by ran-
on a given predictor will decrease as (1) the num- dom measurement error in that predictor, but also
ber of observations (N) increases; (2) the variance by other features. When random measurement
of observed values around predicted values de- error is entailed in a given predictor, X, that
creases; (3) the variance of the predictor increases; predictor is not completely controlled in a regres-
and (4) the correlation between the predictor and sion analysis. Consequently, the unstandardized
other predictors in the regression equation decreases. partial regression coefficient for every other pre-
dictor that is correlated with X will be biased by
PROBLEMS IN REGRESSION ANALYSIS random measurement error in X. Measurement
error may be non-random as well as random, and
Multiple regression is a special case of a very anticipating the effect of non-random measure-
general and adaptable model of data analysis known ment error on regression results is even more
as the general linear model (Cohen 1968; Fennessey challenging than anticipating the effect of random
1968; Blalock 1979). Although the assumptions error. Non-random measurement errors may be
underlying multiple regression seem relatively de- correlated errors (i.e., errors that are correlated
manding (see Berry 1993), the technique is re- with other variables in the system being analyzed),
markably ‘‘robust,’’ which is to say that the tech- and therefore they have the potential to distort
nique yields valid conclusions even when the greatly the estimates of both standardized and
457
CORRELATION AND REGRESSION ANALYSIS
unstandardized partial regression coefficients. Fennessey, James 1968 ‘‘The General Linear Model: A
Thus, if the measures used in regression analysis New Perspective on Some Familiar Topics.’’ Ameri-
are relatively crude, lack high reliability, or include can Journal of Sociology 74:1–27.
distortions (errors) likely to be correlated with
other variables in the regression equation (or with Fox, John 1991 Regression Diagnostics. Newbury Park,
measurement errors in those variables), regres- Calif.: Sage.
sion analysis may yield misleading results. In these
circumstances, the prudent investigator should ———, 1997 Applied Regression Analysis, Linear Models
interpret the results with considerable caution, or, and Related Methods. Beverly Hills, Calif.: Sage.
preferably, shift from regression analysis to the
analysis of structural equation models with multi- Fuller, W. A., and M. A. Hidiroglou 1978 ‘‘Regression
ple indicators. (See Herting 1985; Schumacker Estimates after Correcting for Attenuation.’’ Journal
and Lomax 1996.) This alternative mode of analy- of the American Statistical Association 73:99–104.
sis is well suited to correct for random measure-
ment error and to help locate and correct for non- Gordon, Robert 1968 ‘‘Issues in Multiple Regression.’’
random measurement error. American Journal of Sociology 73:592–616.
REFERENCES Hamilton, Lawrence C. 1992 Regression with Graphics: A
Second Course in Applied Statistics. Pacific Grove, Calif.:
Agresti, Alan and Barbara Finlay 1997 Statistical Methods Brookes-Cole.
for the Social Sciences, 3rd ed. Upper Saddle River,
N.J.: Prentice Hall. Hardy, Melissa A. 1993 Regression with Dummy Variables.
Newbury Park, Calif.: Sage.
Aiken, Leona S., and Stephen G. West 1991 Multiple
Regression: Testing and Interpreting Interactions. Newbury Heise, David R. 1969 ‘‘Problems in Path Analysis and
Park, Calif.: Sage. Causal Inference.’’ In Edgar Borgatta and George W.
Bohrnstedt, eds., Sociological Methodology 1969. San
Allen, Michael Patrick 1997 Understanding Regression Francisco: Jossey-Bass.
Analysis. New York: Plenum Press.
Herting, Gerald 1985 ‘‘Multiple Indicator Models Using
Allison, Paul 1977 ‘‘Testing for Interaction in Multiple LISREL.’’ In H. M. Blalock, Jr., ed., Causal Models in
Regression.’’ American Journal of Sociology 83:144–153. the Social Sciences. New York: Aldine.
Berry, William D. 1993 Understanding Regression Assump- Jaccard, James, Robert Turrisi, and Choi K. Wan 1990
tions. Newbury Park, Calif.: Sage. Interaction Effects in Multiple Regression. Newbury Park,
Calif.: Sage.
———, and Stanley Feldman 1985 Multiple Regression in
Practice. Beverly Hills, Calif.: Sage. Kahneman, D. 1965 ‘‘Control of Spurious Association
and the Reliability of the Controlled Variable.’’ Psy-
Blalock, Hubert M., Jr. 1979 Social Statistics, 2nd ed. New chological Bulletin 64:326–329.
York: McGraw-Hill.
Kelley, Truman Lee 1947 Fundamentals of Statistics. Cam-
Bohrnstedt, George W., and T. Michael Carter 1971 bridge, Mass.: Harvard University Press.
‘‘Robustness in Regression Analysis.’’ In Herbert L.
Costner, ed., Sociological Methodology 1971. San Fran- Kleinman, David G. 1994 Logistic Regression: A Self Learn-
cisco: Jossey-Bass. ing Text. New York: Springer.
———, and David Knoke 1988 Statistics for Social Data ———, Lawrence L. Kupper, and Keith E. Muller 1988
Analysis, 2nd ed. Itasca, Ill.: F. E. Peacock. Applied Regression Analysis and Other Multivariable
Methods. Boston: PWS-Kent.
Carroll, John B. 1961 ‘‘The Nature of the Data, or How
to Choose a Correlation Coefficient.’’ Psychometrica Land, Kenneth C. 1969 ‘‘Principles of Path Analysis.’’ In
26:347–372. Edgar Borgatta and George W. Bohrnstedt, eds.,
Sociological Methodology 1969. San Francisco: Jossey-Bass.
Cohen, Jacob 1968 ‘‘Multiple Regression as a General
Data Analytic System.’’ Psychological Bulletin 70:426–443. McClendon, McKee J. 1994 Multiple Regression and Caus-
al Analysis. Itasca, Ill.: F. E. Peacock.
———, and Patricia Cohen 1983 Applied Multiple Regres-
sion/Correlation Analysis for the Behavioral Sciences, Menard, Scott 1995 Applied Logistic Regression Analysis.
2nd ed. Hillsdale, N.J.: Lawrence Erlbaum Associates. Beverly Hills, Calif.: Sage.
Mueller, John H., Karl F. Schuessler, and Herbert L.
Costner 1977 Statistical Reasoning in Sociology, 3rd ed.
Boston: Houghton Mifflin.
Retherford, Robert D. and Minju Kim Choe 1993 Statis-
tical Models for Causal Analysis. New York: Wiley.
458
COUNTERCULTURES
Schroeder, Larry P., David L. Sjoquist and Paula E. of the remnants of former subcultures (Irwin 1970),
Stephan 1986 Understanding Regression Analysis: An trends suggest that they resist total assimilation
Introductory Guide. Beverly Hills, Calif.: Sage. and retain their cultural diversity and distinct
identity.
Schumacker, Randall E., and Richard G. Lomax 1996 A
Beginner’s Guide to Structural Equation Modeling. Some subcultures diverge from the dominant
Mahwah, N.J.: Lawrence Erlbaum Associates. culture without morally rejecting the norms and
values with which they differ. Others are more
Stouffer, Samuel A. 1936 ‘‘Evaluating the Effects of adamant in their condemnation, clearly conflict-
Inadequately Measured Variables in Partial Correla- ing with or opposing features of the larger society.
tion Analysis.’’ Journal of the American Statistical Asso- Milton Yinger first proposed, in 1960, to call these
ciation 31:348–360. contracultures, envisioning them as a subset of sub-
cultures, specifically, those having an element of
Walker, Helen M. 1929 Studies in the History of Statistical conflict with dominant norms, values, or both
Method. Baltimore: Williams and Wilkins. (Yinger 1960). Indeed, the feature he identified as
most compelling about a contraculture is its specif-
———, and Joseph Lev 1953 Statistical Inference. New ic organization in opposition to some cultural
York: Holt. belief(s) or expression(s). Contracultures often
arise, he noted, where there are conflicts of stan-
Weisberg, Sanford 1985 Applied Linear Regression. New dards or values between subcultural groups and
York: Wiley. the larger society. Factors strengthening the con-
flict then strengthen the contracultural response.
Wright, Sewell 1934 ‘‘The Method of Path Coefficients.’’ Contraculture members, especially from such
The Annals of Mathematical Statistics 5:161–215. groups as delinquent gangs, may be driven by their
experiences of frustration, deprivation, or discrimi-
HERBERT L. COSTNER nation within society.
COUNTERCULTURES Yinger’s conceptualization, although abstract
and academic at first, came to enjoy widespread
The enclaves in which people of the modern era popularity with the advent of the 1960s and the
live no longer resemble the small, integrated, and student movement. Here was the kind of
homogeneous communities of earlier times; rath- contraculture he had forecast, and his ideas were
er, these have been replaced by large societies that widely applied to the trends of the time, albeit
are complex and diverse in their composition. The under another label. Most analysts of contracultures
United States, a prime exemplar, is composed of preferred the term counterculture, and this soon
multiple smaller groups holding characteristics, overtook its predecessor as the predominant ex-
beliefs, customs, and interests that vary from the pression. In 1969 historian Theodore Roszak pub-
rest of society. While there are many cultural lished his The Making of a Counter Culture, claiming
universals binding such groups to the mainstream, that a large group of young people (ages fifteen to
they also exhibit significant cultural diversity. Some thirty) had arisen who adamantly rejected the
of these groups display no clear boundaries de- technological and scientific outlook characteristic
marcating them from the rest of society and fail to of Western industrialized culture, replacing this,
achieve any degree of permanence. Yet those that instead, with a humanistic/mysticist alternative. In
do, and that also share a distinctive set of norms, a more recent update, Roszak (1995) reflected
values, and behavior setting them off from the back on that time, further locating the countercul-
dominant culture, are considered subcultures. Sub- ture phenomenon as an historical aberration that
cultures can be organized around age, ethnicity, arose out of the affluence of post-World War II
occupation, social class, religion, or lifestyle and America. Kenneth Keniston (1971) described this
usually contain specific knowledge, expressions, counterculture as composed of distinct subgroups
ways of dressing, and systems of stratification that (radicals, dropouts, hippies, drug users, communards,
serve and guide its members (Thornton 1997). or those living in communes) rising from the most
Distinctive subcultures within the United States privileged children of the world’s wealthiest na-
include jazz musicians, gangs, Chicanos, gay, col- tion. Jack Douglas (1970) also discussed the social,
lege athletes, and drug dealers. While it was once
hypothesized that these subcultures would merge
together in a ‘‘melting pot,’’ incorporating a mix
459
COUNTERCULTURES
political, and economic background to this move- American Sociological Association (Yinger 1977)
ment and its roots in members’ entrenchment in and a book that serves as the definitive statement
the welfare state and the existing youth and stu- on the topic (1982). He asserted the fundamental
dent cultures. While this movement was clearly import of studying these sharp contradictions to
political as well, Douglas outlined some of its social the dominant norms and values of a society as
dimensions, including rejection of the workaday a means of gaining insight into social order.
world and its idealization of leisure, feeling, open- Countercultures, through their oppositional cul-
ness, and antimaterialism. Richard Flacks (1971) ture (polarity, reversal, inversion, and diametric
and Fred Davis (1971) followed with descriptions opposition), attempt to reorganize drastically the
of the counterculture’s overarching lifestyles, val- normative bases of social order. These alternatives
ues, political beliefs, and ideologies. Ralph Turner may range from rejecting a norm or value entirely
(1976), Nathan Adler (1972), and Erik Erikson to exaggerating its emphasis in their construction
(1968) explored the social psychological implica- of countervalues. As a result, some countercultures
tions of this counterculture, positing, respectively, fade rapidly while others become incorporated
a transformation in the self from ‘‘institution’’ to into the broader cultural value system. Examples
‘‘impulse,’’ the rise of an antinomian personality, of countercultural groups would include the 1960s
where individuals oppose the obligatoriness of the student counterculture (in both its political and
moral law, and the formation of the negative identi- social dimensions); youth gangs (especially delin-
ty. John Rothchild and Susan Berns Wolf (1976) quent groups); motorcycle gangs (such as the Hell’s
documented the vast extension of countercultural Angels); revolutionary groups (the Weathermen
outposts around the country and their innovations of the Students for a Democratic Society, the Black
in child rearing. Charles Reich (1970) emphatical- Panthers, millenialists [Williams 1996]); terrorist
ly stated that this counterculture, consisting most- organizations (such as the Symbionese Liberation
ly of students, was being reinforced by merging Army); extremist racist groups (the Ku Klux Klan,
with nonstudent youth, educated labor, and the skinheads [Baron 1997; Hamm 1995; Young and
women’s movement, already effecting a major Craig 1997] , the Aryan Nation); survivalists (Branch
transformation in Western laws, institutions, and Davidians); punkers; bohemian Beats; ‘‘straight
social structure. There was strong belief that this edgers;’’ Rainbow family; Earth First! (Lange 1990;
movement would significantly and permanently Short 1991) and some extreme religious sects
alter both society and its consciousness (Wuthnow (such as the Amish and the Hare Krishnas [see
1976). After researching one commune in depth, Saliba 1996]).
Berger (1981) later mused about the survival of the
counterculture, acknowledging its failure to meet VARIETIES OF COUNTERCULTURES
earlier expectations, yet examining how its ideals
and values become incorporated into the main- Yinger believed that countercultural groups could
stream culture (cf. Spates 1976). take several forms. The radical activist counter-
culture ‘‘preaches, creates, or demands new obli-
Other subcultural analysts noted more broad- gations’’ (Yinger 1977, p. 838). They are intimate-
ly that these groups are typically popular among ly involved with the larger culture in their attempts to
youth, who have the least investment in the exist- transform it. Members of the communitarian utopi-
ing culture, and that, lacking power within society, an counterculture live as ascetics, withdrawing
they are likely to feel the forces of social control into an isolated community forged under the guide-
swiftly moving against them, from the mass media lines of their new values. Mystical countercultures
to police action. Countercultures were further search for the truth and for themselves, turning
differentiated from subcultures by the fact that inward toward consciousness to realize their val-
their particular norms and values, were not well ues. Theirs is more a disregard of society than an
integrated into the dominant culture, generally effort to change it. These three forms are not
known among group members, and other main- necessarily intended to describe particular groups.
stream subcultures. Rather, they are ideal types, offered to shed in-
sights into characteristics or tendencies groups
Yinger reclaimed theoretical command of the may combine or approximate in their formation.
counterculture concept with his reflective expan- Hippie communities or bohemian Beat groups
sions on the term in a presidential address for the
460
COUNTERCULTURES
combined the mystical and utopian features of religious movements. Other countercultures arise
countercultures in their withdrawal from conven- out of underlying or developing societal stress:
tional society and their search for a higher tran- rapid political or economic change; demographic
scendence. Revolutionary youth gangs, such as the transformations in the population (age, gender,
1960s radicals, the Hell’s Angels, and the punkers, location); a swift influx of new ideas; drastic escala-
fuse the mystical search for new experiences and tion or diminishment of hopes or aspirations;
insights (often through drug use) with an activist weakening of ties to primary support circles (fami-
attack on the dominant culture and its institution- lies, neighborhoods, work groups); and the ero-
al expressions. Survivalists, Amish, and Hare Krish- sion of meaning in the deepest symbols and rituals
nas fuse the radical critique of conventional values of society. These factors are then augmented by
and lifestyles with a withdrawal into an isolated communication among people sharing such expe-
and protected community. riences or beliefs, leading them to coalesce into
normatively and ideologically integrated groups.
Countercultures can be differentiated by their Countercultures can also precipitate social change
primary breaks with the dominant culture. Some if the norms and values they champion are incor-
take odds with its epistemology, or the way society porated into the mainstream. In commenting on
contends that it knows the truth. Hippies and the 1960s student movement, Chief Justice War-
other mystics, for example, have tended to seek ren Burger of the United States Supreme Court
insight in homespun wisdom, meditation, sensory stated that ‘‘the turbulent American youth, whose
deprivation, or drugs, rejecting the rationality of disorderly acts [I] once ‘resented,’ actually had
science and technology. Others assert an alterna- pointed the way to higher spiritual values’’ (cited
tive system of ethics, or the values pursued in in Yinger 1977, p. 848). Lasting influence may not
defining good or striving for the good life. Some, always result from major countercultural move-
like skinheads or KKK members, may be quite ments, as witnessed by the rapid erosion in influ-
conservative in their definition of the good life; ence of Mao’s cultural revolution after his death,
others are libertarian, advocating for people to yet it is possible. This occurs through a cultural
‘‘do their own thing.’’ Still other countercultures dialectic, wherein each existing system, containing
offer alternative aesthetic standards by which fash- antithetical, contradictory ideas, gives rise to the
ion, taste, and beauty are judged. Punk or ac- oppositional values of a counterculture. These are
id rock musical movements, performance or ultimately incorporated into a future new order.
postmodernist art movements, and bohemian or
hippie fashion movements were all aesthetic state- COUNTERCULTURE CASE STUDIES
ments that incorporated a radical rejection of the
standards of conventional taste and its connection While the student movement of the 1960s was
to conventional values. Thus, entire countercultural undoubtedly the largest and most influential coun-
movements may be based on their advocacy of terculture to arise in the United States, a review of
these competing beliefs. three more contemporary American countercultures
may yield further insight into the parameters and
COUNTERCULTURES AND SOCIAL character of these movements. Let us focus on the
CHANGE Hare Krishnas, punks, and survivalists.
Due to their intense opposition to the dominant The Hare Krishna movement, also known as
culture, countercultures are variously regarded the International Society for Krishna Conscious-
as ‘‘engines of social change, symbols and effects ness (ISKCON), is one of the religious movements
of change, or mere faddist epiphenomena’’ (Yinger that became popular in the United States during
1982, p. 285). Examining these in reverse order, the great ‘‘cult’’ period of the 1970s (Judah 1974;
countercultures are often considered mutations Rochford 1985). Its rise after the decline of the
of the normative social order, encompassing such 1960s student movement is not coincidental, for
drastic lifestyle changes that they invoke deep many people who were former hippies or who
ambivalence and persecution. Most major were influenced by or seeking the ideals and values
countercultural mutations appear in the form of of the 1960s turned toward new religions (Tipton
461
COUNTERCULTURES
1982) in search of the same features of communi- survivalists drew on long-standing convictions that
ty, idealism, antimaterialism, mysticism, transcen- an international conspiracy of Jews was taking
dence to a higher plane, and ‘‘a spiritual way of over everything from banking, real estate, and the
life, which stands outside the traditional institu- press to the Soviet Politburo, and that the white
tions found in America’’ (Rochford 1985, p. 44). race was being ‘‘mongrelized’’ by civil-rights legis-
Its primary values conflicting with mainstream lation. A cleansing nuclear war or act of God, with
culture include the rejection of (1) material suc- ‘‘secular’’ assistance, would soon bring the Arma-
cess through competitive labor; (2) education to geddon, eradicating the ‘‘Beast’’ in their midst
promote that end; (3) possessions for sense gratifi- (Coates 1987). Members thus set about producing
cation; (4) authority favoring the status quo; (5) and distributing survivalist literature, stockpiling
imperialistic aggression; and (6) the hypocrisy of machine guns, fuel, food, and medical supplies on
racial discrimination (Judah 1974, p. 16). After the remote farms and in underground bunkers, join-
death of its American spiritual master, Srila ing survivalist retreat groups, and attending surviv-
Prabhupada, in 1977, however, the movement alist training courses (Peterson 1984). Within their
peaked and became more commercialized, trans- retreat communities they rejected the rationali-
ferring its emphasis from self-expression and zation, technologization, secularization, and
uniqueness perpetuation of the sect, thereby be- commodification of society, creating an environ-
coming more of a mass phenomenon. ment of creative self-expression where an individu-
al could accomplish meaningful work with a few
In contrast to the religious and value compo- simple tools. In their withdrawn communities and
nents of the Hare Krishnas’ rejection of main- ‘‘utopian’’ future scenario, men would reclaim
stream culture, the punk or punk rock countercul- their roles as heads of the family; women would
ture of the late 1970s and early 1980s was more of regain mastery over crafts and nurturance. Theirs
a style movement (Hebdige 1981). As Fox (1987, p. is thus a celebration of fantasy and irrationality
349) has noted, ‘‘The punks created a new aesthet- (Mitchell, n.d.). Yet while they isolate themselves
ic that revealed their lack of hope, cynicism, and in countercultures composed of like-minded indi-
rejection of societal norms.’’ This was expressed in viduals, they try to influence mainstream society
both their appearance and their lifestyle. The punk through activism in radical right-wing politics as
belief system was antiestablishment and anarchis- well. Their actions and beliefs, although rejecting
tic, celebrating chaos, cynicism, and distrust of the directions and trends in contemporary society,
authority. Punks disdained the conventional sys- arise out of and represent frustrations felt by
tem, with its bureaucracies, power structures, and embattled segments of the Moral Majority (mainly
competition for scarce goods (Fox 1987). Mem- fundamentalist Christian, white groups).
bers lived outside the system, unemployed, in old
abandoned houses or with friends, and engaged in Scholarly treatment of counterculture move-
heavy use of drugs such as heroin and glue. Hard- ments is not limited to the United States. In the
core commitment was usually associated with field of new social movements research, many
semipermanent alteration of members’ appear- European scholars have looked at organizations
ance through tattoos, shaven heads, or Mohawk that are designed to mobilize forces against na-
hairstyles (Brake 1985). The musical scene associ- tionalistic cultures. These studies, ranging in top-
ated with punks was contrary to established tastes ics from nuclear weapons, ecology, squatters’ rights,
as well and often involved self-abandonment char- gays, women, and other countercultural groups
acterized by ‘‘crash dancing’’ (Street 1986). (i.e., Autonomen or terrorist organizations), have
explored the common denominators inherent in
In contrast to the hippies, Krishnas, and punks, all new social movements. Using quantitative data
the survivalist counterculture was grounded in from protest events collected from newspaper
exaggeration of right-wing beliefs and values. While sources in France, the Netherlands, Germany, and
some of the former groups preached love, survival- Switzerland, Hanspeter Kriesi and others (1995)
ists were characterized by hate. Formed out of outlined the ‘‘new cleavage’’ that exists in these
extremist coalition splinter groups such as neo- Western European societies.
Nazis, the KKK, the John Birch Society, fundamen-
talist Mormon Freemen, the White Aryan Resist- Countercultures thus stand on the periphery
ance, and tax protesters from Posse Comitatus, of culture, spawned by and spawning social trends
462
COUNTERCULTURES
and changes [by their opposition to dominant Western Europe. Minneapolis: University of Minneso-
culture]. As Yinger (1977, p. 850) noted, ‘‘Eve- ta Press.
ry society gets the countercultures it deserves,
for they do not simply contradict, they also ex- Lange, Jonathan I. 1990 ‘‘Refusal to Compromise: The
press the situation from which they emerge. . . . Case of Earth First!’’ Western Journal of Communica-
Countercultures borrow from the dominant cul- tion 54:473–494.
ture even as they oppose it.’’
Mitchell, Richard n.d. Dancing at Armageddon: Dooms-
(SEE ALSO: Alternative Lifestyles; Social Movements; Student day and Survivalists in America. Unpublished paper,
Movements) Oregon State University.
REFERENCES ——— 1991 ‘‘Secrecy and Disclosure in Fieldwork.’’ In
W. Shaffir and R. Stebbins, eds., Experiencing Field-
Adler, Nathan 1972 The Underground Stream: New Life work. Newbury Park, Calif.: Sage.
Styles and the Antinomian Personality. New York: Harper.
Peterson, Richard G. 1984 ‘‘Preparing for Apocalypse:
Baron, Stephen W. 1997 ‘‘Canadian Male Street Skinheads: Survivalist Strategies.’’ Free Inquiry in Creative Sociolo-
Street Gang or Street Terrorists?’’ Canadian Review of gy 12:44–46.
Sociology and Anthropology 34:151–174.
Reich, Charles A. 1970 The Greening of America. New
Berger, Bennett 1981 The Survival of a Counterculture. York: Random House.
Berkeley, Calif.: University of California Press.
Rochford, E. Burke 1985 Hare Krishna in America. New
Brake, Michael 1985 Comparative Youth Culture: The Brunswick, N.J.: Rutgers University Press.
Sociology of Youth Culture and Subcultures in America,
Britain, and Canada. London: Routledge. Roszak, Theodore 1969 The Making of a Counter Culture.
New York: Doubleday.
Coates, James 1987 Armed and Dangerous. New York:
Hill and Wang. ——— 1995 ‘‘Introduction to the 1995 Edition.’’ In
Theodore Roszak The Making of a Counter Culture.
Davis, Fred 1971 On Youth Subcultures: The Hippie Vari- Berkeley, Calif.: University of California Press.
ant. New York: General Learning Press.
Rothchild, John, and Susan Berns Wolf 1976 The Child-
Douglas, Jack D. 1970 Youth in Turmoil. Chevy Chase, ren of the Counter-Culture. New York: Doubleday.
Md.: National Institute of Mental Health.
Saliba, John A. Understanding New Religious Movements.
Erikson, Erik 1968 Identity: Youth and Crisis. New York: Grand Rapids, Mich.: William B. Eerdmans Press.
W. W. Norton.
Short, Brant 1991 ‘‘Earth First! And the Rhetorics of
Flacks, Richard 1971 Youth and Social Change. Chicago: Moral Confrontation.’’ Communication Studies 42:172–
Markham. 198.
Fox, Kathryn J. 1987 ‘‘Real Punks and Pretenders: The Spates, James L. 1976 ‘‘Counterculture and Dominant
Social Organization of a Counterculture.’’ Journal of Culture Values: A Cross-National Analysis of the
Contemporary Ethnography 16:344–370. Underground Press and Dominant Culture Maga-
zines.’’ American Sociological Review 41:868–883.
Hamm, Mark 1993 American Skinheads. Westport, Conn.:
Praeger. Street, J. 1986 Rebel Rock: The Politics of Popular Music.
Oxford: Basil Blackwell.
Hebdige, Dick 1981 Subcultures: The Meaning of Style.
New York: Methuen. Thornton, Sarah 1997 ‘‘General Introduction.’’ In Ken
Gelder and Sarah Thornton, eds., The Subcultures
Irwin, John 1970 ‘‘Notes on the Present Status of the Reader. London: Routledge.
Concept Subculture.’’ In D.O. Arnold, ed., The Soci-
ology of Subcultures. Berkeley, Calif.: Glendessary Press. Tipton, Steven M. 1982 Getting Saved from the Sixties.
Berkeley, Calif.: University of California Press.
Judah, J. Stillson 1974 Hare Krishna and the Countercul-
ture. New York: Wiley. Turner, Ralph 1976 ‘‘The Real Self: From Institutions to
Impulse.’’ American Journal of Sociology 81:989–1016.
Keniston, Kenneth 1971 Youth and Dissent: The Rise of a
New Opposition. New York: Harcourt Brace Jovanovich. Williams, Rhys H. ‘‘Politics, Religion and the Analysis of
Culture.’’ Theory and Society 25:883–900.
Kriesi, Hanspeter, Ruud Koopmans, Jan Willem Dyvendak,
and Marco G. Giugni 1995 New Social Movements in Wuthnow, Robert 1976 The Consciousness Revolution.
Berkeley, Calif.: University of California Press.
Yinger, J. Milton 1960 ‘‘Contraculture and Subculture.’’
American Sociological Review 25: 625–635.
463
The words you are searching are inside this book. To get more targeted content, please make full-text search by clicking here.
Encyclopedia of sociology by Edgar F. Borgatta, Rhonda J. V. Montgomery (z-lib.org)
Discover the best professional documents and content resources in AnyFlip Document Base.
Search
Encyclopedia of sociology by Edgar F. Borgatta, Rhonda J. V. Montgomery (z-lib.org)
- 1 - 50
- 51 - 100
- 101 - 150
- 151 - 200
- 201 - 250
- 251 - 300
- 301 - 350
- 351 - 400
- 401 - 450
- 451 - 500
- 501 - 550
- 551 - 600
- 601 - 650
- 651 - 700
- 701 - 750
- 751 - 756
Pages: