Careers in Print: Books, Journals, and Scholarly Reputations

by Elisabeth S. Clemens, Walter W. Powell, Kris McIlwaine, Dina Okamoto
Citation
Title:
Careers in Print: Books, Journals, and Scholarly Reputations
Author:
Elisabeth S. Clemens, Walter W. Powell, Kris McIlwaine, Dina Okamoto
Year: 
1995
Publication: 
The American Journal of Sociology
Volume: 
101
Issue: 
2
Start Page: 
433
End Page: 
494
Publisher: 
Language: 
English
URL: 
Select license: 
Select License
DOI: 
PMID: 
ISSN: 
Abstract:

CENTENNIAL ESSAY

Careers in Print: Books, Journals, and

Scholarly ~e~utations'

Elisabeth S. Clemens, Walter W. Powell, Kris McIlwaine,

and Dina Okamoto

University of Arizona

Academic reputation rests on publication. But unlike many fields, sociology recognizes both journal articles and books, thereby com- plicating the relation of publication to reputation. Drawing on the sociology of science and organization theory to analyze elite sociol- ogy journals and books nominated for a major prize, the authors show how genre structures scholarly fields and shapes the reception of texts. Method and evidence, not subject matter, distinguish arti- cles from books. Private universities "prefer" books, while scholars trained at public universities are more likely to publish articles. Gender and rank are associated with choice of genre, while citation rates increase with authors' prior publication records. Books gener- ate conversations across subfields and disciplines; articles serve as a currency of evaluation within sociology.

Researcher: I wondered if you could help us fill in some miss- ing data. The Library of Congress catalog doesn't list any- thing under your name, but one of us thought she recalled a book by you.

Author: I don't write books! I'm an article person, I was trained in an article department, and, although this is a book department, I made sure when they hired me that they didn't expect me to write one.'

Among scholars, genre can be a statement of identity. What we write and where we publish may be taken as signals of who we are and how

' For their valuable comments, we thank Craig Calhoun, Paula England, Gerald Marwell, Doug Mitchell, Cal Morrill, Charles Perrow, and the members of the Social Organization Seminar at the University of Arizona. Correspondence should be di- rected to Elisabeth Clemens, Department of Sociology, University of Arizona, Tucson, Arizona 85 72 1.

* The exchanges that begin each section should be read as epigraphs. They are render- ings of conversations that we have had or events that have been described to us in the course of this project.

O 1995 by The University of Chicago. All rights reserved.

0002-9602/96/10102-0006$01.50

AJS Volume 101 Number 2 (September 1995): 433-94 433 we think. In some branches of the academy, entire disciplines are associ- ated with a single genre. Physicists, for example, have been overwhelm- ingly producers of journal articles since the 1890s, which saw "the virtual disappearance of the book as a way of presenting new results" (Bazerman 1988, p. 158). Listen to a group of untenured historians and one learns that, although articles are a nice addition, their futures rest on the com- pletion and quality of "the book." But sociology and some other social sciences differ in recognizing two paths to publication: books and journal articles. Were it simply a question of determining how many articles equal a book, the existence of these two genres would be of limited theoretical interest (although still highly significant in practical terms). However, these paths do not always arrive at the same destination. In- stead, each genre is associated with distinctive career trajectories, institu- tional locations, and patterns of intellectual influence.

Alternative forms of publication not only complicate the merit metric of scholarly prod~ction,~

they signal identity by standing for a series of other salient contrasts: case studies versus hypothesis testing, qualitative versus quantitative, intellectual versus scientist (Wolfe 1990, p. 479). Simpson (1991, p. v) suggests that there are two distinct sociological literatures: books and journals4 Nevertheless, the mapping of genres onto scholarly identities is far from perfect. Article people are found in book departments; book people publish in refereed journals. Although the contrast of book and article appears in the discipline's folklore pri- marily as a classificatory device-sorting individuals, subfields, and de- partments-our concern is whether this contrast of genres can shed any light on the character of this discipline and its relation to others.

There have been numerous attempts to develop a metric that includes both genres in measures of productivity. These efforts underscore the difficulties of simultaneously assessing quality and quantity. For example, "A score of 1 was assigned for each article or essay, and 18 for each book" (Meltzer 1949, p. 26); "For authorship of an original book, a weighting was employed in which every one hundred pages registered as the equivalent of one journal publication, with a maximum upper limit of four quantity counts per book" (Lightfield 1971, p. 129); and another scale in which re- search and theoretical monographs are worth 30 points, an article in AJS or ASR is worth 10, with articles in specialized journals worth still less (Glenn and Villemez 1970; Sturgis and Clemente 1973, p. 170).

This is not to suggest that sociologists publish only in these formats. Chapters in edited volumes are also common and may be the source of innovative work. But our assumption that this is a less important venue is based on the simple observation that tenure at research universities is granted on the basis of books or a set of articles; successful cases based on a collection of chapters are rare. The contrast between sociology and the humanities is suggested by the different rates at which dissertations are published as books. Morton and Price (1989, p. 69) found that 35% of history dissertations were eventually published as books. In contrast, 13% of sociology disser- tations appeared as books, but 39% of the sociologists published part of their disserta- tion research in the form of articles.

"Genre" refers to a type of, or schema for, literary expression and interpretation (Hirsch 1967, chap. 3); it designates the set of expectations associated with a particular literary form.' The absence of a single set of expectations concerning the appropriate form for academic publishing re- flects the multiple functions of publication in science. publication may be a means of conveying information to broader publics, possibly to secure support or legitimacy for adiscipline, or it may be a venue for the communi- cation and evaluation of work among members of a scientific community. Albion Small, the founding editor of the American Journal of Sociology (AJS), was acutely aware of this dual role of scholary publication:

The Journal will thus be primarily technical. It will be devoted to the organization of knowledge pertaining to the relations of men in society into a sociology that shall represent the best American scholarship. On the other hand the Journal will attempt to translate sociology into the language of ordinary life, so that it will not appear to be merely a classification and explanation of fossil facts. As the contents of this number will show, it is not supposed essential to the scientific or even the technical character of thought that it shall be made up of abstractly formulated principles. On the contrary, the aim of science should be to show the meaning of familiar things, not to construct a kingdom for itself in which, if familiar things are admitted, they are obscured under an impenetrable disguise of artificial expression. If sociology is to be of any influence among practical men, it must be able to put its wisdom about things that interest ordinary men in a form which men of affairs will see to be true to life. That form will frequently be the one in which not theorists but men of affairs themselves view the facts concerned. These men are then the most authoritative sociol- ogists. (1895, pp. 13-14)

Claiming both a professional audience and critical lay constituencies for sociological research, Albion Small assigned a dual mission to AJS. The effort to address both topics of public interest and questions of scholarly significance, however, potentially threatened the discipline's ability to control its members' reputations, which are critical components of the organization of work in the sciences (Whitley 1984). Work organizations in which the control of reputation is centralized and clearly bounded differ from those in which it is diffuse or fragmented, although observers may disagree as to which situation is personally or intellectually prefera- ble (Stinchcombe 1994, p. 290). Because scientific work is intended to produce new knowledge or innovation, its quality cannot be judged using routine procedures but requires evaluation by a designated audience.

Although we refer to articles and books as genres, it would be more accurate- though more awkward-to refer to "scholarly articles in the social sciences" and "scholarly books in the social sciences." The former do not evoke the same expecta- tions as a humanistic essay nor the latter the expectations associated with either a textbook or a novel.

Where that audience is limited to other scientists in a particular field, the power of a collegial system of reputational control is enhanced. In the natural sciences, for example, efforts to translate scientific accom- plishments into more general status or celebrity are condemned as "sa- ganization" (Raup 1986, pp. 164-65).~ But when "researchers have a wide variety of legitimate audiences for their work, including educated laymen . . . the need to co-ordinate research results with those of a particular group of colleagues to gain positive reputations is limited, and so contributions to intellectual goals are relatively diffuse and divergent" (Whitley 1984, p. 26). This orientation to multiple audiences may reflect intellectual commitments that value the public relevance of research; a field may also be required to demonstrate its vitality and utility to univer- sity communities, legislatures, funding agencies, and other constituen- cies. Insofar as different genres both embody distinctive expectations and attract somewhat different audiences, a discipline that recognizes multi- ple genres will find it difficult to generate consistent and robust collegial evaluations of reputation. Rather than being simply shorthand for intel- lectual style, the opposition of book and journal sociology is symptomatic of the discipline's distinctive arrangement as a reputational work organization.

Our goal in this article is to examine how these two scholarly genres shape careers, generate professional reputations, and create a structure for the field. We begin by reviewing work in the sociology of science to understand the qualities attributed to each genre as a mechanism for producing and disseminating scientific findings and argument. Because this line of argument has emphasized the role of refereed journals, we then compare books and journals as alternative selection regimes, attending to variations in patterns of access, gatekeeping functions, and the organiza- tion of retrievability. This comparison undermines the strong assumptions about the absence of rigorous evaluation in book publication made in the sociology of science, but also suggests that the influence and acquisition of reputation differ systematically between these two selection regimes.

We focus on elite publications: articles published in our two primary research journals, the American Sociological Review (ASR) and AJS, for one volume year in the period 1987-88, and books nominated for the American Sociological Association's Distinguished Scholarly Publication Award (determined in 1989, awarded in 1990). Compared to other books and journals in the field, these texts are widely distributed and actively promoted. They are, therefore, more likely to reveal the connections among the organization of evaluation, the generation of intellectual con- versations, and the acquisition of reputation.

Current demands that the sciences demonstrate their utility in the form of commer- cial application, however, represent an erosion of collegial control over reputation.

PUBLICATION AND THE PRODUCTION OF KNOWLEDGE

Setting: Departmental tenure review. Chair: We should keep in mind that the first of these articles has received eight citations. Recently promoted associate professor: Is that good or bad?

The organization of scholarly publication has typically been studied from the perspective of natural scientists. Informed by research on publication in physics and chemistry, where refereed journals are the dominant venue for publication, the theoretical frameworks developed for the gen- eral analysis of disciplines typically treat other genres as marginal. In citation studies of the natural sciences, books are routinely eliminated from the research design, although the rationale for this decision varies considerably (see Campanario 1993, p. 348; Cozzens 1985, p. 136; Mo- ravcsik and Murugesan 1975, p. 89). While literature analyses of the social sciences and humanities regularly note the greater prominence of books in citation patterns (Heinzkill 1980), books have been assiduously ignored when sociologists of science address their own discipline. We begin by locating the theoretical sources of this one-sided approach to a two-genre discipline; by treating refereed journals as the crucial mechanism of pro- fessional control, sociologists of science have viewed books as either irrel- evant to the production of scientific knowledge or as evidence of a less than reliable review process. To compare the production system of books and journals more symmetrically, we draw on organization theory, con- cluding that the difference between genres lies less in the rigor of the selection process than in the traits on which selection is based.

Producing Quality and Inequality in Science

Although the printed record has long been used as evidence by historians of science, sociologists were drawn to the study of scientific literature by Merton's account of the professional organization of science as a mecha- nism for the production of novel, objective, and cumulative knowledge. The "ethos of modern science" comprised "four sets of institutional imperatives-universalism, communism, disinterestedness, [and] orga- nized skepticismn-which are "transmitted by precept and example and reenforced by sanctions [and] are in varying degrees internalized by the scientist, thus fashioning his scientific conscience" ([I9421 1973, pp. 269- 70). Viewed from this perspective, the system of scientific publication is an instrument for enforcing these imperatives: "The referee is . . . an example of status judges who are charged with evaluating the quality of role-performance in a social system. . . . Status judges are integral to any system of social control through their evaluation of role-performance and their allocation of rewards for that performance. They influence the motivation to maintain or to raise standards of performance" (Zucker- man and Merton [I9711 1973, p. 460). The refereed journal is a dynamo at the core of scientific endeavor, eliciting new research, ensuring impar- tial evaluation, and disseminating new knowledge.

Merton's theoretical framework poses two basic questions about scien- tific publication. The first concerns the relation of scientific publication to the stratification system of science: Is acceptance for publication a meritocratic process or is it subject to particularistic biases? Are produc- ers of meritorious papers rewarded by positions in prestigious depart- ments or with professional awards? The second turns on the relation of publications to one another over time: Does the publication system pro- duce knowledge that is reliable (and therefore useful for subsequent pub- lications), original or nonredundant, and cumulative? The first question has informed many sociological analyses of publishing in the natural and physical sciences, generating both robust descriptive findings and enduring controversies over how to interpret them. Productivity and prestige are consistently associated, but researchers disagree as to why. Talent and good graduate training may result in both productivity and prestigious employment (Crane 1965), prestigious positions may be a mer- itocratic reward for scientific productivity (Cole and Cole 1973), or presti- gious departments may generate higher levels of productivity, either by providing resources or by instilling norms (Allison and Long 1990). But the difficulty in adjudicating between the effects of prestige and produc- tivity is in part the product of the natural science model of journal publi- cation. Unlike sociology, the referee process in the natural sciences is not double-blind; consequently, a scientist's reputation is a factor in the re- view proce~s.~

The organization of production ensures that reputation, productivity, and publication will be closely related.

While the role of reputation in publication is of considerable interest in the analysis of the natural sciences, it looms larger still when sociologi- cal studies of scientific publication are turned upon their discipline of origin. The question of whether stratification processes are meritocratic is linked to debates over the scientific status of sociology itself. The stakes in this controversy are set by the initial assumptions: if the publication system in the natural sciences is a social control mechanism generating reliable, original, and cumulative knowledge, the operation of the publi-

'The possibility that the lack of a reputation or prestigious affiliation might interfere with a meritocratic review has been the subject of much commentary and some experi- mentation. A resubmission of recently published psychology articles with fictitious names and institutional affiliations resulted in eight rejections and only one acceptance of the nine resubmissions that were not detected (Peters and Ceci 1982).

cation system in sociology can be used to diagnose the scientific status of the discipline.' The impetus toward such disciplinary self-diagnosis was strengthened by Thomas Kuhn's The Structure of Scient$c Revolutions ([I9621 1970). Merton's claim that the referee system functioned as a social control mechanism based on consensus over questions, methods, and quality of findings resonated with Kuhn's distinction between para- digmatic and preparadigmatic disciplines. Sociology's claim to paradigmatic status rested on the proper functioning of its journal system; from this perspective, the discipline's reliance on books was an admission that it remained preparadigmatic:

Today in the [natural and physical] sciences, books are usually either texts or retrospective reflections upon one aspect or another of the scientific life. The scientist who writes one is more likely to find his professional reputa- tion impaired than enhanced. Only in the earlier, preparadigm stages of the development of the various sciences did the book ordinarily possess the same relation to professional achievement that it still retains in other cre- ative fields. And only in those fields that still retain the book, with or without the article, as a vehicle for research communication are the lines of professionalization still so loosely drawn that the layman may hope to follow progress by reading the practitioners' original reports. (Kuhn 1970,

p. 20)~

If sociology is preparadigmatic, this will be evident in the system of journal publication. In the absence of scientific consensus and reliable expert judgment, associations between prestige and productivity will be weak or nonexistent, suggesting that politics or particularism rather than merit is driving career patterns. Research on sociology has supported both sides of this question, documenting associations between produc- tivity and rewards (in the form of recognition and employment; see Lightfield 197 l), as well as enduring patterns of particularism or discrep- ancies between ranking by status and productivity (Sturgis and Clemente 1973; Yoels 1971). The high rejection rates characteristic of sociology journals have also prompted concern, suggesting to some a lack of con- sensus on the part of referees, to others a shortage of pages, and to still others a reason why prominent senior scholars might choose to avoid refereed journals as a venue for their work.

For example, "We hypothesize that quantity of output may be more heavily re- warded in sociology than in physics. When scientists cannot agree upon what high quality is, their concern is likely to be with quantity of output" (Cole and Cole 1971, p 26).

Because the theoretical significance of journal publication lies in its role as a control mechanism or guarantor of quality, the process of book publication is portrayed as correspondingly lax: "For example, if rejection rates are higher in sociological journals than in optical journals, book publication may very well be the outlet most available to sociologists" (Lin and Nelson 1969, p. 49).

TABLE 1 CITATIONSTO BOOKSAND JOURNALS IN SOCIOLOGY

JOURNAL ARTICLES

Journal and Year N References % Books % Journals
ASR 1950 ................................................... 1,016 53.7* 46.3t
ASR 1965 ..................................................... 1,448 61.5* 38.5t
ASR, AJS, and Social Forces 1965-66.. .............. 6,131 63.8* 36.2
ASR 1940-41 ............................................ 569 42 $ 40
ASR 1965-66 ........................................... 1,408 54$ 40
Social Science and Humanities Index, 1970-71 .... 11,130 61.5* 38.5

SOURCES.-Broadus (1952, 1967); Lin and Nelson (1969); Brown and Gilmartin (1969, p. 285); Baugh- man (1974).

* Includes nonserial publications such as unpublished dissertations.
t Includes numbered monographs, yearbooks, etc.
i Figures exclude citations to documents, bulletins, and monographs, which accounted for 18% of

the total in 1940 and 6% in 1966.

The relation between publication and career trajectories may also be assessed by the reception accorded scholarly work. Since the early 1970s when sociology journals were included in the Science Citation Index, citation counts have been viewed as a proxy (albeit problematic) for the quality of scholarly contributions (Chubin 1973; Cole and Cole 1971). Citations also provide a way to address the relation of texts to one another (the degree of knowledge cumulation), the second question implied by Merton's argument. If journal articles cite recent research publications more heavily than classics, this is taken as evidence of progress at the frontiers of the discipline. But while studies of the association between publication and professional status have largely ignored books, studies of the ties linking texts within sociology cannot easily dismiss genres other than refereed journals. Repeated studies of referencing patterns in the sociological literature document the importance of books as resources in scholarly discourse (see table 1). Whereas studies of the natural sciences find that the proportion of total references made to journal articles ranges from 77% for mathematics to over 93% in chemistry, this proportion ranges from 37% to 42% for sociology, placing it in the middle of the social sciences but well above disciplines in the humanities (Heinzkill 1980, p. 357). With respect to the paradigmatic status of sociology and the efficacy of its journals as mechanisms of social control, the persistent prominence of books in the literature can be interpreted either as a threat to the discipline's intellectual integrity or a sign of its absence.

For some members of the discipline, however, the mechanisms that maintain the "scientific integrity" of the journal system also threaten the intellectual vitality of sociology. Where an early editor of ASR confessed his tendency to turn first to the book reviews when a new issue arrived (Hankins 1937, p. 413), the contrast of genres has increasingly served as a touchstone for debates over the intellectual quality and relevance of sociological research.'' Acknowledging the rationale for devolving ASR's book reviews into a separate journal (Contemporary Sociology) in 1972, Norbert Wiley (1979, p. 797) nevertheless argued that "the AJS pattern does make for a more interesting journal in some ways. It dramatizes the split between journal and book sociology, the latter being far more diversified, intellectually open, and close to the pulse of world sociology." In this debate over genre, three concerns loomed particularly large: the increasing presence of a narrow repertoire of quantitative methods (Brown and Gilmartin 1969, pp. 287-89)," the alleged paucity and frag- mentation of theoretical contributions (Wiley 1979, p. 795; Wilson 1979,

p. 805), and the reliance on methodological validity as the sole criterion for review in the absence of common understandings of "generality and novelty" (Zelditch 1979, p. 812). Recently Stephen Cole, now editor of Sociological Forum, echoed this point: "It appears to me that if an author writes an article on a relatively narrow subject, uses commonly accepted methods, and does not try to attack major theoretical issues the chances of the article being accepted are significantly greater than if the author is more ambitious" (1993, p. 337).

Even a brief survey of these debates suggests why the contrast of book and journal sociology endures as a signal of identity within the discipline. For some, the workings of the journal system validate both the internal order of sociology as a profession and its "scientific status" in relation

lo Without identifying specific publication formats as the problem, Alvin Gouldner (1970, p. 18) saw "the call to intellectual convergence and cumulation" as part of "the drive to professionalize sociology . . . an ideology [that] is less congenial to men who see themselves as intellectuals than to those who aspire to be professionals and technicians."

'l Some viewed this turn to quantitative methods as not inherently negative but never- theless problematic when combined with the high rejection rates characteristic of the major journals in sociology: "We are caught between a plethora of manuscripts and a paucity of pages. . . . We often manage this triage-the likely manuscript, the salvageable, and the hopeless-in the simplest way possible: we fix on the adequacy of the methods employed as the clear, immediate basis of evaluation. This puts a high premium on the currently most powerful modes of analysis and tends to relegate matters of theory and substance-the reach and power of ideas-to lesser standing as secondary criteria. . . . The accepted manuscript, then, may become one that treats a very small problem with quite powerful tools. To the extent that this is so, our triage may be based on partial or inadequate criteria. For having shot gnats with howitzers we may have so devastated the landscape that there seems little left to explore" (Wilson 1979, pp. 805-6). But articles that focus on methodology find consid- erable audiences. Pertiz (1983) examined theoretical and methodological articles pub- lished in ASR, AJS, and Social Forces during 1972 and 1973 and found that method- ological articles were cited more frequently.

to other fields. Whereas books are set aside in studies of publishing in the natural and physical sciences because they account for only a small proportion of scholarly citations, sociologists committed to a natural sci- ence model of the discipline downplay the significance of books precisely because they are assumed not to be governed by the organized collegial evaluation typical of the "harder" sciences. But this line of research on journal publishing in sociology has not been accompanied by comparable work on book publishing. We do not actually know that books and articles produce different kinds of knowledge, are subjected to different standards of review, or reach different audiences. We can draw, how- ever, on work in organization theory and the sociology of literature, as well as the reports and reflections of editors, to construct a comparison of how books and journals are produced.

Books and Journal Articles: Alternative Selection Regimes

Books and journals may be compared using a logic of production (Thompson 1967), examining the input (access), throughput (review), and output (retrieval) processes in scholarly publication. Who submits to jour- nals and publishers? How is their work evaluated and what are the key criteria on which editorial decisions are based? How are the articles and books presented to readers? Seen in this light, books and journals evince more similarities than are commonly perceived and their differences are much more subtle. Moreover, practices vary across disciplinary journals much more than is the case with book publishers, where disciplinary differences have limited effect on the production process.

Access.-Manuscripts arrive at journals and publishing houses in dis- tinct ways. At the official journals of professional associations such as ASR, American Economic Review (AER), or American Political Science Review (APSR), editors simply wait for manuscripts. Most of these manuscripts are unsolicited, whereas editors of specialized journals often engage in active solicitation. Neal (1994) reports that in his capacity as editor of Explorations in Economic History, he recruits papers by at- tending conferences and by inviting authors to present at a workshop sponsored by the journal. But whether solicited or not, all papers at scholarly journals are expected to receive comparable treatment. There is a strong expectation that every submitted manuscript will be sent out for review.12 Consequently, who gets published is partly a function of

l2 Many editors report that as many as one-quarter to one-third of papers received could be appropriately rejected by the editor and staff (see Simon 1995), but in practice few papers are rejected without outside review. At AER between July 1, 1992, and June 30, 1993, only 59 papers out of 922 were not sent to an outside reviewer. At ASR 11 out of 608 papers were not sent out for review in 1993.

who submits papers. Patterson et al. (1987) found that as many papers

were submitted to APSR by assistant professors as by full professors and

that the proportion of papers accepted for publication did not vary by

rank.

In contrast, book publishers treat solicited and unsolicited manuscripts very differently. Roughly 75% of manuscripts are unsolicited (Powell 1985; Thatcher 1994), but publishers follow neither the principle of equal treatment nor that of first come, first served. Powell's field study of decision making at two commercial scholarly presses emphasizes that the more personally involved an editor was in acquiring a manuscript, the more likely it is to be published. At one distinguished publishing house, referred to as Apple Press, the numbers illustrated this vividly: of 3,640 unsolicited submissions, 27 were published; of 940 submissions where the author had some form of previous contact with the house, 79 were published; and of the 100 manuscripts that editors had personally solic- ited, 35 found their way into publication. Similar practices are found at specialized commercial monograph houses (Powell 1985, pp. 172-76) and university presses (Topkis 1985; Appel 1994; Thatcher 1994).

At most publishing houses, and especially at elite ones, the demand for editors' time far exceeds its supply. Moreover, waiting for manuscripts to arrive in the mail, as is the case with journal editors, is an ineffective way to obtain high quality manuscripts. Editors report, time and again, that getting out and knocking on doors is the most effective and person- ally rewarding way to find good work (Topkis 1985; Thatcher 1994). By searching for specific titles, editors are attempting to build a "list," that is, to acquire and develop a group of books that relate to each other in an intellectually coherent way so that they constitute more than a mere grouping of titles (Thatcher 1994). This effort at list building has several consequences for who and what gets published. Although many more manuscripts arrive unsolicited than are acquired by an editor, more time is spent on the latter. Put differently, the allocation of an editor's time is a way of organizing his or her obligations. Many unsolicited projects are not reviewed at all, but rejected after a short five to ten minute lookover; other projects receive months of attention. And serendipity, the chance meeting or contact, looms large in the process. Personal rela- tionships, network ties, patronage, and sponsorship all shape the review process in book publishing (Coser, Kadushin, and Powell 1982, pp. 70- 93; Powell 1985, pp. 161-88). A decision to commit resources to publish- ing a book is a business decision, and personal characteristics of the author (i.e., career stage, institutional affiliation, and scholarly reputa- tion) play a critical role in that calculus. In contrast, the decision to review an article is routine, unaffected by the identity of the author. In

sum, access to elite publishing houses is restricted, while access to journal

publication is open.

But in another key respect, books and articles are quite similar: many authors call, but few are chosen. Reports of submissions and acceptances are remarkably constant across social science journals and among univer- sity presses and commercial scholarly presses as well. For example, AER received between 843 and 987 papers a year from 1984 to 1993 and accepted 10%-15% of them. The rate of acceptance fell steadily from a high of 15% in 1984 to an all-time low of 10.4% in 1993, when 900 papers were submitted and 94 published (Ashenfelter 1994, p. 474). The editor of ASR reports an acceptance rate of 13% in 1992 and 12% in 1993. She notes, however, that if revised manuscripts are not considered as separate editorial decisions, a more meaningful statistic would be a probability of 18% that a paper will ultimately be accepted (England 1994, pp. 9-10). The editor of APSR from 1985 to 1991 states that 85%-90% of the papers submitted during his tenure were not accepted for publication (Patterson 1994).

Data for book publishers are less easy to obtain, but Powell (1985, pp. 161-88) reports that Apple Press received 4,680 manuscripts and propos- als during 1975 and 1976, during which time it published 137 books. Of the 2,340 projects and proposals received in 1975, 60 were offered con- tracts. Figures from nonprofit university presses are comparable. In 1976, Columbia University Press accepted 2.5 % of the manuscripts submitted for publication (Powell 1985, p. 161). In 1981, Princeton University Press accepted 5.3% (Darnton 1983, p. 533); more recently, Penn State Univer- sity Press handled 185 submissions in 1990 while issuing 32 books that year (Thatcher 1994).13 But if the rate of acceptance suggests that book publishing is as, if not more, selective than the journal system, more telling differences are found in the criteria and processes that determine

success.

The review process.-A journal editor does little evaluation upon re- ceipt of a manuscript. The review process at most journals is highly routinized, so the main consideration is the selection of appropriate re- viewers. But one key point of variation among scholarly journals is the number of reviewers called upon to review a manuscript. Some journals routinely send a manuscript to three reviewers, while others rely on two and turn to a third in cases of sharp disagreement. Still others begin with a single reviewer and add more depending on the initial report on the

l3 Note that the university press data cover a wide range of academic disciplines from the humanities to the sciences (currently the hottest field in academic publishing, contra expectations), while Apple Press was exclusively a social science publisher.

manuscript. The journal editor's job is essentially to adjudicate among the reviews. It is a process in which more reviews provide more informa- tion for the editor but at a considerable cost of delay in publication for the article. l4

Book editors do a good deal more evaluative work up front. They want to know which intellectual communities might be attracted to a book, how well it is written, and how much work will be involved in publishing it. An editor has to decide whether to spend time on a particu- lar book and whether a manuscript makes a sufficient contribution. Be- cause editors work in several disciplines, they search for manuscripts that contribute to dialogues across fields. For assistance in evaluation, editors turn to members of the scholarly community, most commonly to authors with whom they have previously worked. Initially only one re- view is required. On the basis of a favorable review, most editors can proceed to offer a provisional acceptance. Later, when a completed manuscript is received, many commercial publishers and most university presses will return to the original reviewer as well as to a second reviewer to assess the manuscript.

A negative initial review generally dampens a publisher's interest. Un- like journal publishing, where two subsequent positive reviews might offset a critical one, a book editor is less interested in summing up a variety of reviews. Reviewers are chosen with care, paid for their ser- vices, asked to write lengthy evaluations, and listened to. In part this is because the book editor is likely to have a personal affiliation with the reviewer, and to ignore his or her advice might risk offending someone who has been a source of valuable counsel. Previously published book authors often become members of a publisher's community; they provide referrals, read manuscripts, and at times become friends or close acquain- tances of the editor. In forging this personal relationship, both authors and book editors practice reciprocity. Successful authors serve as talent scouts and advisors but must protect their reputation as trustworthy informants; in return, their opinions carry weight with their editor to an extent that would be considered undue influence at a scholarly journal. l5

l4 The 1993 report of AER, where the typical paper has two referees, clearly illustrates the link between the number of reviewers and the length of the review process. Rejected papers that went to only one reviewer were decided on within an average of 13-14 weeks, while rejected papers with two referees were dealt with in 17-21 weeks. For those papers turned down after reading by three or more reviewers, 26 weeks appears to be the norm, but 69 papers (out of 246) with three or more reviewers took "36-52 +" weeks (Ashenfelter 1994).

l5 Though this special access might be viewed as undue influence and clearly suggests that books are much more particularistic than journals, two points are worth bearing in mind. First, there are many book publishers, so if the industry is competitive and

Book manuscripts are never reviewed blind; the author's identity and affiliation affect how a manuscript is read and received. In this respect, scholarly presses resemble natural science journals more closely than they do social science journals, which typically use a double-blind reviewing process (the referee does not know the author's identity, nor does the author know the identity of the referee). But there is wide variation among journals in this practice and much debate about the merits of different policies. Blank (1991) reports that among 38 well-known jour- nals in different fields, 11 utilize a double-blind procedure. Although scientific objectivity is sometimes conflated with anonymity, the chemis- try, physics, and psychology journals examined uniformly use single-blind procedures where the reputation of the author is a factor in the evaluation of the work. Political science and sociology use double-blind reviews exclusively. Other disciplines, such as biology and history, use both. In a survey of 38 economics journals, 16 reported using double- blind reviews. The consequences of these different practices are suggested by a unique randomized experiment Blank conducted at AER between 1987 and 1989, where half the papers were assigned to single-blind re- views and the remainder to double-blind. She found significant differ- ences in acceptance rates and referee ratings, with double-blind reviews producing lower results in both cases. The differences were telling: accep- tance rates varied between 5.5% for the truly blind, 16.4% for the pseudo-blind (the 45% of double-blind reviews where referees correctly identified the author), and 15% for single-blind. The anonymity of re- view, along with the variations in the number of reviewers, significantly influences acceptance rates. Hargens (1988) has shown that the use of two initial referees with a third in split decisions produces a lower accep- tance rate than a decision process that utilizes one initial referee and a second only when the first recommends rejection.16 In disciplines where the journal acceptance rate is low and outright acceptance rare, a review process that employs double-blind reviews and multiple reviewers may

diverse a broad range of research programs and scholarly trends will be supported. Consequently, the content of book publishing is much more sensitive to changes in industry structure and the business climate (see Coser et al. 1982) than is the content of journals. Second, this particularism entails vesting authority in elites and having established experts or the "stars" of a discipline serve a key gatekeeping role, an arrangement that some scholars have argued is an effective means for establishing consensus in a field (Cole 1993, p. 338; Pfeffer 1993; Stinchcombe 1994).

l6 This effect is magnified in fields where scholarly consensus is not high. Blank (1991) found a low level of agreement on the quality of papers (approx. 25%) among econom- ics reviewers. Former ASR editor Marwell (1992, p. iii) concurs, asserting that referee agreement in most social science journals is around 25%.

well mitigate various forms of reviewer bias. But such a system also slows down manuscript evaluation, increases the likelihood of rejection, and makes it much more likely that the modal "favorable" response to an author will be "revise and resubmit." Thus seemingly minor differ- ences in the processing of submissions by editorial organizations generate wide variation in outcomes.

Differences in reviewing processes may also have consequences for the trajectory of the cumulation of knowledge. Overall, journal editors seek to avoid accepting flawed papers, however interesting or controver- sial, and tolerate rejecting solid or interesting papers that do not survive the review process. Thoughtful commentaries by editors frequently echo this point (Patterson 1994; Cole 1993). Former ASR editor Rita Simon (1994) avers that what may be lost in this process are works of "imagina- tion, innovation, and iconoclasm" that fail to receive "positive appraisal from competent yeomen who are good at catching errors and omissions but might miss a gem, or at least the unusual, the provocative, the outside the mainstream, submission." Book editors are likely to err in the opposite direction, because high standards of technical competence alone do not ensure an audience and because controversy may well gener- ate interest in a book. In sum, review processes for books and articles favor dissimilar kinds of work and treat potential authors in divergent ways.

These differences continue to loom large at the final and least pleasant stage of the review process, rejection. Patterson and Smithey (1990), in a study of 380 authors turned down by APSR, found that most rejected authors are "repeaters." Only one-third of their sample of authors had never before submitted a paper to APSR. Moreover, of those who sub- mitted to APSR previously, 40% had already been published in the jour- nal. Repeaters try again and quickly-30% of their sample had subse- quently sent another paper to APSR and one-third of those met with success. Journal publishing operates like a time at bat. In contrast, book publishing carries the baggage of past experience. Rejection often pre- cludes any further contact between an author and a publisher, and accep- tance makes future contact much more likely. A successful first book nearly guarantees subsequent publication. With books, reputation and past experience cumulate to shape future outcomes in very discernible ways, while journal publication, particularly with double-blind reviews, dampens the effects of reputation and prior productivity.

Retrieval.-Schudson (1989) begins his discussion of the potency of a cultural object with the concept of retrievability. In the absence of a reader, preferably a reader who will subsequently write (Latour 1987, p. 40), a published idea has no scholarly consequences. If published work

fails to "influence and direct the work of others" (Whitley 1984, p.

23), it will not enhance the reputation of the author. Consequently, the

organization of access to publications is critical.

A fundamental source of differences between books and articles lies in their relation to audiences of potential authors. Some of these differences flow directly from the organization of production and distribution. Jour- nals show up regularly in the mailboxes of a significant proportion of a discipline's membership and probably an even larger proportion of those belonging to active authors in the field. This wide and periodic distribu- tion is offset, however, by the scarcity of subsequent reminders to read; articles may be read whenever one finds time to read the latest issue or when one sees that article cited in subsequent publications or mentioned by word of mouth. Books arrive in quite different ways. In communities blessed with good academic bookstores, new publications are readily available, although the key gatekeeping is performed by bookstore buyers who are attentive to commercial as well as scholarly concerns. Books are reviewed in a wide range of journals, newspapers, and magazines; they are advertised in both scholarly and trade journals as well as in the catalogs that flood one's mailbox; they are displayed at professional meet- ings, often for a number of years after their initial publication." Thus citations and word of mouth complement a well-organized system for drawing attention to books. Recall that, in building a list, book editors explicitly consider the potential of a work to speak to an established literature or to generate controversy. Still, in contrast to the regular distri- bution of journals, the visibility of books is more irregular and less bounded by disciplinary affiliation.

Further differences between the retrievability of books and journals are evident if one imagines a visit to a research library. In both cases, one might be tracking a particular citation, but in the case of a general search, the classificatory categories applied to the two genres differ. Ei- ther an old-fashioned card catalog or a new library database will produce all books (and possibly edited collections) by a particular author or under a specific subject heading in the library's holdings, regardless of their disciplinary allegiance. For articles, by comparison, a selection of disci- pline typically precedes the search for a specific author. So long as schol- ars must rely on the cumulative indexes of the journals themselves, the first step in a search has been to decide which field is most likely to have produced discussions of a particular subject. At this point, the technological foundations of intellectual life are evident; new on-line cata-

"Whereas the book displays are frequently the most crowded room at the ASA annual meetings, at a meeting of biomedical researchers the poster sessions presenting new findings draw the largest numbers.

logs allow for subject area searches of the journal literature that cross

both disciplinary lines within the academy and the boundary between

the scholarly and popular press. Insofar as retrievability shapes the prob-

abilities of reading by various audiences, we expect that these technologi-

cal changes will soon influence the networks established among texts

through citation.

In addition to reviews and catalogs, citations are another method of

retrieving texts. In searches for the key works by a scholar in a discipline

other than one's own, the Social Sciences Citation Index (SSCI) consti-

tutes a guide to reputation, pointing to the most widely cited work of

a particular author. The SSCI is asymmetric, however, reporting only

citations contained in journal articles.

Thus differences by genre in the distribution and retrieval of scholar- ship potentially offset one another. Books are not routinely distributed to many members of the discipline, but they continue to attract attention selectively after publication via reviews and advertisements. The pub- lished discussion of articles, by contrast, is a relatively rare event. Book publishers are keenly concerned with marketability, a consideration jour- nal editors can eschew. How these cross-cutting considerations affect visibility is an empirical question, but we note that Hargens (1991, p. 347) reports that a study he conducted with David Botts found that "the median number of citations books receive approximately equals the median number going to the articles and research notes published by ASR."

Taken together, these observations recall Turner's (1960) discussion of sponsored and contest mobility. The book evaluation process is a system in which reputation and patronage are key, projects are nurtured on the basis of promise, and the rewards of accumulative advantage accrue to those who are productive. Journal evaluation is a more open competition, a contest of resilience and repeated effort. One should not conclude, however, that the key difference between sociology and the natural sci- ences lies in the intrusion of particularism via the organization of book publishing. Both types of mobility also operate in the natural sciences but with one critical difference in the system of publication. Given the predominance of journal publication under single-blind procedures, both the merit of a single text and the reputation of the author are factored into each decision to accept or reject-a situation found in sociology solely in arenas such as National Science Foundation (NSF) reviews and book publishing, although only the former practices equal treatment of submissions. For example, one referee of what would become the founda- tional article in the debate over asteroid impacts and terrestrial extinction altered the review form in order to judge the submission as both "excel- lent and exciting, merits rapid publication (potential)" and "mediocre or

poor, should not be published in Science (actual)." Assessing the paper as "sloppy and incomplete, however brilliant," the referee remarked that, had it been written by a graduate student, he "would give it back to be done right!" But the lead author was a Nobel laureate with a long reputation for imaginative arguments, so the reviewer concluded that "it would be a feather in your cap to publish it in some form. Therefore, I urge that you do everything possible to convince the authors to revise and rewrite for eventual publication in Science" (Raup 1986, pp. 68-69; see Clemens 1986). Thus the organization of publication in many of the natural sciences gives journal reviewers and editors greater scope for cultivating significant papers, even if initially flawed, in the manner of book editors in sociology.

TEXTS, AUTHORS, CAREERS

Researcher: You're part of our sample of article authors, and

I was hoping that you could provide us with some biograph-

ical information we haven't been able to track down.

Author: How can you count me as an article person? I write

books!

If disciplines are understood as reputational work organizations, sociol- ogy may be characterized by its differentiated (rather than unified and clearly bounded) system of evaluation. Recall Albion Small's desire that AJS publish both work of a high theoretical and technical standard- texts that would garner reputation within the discipline-and work that would engage critical lay audiences who are credited with legitimate but nondisciplinary bases for evaluating sociological work. This dual orientation continues to shape the development of the field, because pub- lication is among the most important terrains upon which reputations are earned and careers are constructed. By drawing on disciplinary folklore, research in the sociology of science, and organizational analyses of schol- arly publishing, we develop a set of expectations (shown in italic below) about the relations between publication and careers in sociology. We then assess these claims by examining one sample of highly visible texts.

Expectations

Scholarly publication is a complex phenomenon, involving texts, rela- tions among texts, authors, and professional careers. With respect to the texts themselves, disciplinary folklore provides fairly clear-cut expecta- tions about the differences between genres. If there are book subfields and journal subfields, genres should difjcer by subject matter. The same patterns, the folklore tells us, should hold for evidence and method:

journals should publish more quantitative work, while ethnographic and comparative-historical work are more likely to be published in book for- mat. Analysis of the organization of production also predicts differences between genres. Given the divergent decision criteria employed by editors as well as the organization of distribution and retrieval, we expect that books will be cited more often, and more widely, than articles. In addi- tion, articles should receive a larger proportion of their citations in socio- logical journals, while books should receive a larger proportion from out- side the discipline.

Turning to authors, the sociology of science literature and the produc- tion model converge in their portrait of the linkage of careers and publish- ing strategies. Given the significance of network ties to disciplinary elites in book publishing, we expect that article authors will come from a wide range of universities-elite and nonelite, public and private-while book authors will come from a disproportionately elite institutional base. The links between institutional type and publication patterns, however, may be the result of either the socialization and cultural capital acquired dur- ing graduate school or monitoring, sponsorship, and evaluation for pro- motion by one's employer. The first would predict that one's graduate institution will be associated with publication strategy; the second argu- ment suggests that publication format will be associated with one's insti- tution at the time of publication. The role of reputation and network ties in book publishing also generates expectations concerning the personal characteristics of authors. First, a higher proportion of article authors will be junior scholars; conversely, book authors should be more senior.

This should hold whether seniority is measured by years since degree or by faculty rank. Second, because book publishing has more restricted access, we would anticipate that book authors will have a more substan- tial publication record. Third, insofar as network ties are conducive to homosocial reproduction (Kanter 1977), we expect to find that more fe- male authors will publish under the contest regime of journal publication than under the sponsored mobility regime of book publishing, although this association may be diluted by gendered differences in the choice of topics, methods, and evidence. Finally, to the extent that there are sepa- rate book and journal sociologies, as alleged by the discipline's folklore, we expect that authors' publication records will be specialized by genre and exhibit little crossover.

Beyond the contrasts of genres and authors, citation patterns should reflect both the organization of scholarly work and disciplinary reputa- tional processes. Rather than being an unproblematic indicator of the quality of a given piece of work (an interpretation implicit in the use of citations as a productivity or prestige measure), citations should vary with the number of scholars working on a particular topic, whether within sociology or in other disciplines. A work on Weberian methodol- ogy that is highly cited relative to other publications in the history of theory may nevertheless receive fewer citations than an average work of medical sociology, which can reach audiences of social workers, physi- cians, and policy analysts as well as sociologists (see Ennis 1992). There- fore, we expect that subfields will exhibit strong "adjacency" effects, with higher citation rates for work in areas linked to interdisciplinary projects or applied research, such as life course and criminology. Conversely, core areas such as sociological theory should be cited less outside the discipline

(Crane and Small 1992). In addition, if there are book and journal sub- fields, the more extensive scope of distribution and retrieval for books suggests that adjacency effects will be stronger in book subfields than in journal subfields.

Citation patterns, however, are not independent of the system of aca- demic stratification and the organization of scholarly production. To the extent that reputation matters both in the probability of publishing a book and in the decisions of readers to buy, borrow, read, and reference, we also expect that citations will reflect the institutional visibility and publication career of the author. First, citation counts should increase with the prestige of the author's institutional affiliation. Second, because readers may choose to read on the basis of an author's prior record or go back to a publication on the basis of the author's subsequent reputation,

citation counts should be directly related to the volume of high visibility work published over an author's career. Finally, because books continue to be embedded in a complex gatekeeping system, we expect that the number of citations should increase with the number of book reviews and that the number of disciplines citing a book should increase with the number of disciplines that have reviewed it.

Data and Methods

Our concern is with how genre shapes careers and, in turn, structures a field. As such, we are less interested in the factors associated with getting published than in the social forces that shape the production and recep- tion of scholarly work. Whereas the former question might be addressed by a random sample of publications, this approach would be ill-suited to an exploratory analysis of scholarly impact because a significant pro- portion of articles receive few if any citations.18 In order to elucidate the

'' The actual extent of uncitedness is the subject of some contention. In the natural sciences, 62% of the authors whose work was cited between 1961 and 1980 received fewer than five citations (Zuckerman 1988, p. 527). An additional number may have received no citations at all; reports from the Institute of Scientific Information state that 55% of papers published between 1981 and 1985 had not been cited at all in the

linkages among authors, their institutions, publication outlets, and the reception of scholarly work, we examine a sample of published work that comes with some prospect of visibility. This sample permits us to draw on multiple data sources in order to reconstruct the connections among genres, authors, and the reception of texts. We have focused on publica- tions that play a large role in shaping careers and structuring conversa- tions within sociology, but the limitations of the sample mean these analy- ses are necessarily exploratory.

To assess the contribution of journal articles to scholarly dialogue, we select articles appearing in a volume year of ASR and AJS between 1987 and 1988.19 To gauge the consequences of publishing books in sociology, we examine the 80 books nominated for the American Sociological Asso- ciation's Distinguished Scholarly Publication Award decided in 1989 and awarded in 1990. Most were published in 1987 and 1988, with a few from 1989 (see app. A).'' Not a11 were written by sociologists, but the fact of nomination suggests some reasoned judgment that they represent a distinctive contribution to sociological scholarship. Not all authors ap- pearing in AJS and ASR were sociologists either. Of the first authors for whom we could identify a departmental affiliation in 1988, 16% of book authors and 9% of article authors were located in departments other than so~iology.~~

Publications do not map precisely onto disciplines. Just as sociologists may seek broader audiences for their work, other audiences also try to enroll sociologists in their projects; a survey of U.S. political scientists, for example, found ASR and AJS ranked second and fourth,

first five years since publication (Campanario 1993, p. 347; Hamilton 1990), but criti- cisms point out that these studies do not carefully specify what counts as a "paper" and appear to have included book reviews and other rarely cited genres. Reacting to these highly visible claims of low levels of citing, a study of sociology articles found that 43% of a random sample of articles and research notes were cited in the first year after publication and that after six years 83% had been cited at least once (Har- gens 1991, pp. 345-46).

l9 The AJS authors were all those with articles appearing in AJS, vols. 93 and 94 during the months of July, September, and November 1987 and January, March, and July 1988. (The May 1988 issue included four articles as a special tribute to Otis Dudley Duncan and therefore did not reflect standard selection processes.) The ASR authors were those with articles appearing in vol. 53 during the months of February, April, June, August, October, and December 1988.

20 In a similar study, Sullivan (1994) compared a random sample of 27 of the books nominated for the prize in 1986 to a sample of articles published in the same year.

The proportion employed in departments other than sociology increases to 20% for the set of all article authors, suggesting the importance of coauthorship as a vehicle for interdisciplinary work. We did not count authors with joint appointments in sociology departments as "outside" (e.g., those with a primary appointment in a business school); therefore, these figures represent a conservative estimate.

respectively, among the top social science journals (Crewe and Norris

1991, p. 525).

The decision to select articles and books published primarily in 1987 and 1988 allows us to track the influence of publications over a five-year period. To the extent that sociology resembles journal publication in the natural sciences, a majority of citations will be found in this period (MacRae 1969). To measure the scholarly impact of publication, we rely on two basic indicators: citations in scholarly journals (of both articles and books) and book reviews. With respect to the former, we gather both the absolute number of citations received through 1993 (volume of impact) and the number of disciplines in which the work was cited (breadth of impact). In order to assess the volume of impact, we use the SSCI and the Science Citation Index, both published by the Institute for Scientific Information. By using these indexes, we capture the number of articles that cite a given publication rather than the absolute number of individ- ual citations that would be generated by counting multiple citations within a single article. We include citations appearing in articles, editori- als, review articles, surveys of previously published literature, and tech- nical notes but not citations in either letters to the editor or book reviews. Note that we are using a primarily journal medium (SSCI) to assess the impact of both articles and books. To the extent that distinctive book and journal worlds exist, the impact of books is therefore likely to be underestimated.

In order to capture the breadth of impact, we assign the citing journals to 25 disciplinary categories as follows: anthropology, applied social sci- ences, area studies, business, computer science, demography, education, economics, engineering, geography, history, interdisciplinary social sci- ence (e.g., race and ethnic studies, women's studies), language, law, library and information science, life course, medicine, opinion, philoso- phy, political science, psychology, natural science, social work, sociology, and statistics." While most of these correspond to traditional university departments, the categories of applied social sciences, area studies, inter- disciplinary social science, and life course designate multidisciplinary communities within academia. "Opinion" designates any publication directed at a primarily nonacademic audience. Finally, we note what percentage of each author's citations were self-citations. Contrary to the

22 This was not a straightforward task. A typical difficult case would be deciding the discipline for the Journal of Health and Social Behavior. To resolve such cases, we consulted the 1990 edition of the Index and Abstract Directory, which lists journals by subject, giving a brief description of content for some. If this did not provide sufficient information, we went directly to the journals, examining both article content and self-descriptions given in the journal's front matter.

folklore, these turned out to be relatively rare. In the few cases where self-citations exceeded 30%, the total number of citations was consistently

In the previous section, we argued that the character of the work would influence its reception. To code content, we rely on the categories currently employed to classify books reviewed in Contemporary Sociol- ~gy.'~

Each work is also coded with respect to its primary evidentiary base: original data, primarily qualitative (e. g., original ethnography); secondary evidence, primarily qualitative (e. g., nonquantitative compar- ative historical); original data, primarily quantitative (e.g., original sur- veys or experiments); secondary evidence, primarily quantitative (e.g., secondary analyses of the census); and textual criticism (e.g., commentar- ies on the classics).

Work in the sociology of science suggests that the institutional affilia- tion of an author has a significant effect on the likelihood of publication. The presumption is that the status of the institution provides an aura or imprimatur on the quality of the work. Applying this argument to sociol- ogy raises the puzzle of whether the status of the institution or of the department is more relevant. After many discussions that suggested that the status of the institution was both more enduring and more salient for those outside a given discipline than the reputation of a particular department, we opted for the former. Institutions are classified in six categories: elite public university, nonelite public university or college, elite private university, nonelite private university or college, non-U.S. institution, and other (research institute, government agency, or think tank). We code as elite the top 25 institutions based on the "academic reputation" component of U.S. News and World Report's 1993 ranking of colleges and universities. Unlike this explicitly reputational measure, other rankings such as NSF's list by research funds are heavily skewed by the presence of medical schools and Department of Defense research support.2s

Recognizing that an institutional affiliation may take various forms,

23 A study of articles in high-energy physics found that "no more than 7% of the citations" were made by first authors; the proportion would increase if citations by all coauthors were included (Chubin and Moitra 1975, p. 430).

24 We divided "theory" and "methods," which Contemporary Sociology treats as a single category. The category of teaching and research materials was dropped, and our sample included no publications in the area of sociological practice. For a list of the subject categories used in this study, see fig. 1.

25 We should note that the U.S. News rankings exclude our own university from the top 25, whereas either the NSF listing or departmental rankings would have included Arizona. For graduate students who have not yet received the Ph.D., we coded the institution where they were enrolled.

we collect data on where and when each author received the Ph.D. and

whether or not it was in sociology. We also note each author's gender,

institutional affiliation, department, and rank26 for the years 1988 and

1993. The data on institutional affiliation and rank are drawn largely

from the 1988 and 1993 editions of the Guide to Sociology Graduate

Departments and the 1990 edition of the American Sociological Associa-

tion Member Directory. In cases where the authors are not sociologists,

we consult the 1994 National Faculty Directory as well as Dissertation

Abstracts on Disc. For foreign scholars, we determine current institu-

tional affiliation by consulting The World of Learning (1994).

Reputation and the reception of a work may be the product of one's publication record as well as institutional affiliation. Invoking the "Mat- thew effect" in science (Merton [I9681 1973), this explanation predicts that citations will vary with the author's prior publication record. Thus, an article by someone with prior publications in ASR or AJS should receive more attention than one by a first-time author, even though both appear in the same volume. It is also possible that subsequent scholarly success may lead readers to recover the "early works" of a now promi- nent scholar. To assess this argument, we construct various measures of an author's productivity. To explore differences stemming from the ini- tial reception of a piece, we count publications before 1989. For journals, this count includes all articles appearing in either AJS or ASR." The logic of this restriction rests on our interest in visibility, not simple quan- tity. To generate a count of books published before 1989, we use the Library of Congress Information System (LOCIS), an on-line search ser- vice that provides information about the Library of Congress record. The question of what counts as a book deserves some explanation. Our inter- est is in the volume and visibility of scholarly publication; thus we do not count textbooks, novels, anthologies, collections of previously published works, subsequent editions or reprints of previously counted books, translations of previously counted books, reports or pamphlets of any sort, and nonacademic work. All books authored, coauthored, edited, or

26 We classified rank as: graduate student, nonranked faculty member (e.g., lecturer),

assistant professor, associate professor, full professor, and other. ''We began by using Sociofile records for each author, but this was inadequate be- cause the database begins in 1972 and we detected a number of omissions, particularly in the early years. To construct more complete records of career publication, we supplemented the Sociofile records using AJS and ASR indices that cover their entire history through 1970 and the Cumulative Index of Sociology Journals, 1971-85. We count only articles and research notes; we exclude comments, replies, and review essays because the first two reflect a scholar's propensity to generate or join in contro- versies and the latter is likely to mark an author's centrality in the book world of sociology.

coedited are counted. An author who published at least one book or

article before 1989 is coded as a "veteran." Others are "rookies." We

rely on the same sources to construct a measure of career productivity

(up to and including 1993). We divide the total set of authors (N = 230)

into quartiles twice, first using the number of AJS and ASR articles and

second, the number of books they have published. For articles, the

rounded quartiles are 0, 1, 2-3, and 4 or more; for books, they are 0, 1,

2-4, and 5 or more. We then combined these scores to create a measure

of career productivity, which ranges from 1-6. (An author with only one

book or article would have a score of 1; one who has published 4 or more

articles and 5 or more books would score 6.)

The identification of relations between the characteristics of authors and those of texts or their reception is complicated by coauthorship. In analyses of the relation of author characteristics to publication in different genres, we present results both for first authors alone and for the entire set of authors, recognizing that coauthorship is found more frequently in some styles of work and that the effects of employment location will differ for ranked faculty and for their graduate student coauthors. When we look at the relation between author characteristics and the reception of a text, we restrict the analysis to first authors on the assumption that the reputation of additional authors is less salient to those reading and citing. Given the folklore concerning the personality of different journals, we also look consistently for differences between AJS and ASR but report these only when the findings are of substantive interest.

In linking author characteristics to the reception of texts, we encoun- tered another dilemma: not only can texts have more than one author, authors can publish more than one text. Our sample includes 80 books written by 78 different first authors and 12 additional authors, 90 articles with 88 different first authors and 57 different additional authors (one of whom is also a first author; one person appears as an additional author on two articles). Four people appear in our sample as both book and article authors. Consequently, the size of the sample shifts with the unit of analysis. When comparing the texts themselves, we look at 80 books and 90 articles, analyzing them with respect to first authors and adding comparisons with the set of all authors when relevant. We limit our analyses of careers to 230 discrete individuals, although there are 240 combinations of text and author; multiple appearances in our sample become part of the information on scholarly careers.

RESULTS

We begin with the texts themselves, turning subsequently to their authors and then to the reception of the work. Figure 1 shows little evidence of

% of Publications, by Subject Category

Urban and

Political Processes &

Demography

r

Medical
Sociology

TABLE 2 THE RELATION OF TYPEOF EVIDENCETO CHOICEOF GENRE

Original, qualitative ........................................ 12.2 11 37.5 30
Secondary, qualitative.. .................................... 2.2 2 21.3 17
Original, quantitative* .....................................35.6 3 2 8.8 7
Secondary, quantitative ....................................34.4 31 5.0 4
Textual analysis? ........................................... 15.6 14- 27.5 2 2 -
Total ......................................................... 90   80

* Includes the assembly of data from multiple sources to produce new data sets. "includes work that relies on other academic writings as the primary source of evidence.

the purported substantive differences between book and article sociology. Compared in terms of their subject matter, there are not striking differ- ences between the two genres. A chi-square test, with some categories combined to correct for small cell size, found no significant association (P> .lo) between genre and subject area. Both microsociology and mac- rosociology tend to be more journal based, but the numbers in these categories are small. As one might expect, theory is much more of a book field than an article field, as are political sociology and the life course. The more important point of divide between the genres is found in method and evidence. Table 2 highlights a significant association be- tween genre and evidentiary base (P < .001). Books are much more likely to employ qualitative data, while journal articles more frequently utilize quantitative data.

Books are clearly cited more frequently than journal articles, by a ratio of 3 :1 (see table 3). We present means, standard deviations, and medians because several books garnered an exceptional number of citations, thereby skewing the distribution. The Truly Disadvantaged, by William

J. Wilson, received 512 citations; Science in Action, by Bruno Latour, was cited 238 times; The Moral Dimension, by Amitai Etzioni, garnered 133 cites; and Urban Fortunes, by Harvey Molotch and John Logan (the book that won the Distinguished Scholarly Publication Award), landed

115. To facilitate analysis of the citations, we divide the two samples into quartiles (see app. B). The differences in the number of citations received by articles and books are not large at the lower end of their respective distributions (quartiles 1 and 2) but increase markedly for the most frequently cited texts (quartiles 3 and 4). We look at the citation data in a variety of ways, but the outcomes are consistent: books are always cited more often than journal articles. Given that the SSCI treats

TABLE 3

GENRE N Mean SD Median Ql* Q2* Q3* Q4*

Articles:

Allcitations ....................... 1,022 11.4 9.2 9 2.5 7.1 12.4 24.7

Citations within sociology ..... 549 6.1 5.1 4 1.0 3.6 6.6 12.9

Citations outside sociology .... 473 5.3 5.9 3 .4 2.5 5.3 13.6

Books:

All citations ....................... 2,652 33.2 65.2 14 4.4 10.7 21.1 95.7

Citations within sociology ..... 699 8.7 14.9 5 1.1 3.9 7.4 22.1

Citations outside sociology .... 1,953 24.4 51.6 8 2.4 6.5 13.9 75.1

NoTE.-T~~ sample includes 90 articles and 80 books cited a total of 3,674 times.

* Because of the skewed distribution of citations, we divided the article and book samples into quartiles (Q). Here we report the mean number of cites within quartiles. App. B provides the range for each quartile.

only citations in journal articles, we found the regularity of this pattern somewhat surprising, although it is consistent with the predominance of citations to books in sociological journals (see table 1). We expected that citation patterns might diverge by content or evidentiary base, but, as we shall see below, the pattern of more total citations for books is constant.

Some of the difference in the citation patterns is explained by the underperformance of a sizable number of journal articles. Five articles were not cited at all within five years of publication, while half the articles received nine or fewer cites over the five-year period. The most cited articles (see app. C) are notable for attracting attention from outside the discipline of sociology. Whereas the mean for sociology citations to articles in the top quartile is 13 times greater than the mean for the lowest quartile, this multiplier increases to 34 for citations coming from out- side sociology. Articles resemble books in that there is a large potential for citations from nonsociologists. But unlike books, only the most cited articles receive a majority of their citations from outside the disci- pline.

The influence of other disciplines on the citations to books is even greater, as we expected. We initially thought that this might be because several of the authors of the most cited books are from outside sociology (e.g., Hayduk and Zuboff; see app. D), but when we divide the book sample into quartiles the appeal of sociological books to scholars in other disciplines remains. Regardless of the total number of citations, books

TABLE 4 THE INFLUENCE ON CHOICE

OF PH.D. INSTITUTION OF GENRE

FIRSTAUTHORS ALL AUTHORS Articles Books Articles Books INSTITUTION % N % N % N % N

Elite private.. .... 31.5 28 42.8 33 33.8 49 44.3 39
Elite public ....... 27.0 24 24.7 19 27.6 40 23.9 21
Otherpublic ..... 31.5 28 18.2 14 31.0 45 19.3 17
Other private .... 5.6 5 2.6 2 4.1 6 2.3 2
Non-U.S. ......... 4.5 4- 11.7 9- 3.5 -5 10.2 9-
Total* ..........   89   77   145   88

* Data are missing for the Ph.D. institution of three article authors (one first author) and four book authors (three first authors).

receive a majority of their citations from outside the discipline. On aver- age, books in the least-cited quartile receive 54.5% of their citations from outside sociology; the most cited receive 78.5% of their citations from other disciplines. For articles in the lowest quartile, all but 16% of their citations are found in sociology journals; the most-cited articles receive 55% of their citations outside of the field. Note that within each quartile the books in our sample also receive a greater absolute number of cita- tions within sociology than do the journal articles. While this last finding may reflect greater length as well as genre characteristics, the differences in total pages cannot explain the proportionately higher visiblity of books outside the discipline.

Turning to authors, we first consider whether scholars located in dis- similar positions in the scientific stratification system produce different kinds of work. We anticipated that book authors would come from more elite institutions, because prestige and sponsorship play such a key role in book publishing. The competition for article space is more open; thus article authors should come from a wider universe of institutions. Look- ing at where the first authors received their degrees (table 4), we see these expectations are supported: a higher percentage of article authors received their degrees at nonelite public schools (31.4% vs. 18.2% for book au- thors), while book authors were more likely to earn their degrees at elite private schools (42.8% vs. 3 1.4% for article auth~rs).'~ Among foreign

28 The association between institutional status and genre varies by discipline. In eco- nomics, which is a predominantly journal field, scholars affiliated with elite private universities predominate among the authors of articles in prestigious journals (Bairam 1994). But part of the folklore of that discipline is that only senior scholars write books.

scholars, books are the primary medium for reaching U.S. audiences.

Because article authorship often involves several colleagues at the same

institution, sometimes of different rank, we analyzed the first author and

all author samples separately, but the results are similar.

The connection between institution and publication pattern is ampli- fied when we look at the specific universities where people received their degrees, in order to consider whether individual departments nurture distinctive "cultures of production" (see table 5). There appears to be an important socialization effect, with universities such as Berkeley and Harvard turning out scholars who publish books, while N. Carolina, Wisconsin, and Washington graduates publish articles. Columbia and Michigan appear to encourage both venues. Chicago, interestingly, fa- vors books, but when the Chicago authors turn to articles, they publish in ASR.In general, the comparison of AJS and ASR authors reveals that the latter have been educated at a more varied assortment of institutions. Of course, the small number of authors from any given school means that these findings are only suggestive, but even a cursory glance at the publishers of the books in our sample underscores the link between institutional status and publication that is generated by the personal networks involved in the solicitation and review of manuscripts. The University of California Press issued 14 of the books on our list, while the University of Chicago Press contributed 13. More than one-third of the books nominated for the award come from presses affiliated with elite universities in the United States and, adding in Cambridge and Oxford, the elite university press share is nearly one-half (see app. A).

When we turn to current affiliation (table 6), the importance of elite status is dampened. In part this reflects the fact that the majority of Ph.D.'s in our sample were received at elite institutions. But recall the opening exclamation from the article-based scholar in a book department: one's socialization might well conflict with the norms of one's current employer. Indeed a comparison of tables 4 and 6 suggests that the modal career path for our authors leads from an elite graduate institution (public or private) to a job at a nonelite public university. With respect to gradu- ate training, we found no concentrations of book authors at nonelite public universities (see table 5), but many clusters of book authors appear among those employed at these institutions (see table 7). A few depart- ments both produce students and employ scholars associated with partic- ular genres. Chicago, Harvard, and Columbia continue to be strong book departments, and there is a notable concentration of book authors at a handful of elite private universities. Wisconsin, Indiana, and N. Carolina are still prominent in the article arena. But table 7 also reveals many newcomers, institutions represented in prominent publications by their

TABLE 5

AJS ASR BOOKS TYPEOF INSTITUTION School Authors School Authors School Authors

Elite private ........................ Columbia 2 Chicago 3 Columbia 8 Cornell 2 Columbia 3 Harvard 8 Princeton 2 Northwestern 3 Chicago 6 Stanford 2 Pennsylvania 3 Yale 2

Elite public ......................... Wisconsin 4 Michigan 5 Berkeley 13 Texas 3 Wisconsin 5 Michigan 5

N. Carolina 2 Berkeley 2

N. Carolina 2

Other public ....................... Washington 3 Michigan St. 2 Indiana 2 Colorado 2

Other private ...................... Southern Cal. 2

No~~.-wereport data for first authors only and list only those universities represented by at least two authors in each publication format.

TABLE 6

1988 INSTITUTION Elite private ..............................................

Elite public ....................

.......

...............

Other public ..............................................

Other private .............................................

Research institute or non-U. S. .......................

Total ....................................................

NOTE.-We report data for first authors only; there are few book coauthors, and nearly half of the article coauthors are located at the same institution.

* Data are missing for three book authors.

faculty rather than their graduates: Ohio St., Illinois, Maryland, Kansas, Minnesota, and Arizona. 29

Unlike real estate, location is not everything in academia. What one writes and where one publishes are also shaped by one's history in the profession. There are numerous reasons to expect that article authors should be younger than book authors. Access to scholarly presses depends in part on reputation, which must be built from previous writings. At- tending a graduate program with active collaborative research shops pro- vides opportunities for article publication at an early career stage. Also, many departmental decisions about tenure are heavily weighted toward the quantity of output a junior scholar produces. Publishing a variety of articles in top journals is a clear path to tenure, while letting one's chances rise or fall on a single book is a more risky career strategy. Not surprisingly then, we find the article authors to be younger than book authors, with the mean Ph.D. date of 1977 for the former group and 1972 for the lattere30 Table 8 clearly illustrates this division: more than one-third of journal authors are found in the graduate student and assis- tant professor ranks, while none of the book authors are graduate stu- dents and more than half are full professors. Contrary to the discipline's folklore, however, many full professors continue to publish in the leading refereed journals, particularly in ASR.

29 Authors affiliated with these institutions appear at least four times in our total sample of first authors. Neither of the Arizona article authors in our sample had an appointment in the sociology department, underscoring the interdisciplinary character of even our discipline's flagship journals.

30 These are rounded figures. The exact figures are 1,972.15 for book authors and 1,976.57 for article authors. There is a significant (P < ,001) difference between the academic age of the authors in the book and journal samples.


TABLE 8
  AJS AUTHORS   ASR AUTHORS     BOOKAUTHORS
First All First All   First All
1988 RANK % N % N % N % N % N % N
Graduate student ........ ...... 0 0 5.5 3 0 0 6.5 6 0 0 0 0
Assistant professor.. ......... . . 38.9 14 32.7 18 27.8 15 25.8 24 15.0 12 14.1 13
Associate professor .......... . . 16.7 6 21.8 12 24.1 13 23.6 22 22.5 18 21.7 20
Full professor... . . . ....... . . . .. . 30.6 11 29.1 16 46.3 25 35.5 33 52.5 42 51.1 47
Other* ............................ 13.9 5- 11.0 -6 1.8 1 8.6 -8 10.0 8- 13.0 12-
Total N........................   36   5 5   54   93   80   92
* Nonranked faculty (e.g., instructor), independent scholar, research position, etc      

Given that book authors are more senior and access to publishers is more restricted, book authors should have a more substantial career pub- lication record. Looked at as a simple comparison of rookies (no prior articles in AJS or ASR and no prior books) and veterans (more than one book or article prior to 1989), there is very little difference; 23% of the first authors in both categories had previously published neither a book nor an article in AJS or ASR. But when we analyzed the two groups using the career productivity measure, which weights article and journal publications as well as various combinations thereof, we found more complex patterns (see table 9). The distribution of article and book au- thors across different productivity levels is not strikingly different, al- though first authors in the article sample are more likely to have a score of 4 or greater (45.5% vs. 35.9% for book authors). These are also sugges- tive differences between ASR and AJS-the former publishes a lower proportion of both novice authors and senior scholars with a significant publishing record in both genres (33.3% of AJS first authors had scores of 1 or 2 vs. 22.2% for ASR first authors; 33.3% of AJS first authors had scores of 5 or 6 vs. 24.1% for ASR). This result is certainly consistent with the folklore that describes ASR as a "quantitative" or even "normal science" publication and therefore likely to attract highly productive article specialists, given the association of genre with quantitative evi- dence and methods. AJS, often perceived as the more "literary" journal, does publish more authors with a strong record as book authors (a score of 5 or 6 means that an author of four or more major journal articles has published at least two books or that the author of two or three major journal articles has published five or more books). Given that we know that books and articles differ by method and evidence, AJS appears more open than ASR with regard to content and style of sociology. In effect, AJS serves as a "big tent" for diverse methodologies. But authorship in AJS is more restricted to scholars from elite institutions. In constrast,

ASR tends to publish articles that are more restricted with respect to evidence and method but opens its big tent to authors from a wide array of universities.

Thus far we have found that journals and books do not differ in terms of subfield but do diverge with regard to method and evidence. Partici- pants in these publishing systems also differ; article authors are younger on average and more likely to be affiliated with nonelite public universi- ties. But the assumption that there are two separate domains in the sociological literature also requires scrutiny. How much cross-traffic is there between books and journals? We hypothesized that publication records would be specialized by genre, and there is substantial evidence to support this. Looking at our article authors, 62 of the 144 authors (43.1%) had not published a book by 1993; of the 90 book authors, 41

TABLE 9
  AJS AUTHORS     ASR AUTHORS     BOOK AUTHORS
First   All First All   First All
CAREERPRODUCTIVITYSCORE* % N % N % N % N % N % N
1....................................... 22.2 8 27.3 15 3.7 2 12.9 12 15.0 12 15.2 14
2 ....................................... 11.1 4 9.1 5 18.5 10 19.4 18 20.0 16 21.7 20
3 ....................................... 19.4 7 20.0 11 31.5 17 29.0 27 27.5 22 25.0 23
4 .......... ......... 13.9 5 12.7 7 22.2 12 19.4 18 8.8 7 8.7 8
5 ....................................... 19.4 7 20.0 11 14.8 8 11.8 11 16.3 13 17.4 16
6 ....................................... 13.9 -5 10.9 -6 9.3 -5 7.5 -7 12.5 -10 12.0 -11
Total N .......................... .   36   5 5   54   93   80   92
* Career productivity is a six-point scale based on the sum of article productivity and book productivity      

(45.6%) had not published in AJS or ASR by 1993. This suggests a strong separation into different realms of publication, but we also found that a number of authors were extremely prolific and moved freely between different genres. Ten authors in our sample had published 10 or more books in their careers through 1993; three of these scholars are in our sample as "article" authors. Eight authors in our sample had published 10 or more AJS or ASR articles through 1993; one of these is in our sample as a "book" author. Seen in another fashion, the 10 most prolific authors (4.3% of the sample) penned 158 of the 549 books (22%) written by the 230 individual authors in our total sample. Similar concentration is found in the journal arena, where our authors generated 709 articles in AJS and ASR. Eight authors (3.5%) were responsible for 116 of the articles (16.4%), with Leo Goodman leading this list. Moreover, the 21 authors in our most prolific category (author or coauthor of five or more books and four or more AJS or ASR articles) are exceedingly vis- ible sociologists: D. J. Black, J. S. Coleman, N. K. Denzin, A. Etzioni,

C. S. Fischer, F. F. Furstenburg, G. Hage, M. T. Hannan, R. M. Hauser,

M. Hechter, D. Knoke, E. 0. Laumann, N. Lin, G. Marwell, A. Portes,

T. J. Scheff, M. Semyonov, R. G. Simmons, W. F. Whyte, E. 0. Wright, and R. J. Wuthnow. Of these, the late Roberta Simmons is the only woman. We should note that the length of this list almost doubled when we supplemented the current on-line database with older AJS and ASR indexes to include article publication before 1972. There is, it would seem, some truth to the folklore that eminent scholars stop publishing in the leading refereed journals.

The paucity of women in the ranks of the most prolific authors is striking and illustrates the complex ways in which the publication system shapes careers. Following arguments made by Kanter (1977) and many authors in research on work organizations, we argued that women would perform better under the contest system of journals than the sponsored regime of books. The results shown in table 10 are mixed in this regard. Women make up about 9% of the ASR first authors and 31% of the book first authors. When all authors are considered, the percentage of female ASR authors decreases, while the number of female book authors rises to 33%. Authorship at AJS does offer some support for the idea that journals are a contest system-33% of AJS authors are female.31 Half of these women were located at elite institutions, however, compared to

Because the percentages seemed particularly skewed given the composition of our discipline, we also looked at the most recent volumes of AJS and ASR. The overall representation of female authors persists, although the journals are reversed: 90% of the first authors in vol. 98 of AJS and 80% of the first authors in vol. 58 of ASR are male.

TABLE 10
THERELATIONSHIPBETWEEN GENREAND GENDER

AJS:

First authors .................................................. 67 24 33 12

.........................All authors

.......................... 65 36 35 19

ASR:

First authors ................................................... 91 49 9 5
All authors................................................... 92 86 8 7

Book: First authors .................................................... 69 5 5 3 1 25 All authors............

...

..................................... 67 62 33 30

almost 29% for all article first authors and 33% for male first authors in AJS. Even in an apparent contest system, institutional status may have consequences. The intermediate position of AJS relative to books and ASR is also suggested by the relation of gender to patterns of coauthorship; 42% of additional book authors are women, compared to 37% in AJS and 5% in ASR. Closer examination of graduate student coauthors suggests one mechanism by which these gender differences are repro- duced; of the nine graduate students (all coauthors of articles) in our sample, there is only one woman (in ASR). To the extent that socializa- tion in graduate school shapes future publishing strategies (revisit tables 4 and 5), differential exposure to professional publication early in one's career may have lasting consequences. Recall also that who gets pub- lished in journals is partly a function of who submits.32 We do not have this information for our sample, but information provided by the current editor of ASR sheds some light on submissions; recently, 25% of the first authors submitting to ASR have been female, 66% male, and for 9% gender could not be determined from the name alone. For the same period, the rates at which male and female first authors received rejec- tions, revise and resubmit, or (conditional) acceptance never differed more than 5%, and the direction was not consistently in favor of one

32 Conversations with former ASR editor Marwell (personal communication, Decem- ber 1992) confirm this point. Despite various efforts to attract more submissions by female authors, including presenting articles by women in the ostensibIy more visible lead article slot, he reports that the number of submissions by women did not increase appreciably. Conversations with a handful of recently tenured women at various universities amplify this point: they report submitting only a few articles to either AJS or ASR thus far in their careers and speak of matching articles to audiences.

IN THE RELATION TO GENDER

TABLE 11 THE ROLE OF EVIDENCE OF GENRE

MALE   FEMALE
EVIDENTIARYBASIS %   N % N
Qualitative: AJS ......................................................... ASR ........................................................ Book........................................................        
Subtotal.................................................... Quantitative: AJS ......................................................... ASR ........................................................ Book........................................................        
Subtotal.................................................... Textual analysis: AJS ......................................................... ASR ........................................................ Book........................................................        
Subtotal.................................................... N ...............................................................        
NoTE.-W~report data for first authors only.        

group or the other (Paula England, personal communication, August 1994).~~

To evaluate the representation of women in the major journals adequately, we would need to know the distribution of men and women across type of university, rank, subfield, and methodological orientation.

Further comparison of publications by men and women underscores the importance of method and evidence in shaping publication strategies. In table 11, we examine differences in the evidentiary basis of publica- tions by men and women. To note only the most striking results, women were over eight times as likely to be first authors of qualitative pieces in US; men were six times more likely to be first authors of quantitative pieces in ASR. The association of gender and evidentiary base is also clear for books; 45.2% of the female first authors in our sample wrote books based on qualitative evidence. By comparison, the modal format for male first authors was the quantitative article in ASR. As we shall see when we turn to the association of evidentiary base, genre, and rates of citation, these differences have powerful implications for the career trajectories of men and women.

33 These figures are based on the combined periods of 1992 and July-September 1994.

Taken together, these findings suggest a much more variegated view of our field than as a discipline bifurcated into the worlds of journal articles and books. Admittedly, almost half of our sample of book authors have not yet appeared in the most visible article arena and more than one-third of the AJS and ASR authors have not, as yet, published a book. But a select group operates in both arenas and publishes extensively, confirming observations in the sociology of science that a limited number of scholars do an inordinate amount of the most visible research in many fields and that women are comparatively rare in the ranks of these influ- ential scholars (Cole 1970; Zuckerman, Cole, and Bruer 1992). These results are also consistent with DiMaggio's (1987) analysis of the relation of "artistic classification systems" (or the system of genres) to social stratification. Rather than finding a relation between a clear hierarchy of genres (e.g., high culture vs. popular culture) and class position, Di- Maggio argues that "the well educated and persons of high occupational prestige do and like more of almost everything. The reason for this find- ing, so at odds with conventional notions about the isomorphism of taste and class . . . , is that wide-ranging networks require broad repertoires of taste" (1987, p. 444). Our findings suggest that this is no less true in academia. Just as the multilingual have an advantage over the monolin- gual in the realm of world commerce, so scholars able to master multiple genres are more likely to gain entry to multiple conversations, both inside and outside the academy.

This conversational metaphor reminds us that the reception, as well as the production, of scholarly literature matters for the structuring of a field. To analyze reception, we use citation counts-an imperfect mea- sure, but one that offers a useful window into how scholarship is read and utilized. The level of attention paid to a publication is certainly an indicator of importance and quality, but there are sharp differences in the frequency with which different subfields are cited. The size of the potential audience of readers and citers varies by subfield. Having argued that research with an appeal to an inter- or multidisciplinary audience will be more widely cited, we begin with an examination of how work in different subfields is received (see table 12). Our argument is supported by the reception of work in the life course area; this is the most visible article subfield and one of the most recognized book areas as well, but life course publications receive comparatively few citations within sociology regardless of genre. Social hierarchies, a category that includes work on race, gender, and ethnicity, is also widely cited in both formats. Theory gets much more attention in book form, but its primary audience comes from inside the discipline. Microsociology and culture are the only book subfields that receive more attention from sociologists than from other disciplines. All journal subfields save for life course receive more citations

from within the discipline. In every category save for microsociology,

books receive more outside attention than do articles. Finally, table 12

illustrates vividly that citations to books overwhelm the number of cita-

tions to articles.

These differences among subfields should not be conflated with the styles of work that disciplinary folklore associates with each publication format. The most typical books (those based on qualitative evidence) are not the most cited books: books based on quantitative evidence are more highly cited than those based on qualitative evidence (fig. Z), and those presenting original quantitative findings receive by far the most citations from outside sociology. Remember that books are the most fungible for- mat (i.e., reviewed and marketed widely, cataloged by subject rather than discipline) for presenting sociological research to a broad audience. These results remind us that scholars from other disciplines turn to sociol- ogy for its systematic empirical evidence. Echoing our analysis of career productivity, this suggests that the ability to combine the styles of evi- dence and argument associated with different genres characterizes the most visible sociological work.

What effect does academic affiliation have on the visibility of research? Because the average Ph.D. year was 1972 for the book authors and 1977 for journal authors, we would expect to find little continuing effect of the prestige of one's graduate institution. We found none and do not report the results here. But figure 3, which groups citations by the 1988 institution of the first author, suggests a pronounced prestige effect for both journal articles and books, with the most citations going to authors at elite private schools. Note that this advantage is driven by the higher numbers of citations from outside sociology in both cases. These schools are also disproportionately located in urban areas, thus linking authors both to rich local networks of scholars and serving as nodes in the na- tional and international circuits of academic visitation (see Wolfe 1990). The greater number of outside citations received by book authors who are at private schools-whether elite or not-also suggests that these institutions may foster closer ties among disciplines, through either per- sonal networks or interdisciplinary seminars, thus enhancing the network potential of their faculty.

Although institutional status augments individual reputation, presti- gious appointments may also reward hard work and important contribu- tions. To examine whether a scholar's past record influences the reception of current publications, we calculate the career productivity (on a scale from 1 to 6) for each first author. We then take the number of citations for each text and divide them into citation quartiles (see app. B). If frequency of citation is related to the amount of highly visible work an author has published over his or her career, then the mean career

a,

U

3

%

.3

'a,

I.

P

Lo

.-E 2

gas

5

Article authors.. ............... 2 .58 2.80 3.38 4.15
Book authors .................. 2.96 3.30 3.60 4.10

* Citation quartiles are based on the total citations received by an article or book (see app. B). For articles, Q1 equals 0-4 cites, Q2 equals 5-9 cites, Q3 equals 10-15 cites, and 44 equals greater than 15 cites. For books, Q1 equals 0-7 cites, Q2 equals 8-14 cites, Q3 equals 15-33 cites, and Q4 equals greater than 33 cites.

t Career productivity scores range from "1" (one article or book) to "6" (four or more articles and five or more books).

productivity score ought to increase as we move from the less frequently cited quartile to the more frequently cited quartile. The results, presented in table 13, provide support for the productivity hypothesis or the Mat- thew effect. For both book and article authors, the more an author has published, the more frequently that author's work is cited. When we examined the association of career productivity with citation counts for all first authors, however, the relationship proved significant only for articles (P = .005). This finding is consistent with the greater embed- dedness of articles within disciplines; readers are likely to have more information about prior productivity that they can use as a guide in selecting what to read and cite. Thus the role of prior reputation in shaping the reception of articles also qualifies our portrayal of sociological journals as a system of contest mobility-the review process may be organized to exclude an author's reputation, but such procedural blinders do not apply to the reception of texts.

Thus far we have proceeded as if characteristics of texts and authors directly influence the reception and citation of sociological work, forget- ting the distribution and gatekeeping systems responsible for the avail- ability and retrievability of texts. For books, it is possible to assess the impact of one element of the gatekeeping process by examining the rela- tion between book reviews and citation patterns. Within our sample, books received a minimum of two reviews and a maximum of 39; the median number of reviews was six. The number of reviews was signifi- cantly associated with the total citations of a book (P = .0001). Only one book in the top citation quartile received the minimum of two re- views. The book review process also appears to be important in account- ing for the broader extradisciplinary reach of books in sociology. Only three of the 80 books were reviewed in just one discipline; the median number was four. The Transformation of Old Age Security, by Jill Qua-

dagno, and George Washington: The Making of an American Symbol, by

Barry Schwartz, were reviewed in the maximum of eight different fields.

More than one-third of the books were also reviewed in at least one

opinion journal or newspaper, thus receiving some exposure to nonaca-

demic audiences.

The gatekeeping function of reviewing has clear consequences for the visibility of a book. The number of disciplines whose journals review a book is significantly associated with the number of disciplines by which a book is cited (P = .0034). Only one book was never cited, while five were cited solely in one discipline. The median number of disciplines citing was five and the maximum was 20, reached by both Latour's Science in Action and Zuboff's In the Age of the Smart Machine, followed by Wilson's The Truly Disadvantaged, which was cited in 17 fields, and Hayduk's Structural Equation Modeling with LISREL in 16; Etzi- oni's The Moral Dimension and Logan and Molotch's Urban Fortunes (the winner of the publication award) were cited in 12.

These measures of multidisciplinary reach confirm the differences be- tween genres evident from the number of citations outside sociology. Of the articles in our sample, only "Intergenerational Consequences of Fam- ily Disruption" by McLanahan and Bumpass was cited in 10 areas. Three other articles were cited in eight fields: Hilgartner and Bosk, "The Rise and Fall of Social Problems"; Osgood, Johnston, and O'Malley, "The Generality of Deviance in Late Adolescence and Early Adult- hood"; and Sampson, "Urban Black Violence" (see app. C). These spe- cific examples of widely cited articles and books support our contention that subfield influences citation rates; each of these texts speaks to a multidisciplinary community defined by interests such as public policy, science and technology studies, life course, and methodology.

Viewed through the lens of its publications, sociology appears as nei- ther a clearly bounded, collegial "reputational work organization" nor as a discipline divided into separate worlds of books and refereed journals. Instead, our analyses generate an image of sociology as a discipline em- bedded in multiple fields, potentially speaking to many audiences. In this respect, Albion Small was prescient. But the heterogeneity of publishing within sociology is not neutral with respect to the careers of sociologists and the development of the field. Refereed journals and books represent distinctive selection regimes that may be more or less open to scholars with different graduate training, at different institutions, and working with different styles of evidence. As we suggested earlier, refereed jour- nals may be understood as a contest regime, open to comers from a variety of institutional locations, although perhaps less ecumenical (in either attracting or selecting submissions) with respect to types of evi- dence and analysis. The system of book publishing more closely resem- bles a system of sponsored mobility in which network ties and institu- tional affiliation influence access, but it also resembles refereed journals in the natural sciences in its reliance on highly published senior scholars as the key gatekeepers in the review system and the inclusion of an author's reputation as an element to be weighed by the referees.

CONCLUSIONS AND IMPLICATIONS

Setting: Lunch at a tavern.

Participants: Senior faculty member from a large public uni- versity and five junior faculty from an elite private uni- versity.

Seniorfaculty member: So, why don't each of you tell me what you're working on? Junior faculty (in unison): My book!

Our results support the Mertonian view that science is a social institution, but the folk knowledge concerning how the institutional structure of sociology operates is, we learn, often misguided. In several key respects, our findings transform simple conceptions of the discipline into an Esch- erlike portrait, bringing into question many of the categories commonly used to talk about scholarship. For example, books and journals do not differ markedly in terms of subject category; their primary point of diver- gence is in method and evidence. But here's the rub: while books are more likely to use qualitative evidence, the most widely cited type of book presents quantitative results. And subject category matters not because of intellectual content but because of the size of the audiences found in related disciplines for different varieties of sociological work. These find- ings suggest not only the need for a more nuanced description of the structure of the discipline but also the possibility of more effectively meeting the dual goals spelled out by Albion Small (1895, p. 14) a century ago: "the aim of science should be to show the meaning of familiar things, not to construct a kingdom for itself in which, if familiar things are admitted, they are obscured under an impenetrable disguise of artifi- cial expression. If sociology is to be of any influence among practical men, it must be able to put its wisdom about things that interest ordinary men in a form which men of affairs will see to be true to life."

With respect to the organization of the discipline, our findings suggest that sociology should not be understood as a literally "invisible college" (Crane 1972), or network of scholars that transcends place and distance. Location does matter. We found that institutional origins had a strong influence on choice of genre, with a clear separation between elite private universities, which favor books, and public institutions, which encourage articles. The heat generated by the contrast of genres may, therefore,

indicate a discipline-specific form of class warfare. Prestige operates in an unusual fashion, however. Apparently graduates of elite private schools learn the ropes in a different manner than graduates of public (and presumably larger) schools. But while graduate programs shape the choice of genre, the effects of socialization may be suppressed by the markets (and networks) for jobs and ideas. When we turn to the reception of texts as measured by citations, the prestige of an author's current affiliation matters for articles, but book citations are less sensitive to academic status. Book authors at private institutions, whether elite or not, tend to receive more citations. This finding suggests that location is important because institutions of different types are embedded in distinc- tive scholarly networks. Books are more likely to be the coin of the realm at private institutions, while articles are the primary currency at public universities. At private universities, local reputations may be valued as much as one's disciplinary standing. A scholar's status on her or his home campus is more influenced by the publication of a book, because books are more accessible to colleagues in other fields. Intellectual discus- sions at private universities typically span departments, and a review in a noted journal of opinion is a clear signal of importance. At the large public universities, the cultures of production are less oriented to the university community than to the department's standing in the disci- pline's hierarchy (namely, the frequency with which department chairs produce "bragging sheets" that compare departments by per capita pub- lications in the top journals). Whether the goal is attracting graduate students or protecting department budgets, publication in elite journals provides a highly visible marker of status within sociology.

Just as public and private institutions appear to support distinctive cultures of production, the careers of individual authors provide further evidence of the partially differentiated worlds of book and journal sociol- ogy. A sizable number (almost half of the book authors and more than one-third of the article authors) have not yet crossed over into the other genre, but more than half of the article authors did publish books and some of the articles written by younger authors may well yet culminate in a book. More significantly, a small handful of prolific and frequently influential sociologists operate in both markets. Curiously, however, early experience with article publishing proved not to be a harbinger of professional socialization and success. More than half of the graduate student coauthors had not completed their degrees five or more years later, and several had left the field.

The reception of texts reveals further intricacies in the publication system. Recall why the system of publication has been central to work on the sociology of science: refereed journals constitute a mechanism for the expert evaluation of work and the distribution of these certified results in a way that allows for the cumulation of knowledge within a particular discipline. Both the expert evaluation and the subsequent reception and referencing of these works then contribute to the construction of scientific reputation and the organization of career trajectories. We find, however, that the citation patterns for sociological texts do not conform to this model of discipline-governed reputation. Instead, outsiders to the disci- pline play a critical role in determining the success of a text as measured by the frequency with which it is cited. The higher citation rate for books is largely attributable to their reception outside of sociology. Moreover, the ratio of outside- to inside-sociology citations increases with the total number of citations. This outreach quality of sociology cannot be easily dismissed, because both legitimacy and resources result from attention by other disciplines as well as from outside the academy. But the multi- plicity of audiences unquestionably generates an enduring tension be- tween the construction of broad-based scholarly reputations and the ca- pacity of the discipline for the intellectual self-governance that some claim ensures the cumulation and coherence of knowledge.

These centrifugal forces are felt with particular keenness when we look across subfields within sociology. Those areas typically deemed central to the discipline and accorded intellectual prestige by sociologists are precisely those least likely to connect with other disciplines and audi- ences. Thus works in core areas such as sociological theory are handi- capped with respect to the allegedly neutral index of citation counts. The opposite is true of applied areas, which are often seen as less central to the enterprise of sociology but which are better positioned to garner atten- tion from a range of adjacent fields, both academic and practical. The inverse relation of prestige within the discipline and the potential for academic reputation is illustrated by the audiences for some of the most highly cited texts. Among the 10 most cited articles, the proportion of cites from sociological journals ranged from 13% for J. Richard Udry's "Biological Predispositions and Social Control in Adolescent Sexual Be- havior" to 86% for "The Transformation of the American Class Struc- ture, 1960-1980," by Erik Olin Wright and Bill Martin, a piece that speaks to central issues in stratification and Marxist theory. Among the 10 most cited books, Bruno Latour's Science in Action received only 11% of its citations from sociology journals, while Andrew Abbott's The System of Professions received the largest proportion of citations (41%) from within the discipline. So while books regularly reach a broader range of audiences than do articles, even among books there is consider- able variation in the embeddedness of texts in central disciplinary con- cerns.

How texts are received shapes both scholarly recognition and disciplin- ary reputation. Yet, it is difficult for sociologists to develop consistent standards of evaluation, because the relation of texts to audiences is so variable. But reputation is the key resource for organizations and fields that lack more immediate and discretionary forms of bargaining power. According to sociologists of science (Merton 1973; Cole and Cole 1973; Whitley 1984), recognition is the primary reward for scientific productiv- ity, but the results of our exploratory study suggest that this universalistic norm comes into direct conflict with a disciplinary structure in which productivity and reputation may be orthogonal to one another. Sociology is a field with competing views of what kinds of scientific activity are appropriate. Particular conceptions of what is high quality research are localized in departments, the organizational units that influence resources, careers, and intellectual community most directly. Departments control graduate training, career advancement, and the legitimation of particular styles of research. The performance criteria by which sociolo- gists evaluate high-impact work, however, are subject to contestation. We find notable differences between public and private universities and across subfields, in terms of what genre of work is likely to be highly regarded. Moreover, the great majority of attention to sociological re- search (as measured by citations) comes from outside the discipline.

We draw several implications for sociology as a reputational work organization from these findings. Discipline-wide authority is rendered fragile by the presence of multiple standards of evaluation as well as by shifts in external pressures and rewards. Articles and books constitute different routes to reputation: articles may provide recognition by one's disciplinary colleagues, while books offer the possibility of more universal academic renown. We know from research on collective memory, how- ever, that selection for commemoration derives less from past achieve- ment than from relevance to contemporary issues (Lang and Lang 1988; Schwartz 1987). Consequently, books that are widely discussed may not speak to core topics in the discipline. Books appear to facilitate entry into intellectual and lay communities that transcend immediate networks of colleagues, but the discipline has little authority to police these linkages or to shape the standards by which work is evaluated. Articles establish disciplinary authority but seem to do so at the cost of appealing to broad audiences.

These findings have practical implications as well. Above all, we find that the system of publication in sociology generates problematic indica- tors of quality, novelty, or the other factors thought to underlie scholarly reputation. The number of a scholar's publications is not a reliable indi- cator of quality, because there is huge variation in the number of citations received by articles appearing in the most prestigious journals and by books nominated for the most prestigious awards. Nor are citations an unproblematic indicator, because they vary in systematic ways by genre, subfield, and institutional location.

The publishing system of sociology is a crucial terrain upon which ideas are conveyed and reputations constructed. So, one might ask, would it be desirable for sociology to be both more coherent internally and to have more influence on other disciplines or lay audiences? If the answer is "yes," this study suggests several strategies. First, quantitative people should be encouraged to write books and book people should be encouraged to use quantitative data, recognizing that systematic quanti- tative evidence is a valued product of our field so long as it is presented in a form that is accessible to nonspecialists. Second, qualitative people should think about how to achieve greater cumulation. This need not mean embracing a model of "normal" science; efforts to build denser linkages among texts (whether or not they are limited to the discipline) would be equally valuable. Whether driven by an ethnographer's taste for interpretation over theory development (Snow and Morrill 1994) or by a literary "mania for originality" (Anheier and Gerhards 1991), the absence of strong and dense ties among texts (as represented by the prac- tice of citation) may facilitate the marginalization of some types of schol- arship. But one might also answer no, that integration and influence are not ultimate goals. Although "the wide variety of substantive subject matter in disintegrated disciplines, and the strong boundaries around each substantive specialty, means that people cannot get interested in each other's work," Arthur Stinchcombe (1994, pp. 290-91) argues none- theless that "this disintegrated state of sociology represents the optimum state of affairs, both for the advance of knowledge and for the expansion of mind of undergraduates" even if it renders the discipline organization- ally vulnerable to university administrators. If in sociology, "the advance of knowledge goes on with many different methods, many different theo- ries, and with many different relations to ideological, granting agency, and theoretical objectives" (pp. 290-91), the selection criteria embedded in different publishing systems should be examined critically, with an eye to identifying undue constraints on fruitful intellectual variation.

Short of changing the structure of the field, these findings also suggest how individuals might adapt to the existing system of opportunities and incentives. These implications are particularly clear with respect to the different relation of male and female scholars to the publication system. Women should be aware of the possible career consequences of opting for qualitative methods. Journal referees as well as faculty in charge of methods sequences should be sensitive to the consequences of canonizing a particular range of methodologies. Those doing qualitative work could also respond collectively by changing the practices that make it difficult to secure reputation (e.g., the failure to acknowledge intellectual influ- ence; see Anheier and Gerhards 1991) or by mobilizing new audiences. In this respect, it is significant that the one way in which female authors appear advantaged over their male counterparts is that books by women are more likely to receive reviews in nonacademic publications (44.8% of the first authors who received outside reviews were female, as opposed to 20% of all first authors of books). Access to feminist publications aimed at an educated but not specifically academic audience accounts for much of this difference. Finally, for both male and female scholars at public universities, the lack of access to interdisciplinary networks and audi- ences is a point of concern. Finding ways to promote interdisciplinary conversation, whether through faculty seminars or support for attending conferences, would be beneficial from either a careerist or an intellectual perspective.

Thus, rather than conceiving of sociology as a distinct intellectual en- terprise, it may be more useful to view it as a field of practice nested in multiple institutions. In this field, defined by the mutual orientation of scholars and employers, authors and audiences, books and articles play different roles. Books are high-stakes endeavors that, when successful, are effective in enrolling allies from neighboring fields. Articles, in con- trast, discipline the troops, generating a common currency of evaluation, be it in comprehensive exams or tenure decisions. To the extent that we care about scholarly reputation, both our discipline's and our own, nei- ther genre should be ignored.

APPENDIX A

Books Nominated for the Distinguished Scholarly Publication Award, 1989

Abbott, Andrew. The System of Professions. Chicago: University of Chi- cago Press, 1988. Abramovitz, Mimi. Regulating the Lives of Women. Boston: South End Press, 1988. Adam, Barry. The Rise of a Gay and Lesbian Movement. Boston: Twayne, 1987. Alexander, Jeffrey. Action and Its Environments. New York: Columbia University Press, 1988. Alexander, Jeffrey. Twenty Lectures. New York: Columbia University Press, 1987. Archer, Margaret. Culture and Agency. New York: Cambridge Univer- sity Press, 1988.

Baumgartner, M. P. The Moral Order of a Suburb. New York: Oxford University Press, 1988. Biggart, Nicole. Charismatic Capitalism. Chicago: University of Chicago Press, 1989. Black, Donald. Sociological Justice. New York: Oxford University Press, 1989. Bloom, Jack. Class, Race, and the Civil Rights Movement. Bloomington: Indiana University Press, 1987. Brustein, William. The Social Origins of Political Regionalism. Berkeley and Los Angeles: University of California Press, 1988. Campbell, Colin. The Romantic Ethic and the Spirit of Modern Consum- erism. New York: Blackwell, 1987. Cancian, Francesca. Love in America. New York: Cambridge University Press, 1987. Coleman, James S., and Thomas Hoffer. Public and Private High Schools. New York: Basic, 1987. Converse, Jean. Survey Research in the U.S., 1890-1960. Berkeley and Los Angeles: University of California Press, 198 7. Daniels, Arlene. Invisible Careers. Chicago: University of Chicago Press, 1988. Darby, John. Intimidation and the Control of ConJEict in Northern Ire- land. Syracuse, N.Y.: Syracuse University Press, 1987.

Denzin, Norman. The Alcoholic Self. Newbury Park, Calif.: Sage, 1987.

Denzin, Norman. The Recovering Alcoholic. Newbury Park, Calif.: Sage, 1987. Dubin, Steven. Bureaucratizing the Muse. Chicago: University of Chi- cago Press, 1987. Ebaugh, Helen. Becoming an Ex. Chicago: University of Chicago Press, 1988. Epstein, Cynthia. Deceptive Distinctions. New Haven, Conn.: Yale Uni- versity Press, 1988.

Etzioni, Amitai. The Moral Dimension. New York: Free Press, 1988.

Fantasia, Richard. Cultures of Solidarity. Berkeley and Los Angeles: University of California Press, 1988. Flacks, Richard. Making History. Berkeley and Los Angeles: University of California Press, 1988.

Furstenberg, Frank, J. Brooks-Gunn, and S. Philip Morgan. Adolescent Mothers in Later Life. New York: Cambridge University Press, 1987. Gans, Herbert. Middle American Individualism. New York: Free Press,

1988.

Gartman, David. Auto Slavery. New Brunswick, N. J.: Rutgers Univer- sity Press, 1988.

Gitlin, Todd. The Sixties. New York: Bantam, 1987. Glassner, Barry, and Julia Loughlin. Drugs in Adolescent Worlds. New York: St. Martin's, 1987. Gottdiener, Mark. The Decline of Urban Politics. Newbury Park, Calif.: Sage, 1987. Greenberg, David. The Construction of Homosexuality. Chicago: Uni- versity of Chicago Press, 1988. Halliday, Terence. Beyond Monopoly. Chicago: University of Chicago Press, 1987. Halpern, Sydney A. American Pediatrics: The Social Dynamics of Pro- fessionalism. Berkeley and Los Angeles: University of California Press, 1988. Hayduk, Leslie. Structural Equation Modeling with LISREL. Baltimore: Johns Hopkins University Press, 1987. Hechter, Michael. Principles of Group Solidarity. Berkeley and Los Angeles: University of California Press, 1987. Johnson, Miriam. Strong Mothers, Weak Wives. Berkeley and Los Angeles: University of California Press, 1988.

Katz, Jack. The Seductions of Crime. New York: Basic, 1988.

Klatch, Rebecca. Women of the New Right. Philadelphia: Temple Uni- versity Press, 1987. Landry, Bart. The New Black Middle Class. Berkeley and Los Angeles: University of California Press, 1987. Larabee, David. The Making of an American High School. New Haven, Conn.: Yale University Press, 1988. Latour, Bruno. Science in Action. Cambridge, Mass.: Harvard Univer- sity Press, 1987. Laumann, Edward, and David Knoke. The Organizational State. Madison: University of Wisconsin Press, 1987. Lawson, Annette. Adultery: An Analysis of Love and Betrayal. New York: Basic, 1988. Lepenies, Wolf. Between Literature and Science. New York: Cambridge University Press, 1988. Logan, John R., and Harvey L. Molotch. Urban Fortunes. Berkeley and Los Angeles: University of California Press, 1987. Lyman, Stanford, and Arthur Vidich. Social Order and the Public Phi- losophy. Fayetteville: University of Arkansas Press, 1988. Marx, Gary. Undercover. Berkeley and Los Angeles: University of Cali-

fornia Press, 1988.

McAdam, Doug. Freedom Summer. New York: Oxford University Press, 1988.

McNall, Scott. The Road to Rebellion. Chicago: University of Chicago Press, 1988.

Melbin, Murray. Night as Frontier. New York: Free Press, 1987. Newman, Katherine. Falling from Grace. New York: Free Press, 1988. Quadagno, Jill. The Transformation of Old Age Security. Chicago: Uni- versity of Chicago Press, 1988. Ragin, Charles. The Comparative Method. Berkeley and Los Angeles: University of California Press, 1987. Reinarman, Craig. American States of Mind. New Haven, Conn.: Yale University Press, 1987. Rosen, Ellen. Bitter Choices. Chicago: University of Chicago Press, 1987. Rule, James. Theories of Civil Violence. Berkeley and Los Angeles: Uni- versity of California Press, 1988. Scheppele, Kim Lane. Legal Secrets. Chicago: University of Chicago

Press, 1988. Schwartz, Barry. George Washington. New York: Free Press, 1987. Schwartz, Gary. Beyond Conformity or Rebellion. Chicago: University

of Chicago Press, 1987. Sica, Alan. Weber, Iwationality, and Social Order. Berkeley and Los Angeles: University of California Press, 1988. Simon, Barbara. Never Mam'ed Women. Philadelphia: Temple Univer- sity Press, 1987. Simmons, Roberta, and Dale Blyth. Moving into Adolescence. Hawthorne, N.Y.: Aldine, 1987. Smith, Barbara. Digging Our Own Graves. Philadelphia: Temple Univer- sity Press, 1987. Smith, Dorothy. The Everyday World as Problematic. Boston: North- eastern University Press, 1987. Sullivan, Deborah, and Rose Weitz. Labor Pains. New Haven, Conn. : Yale University Press, 1988. Sweet, James A., and Larry L. Bumpass. American Families and House- holds. New York: Russell Sage, 1987. Szelenyi, Ivan. Socialist Entrepreneurs. Madison: University of Wiscon- sin Press, 1988. Tropman, John. Public Policy Opinion and the Elderly, 1952-1978. Westport, Conn.: Greenwood, 1987. Turner, Jonathan. A Theory of Social Interaction. Stanford, Calif.: Stan- ford University Press, 1988. Van den Berg, Axel. The Immanent Utopia. Princeton, N.J.: Princeton University Press, 1988. Vanneman, Reeve, and Lynn Weber Cannon. The American Perception of Class. Philadelphia: Temple University Press, 1987.

Wallerstein, Immanuel. The Modern World System, vol. 3. San Diego, Calif.: Academic Press, 1988. Warner, R. Stephen. New Wine in Old Wineskins. Berkeley and Los Angeles: University of California Press, 1988. Westney, D. Eleanor. Imitation and Innovation. Cambridge, Mass.: Harvard University Press, 1988. White, Merry. The Japanese Educational Challenge. New York: Free Press, 1987. Whyte, William Foote, and Kathleen King Whyte. Making Mondragon. Ithaca, N.Y.: ILR Press, 1988. Williams, Bruce. Black Workers in an Industrial Suburb. New Bruns- wick, N. J.: Rutgers University Press, 1987. Wilson, William J. The Truly Disadvantaged. Chicago: University of Chicago Press, 1987. Zuboff, Shoshana. In the Age of the Smart Machine. New York: Basic, 1988.

APPENDIX B

TABLE B1

GENREBY CITATIONQUARTILES

ARTICLES BOOKS
N of Texts Cite Range N of Texts Cite Range
Total citation quartiles:      
Q1. . . . . . . . . . . . . . . . . . . . . . . . . . . . .      
Q2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . .      
Q3. . . . . . . . . . . . . . . . . . . . . . . . . . . . .      
Q4. . . . . . . . . . . . . . . . . . . . . . . . . . . . .      
Sociology citation quartiles:      
Q1. . . . . . . . . . . . . . . . . . . . . . . . . . . . .      
Q2 . . . . . . . . . . . . . . . . . . ... . . . . .      
Q3. . . . . . . . . . . . . . . . . . . . . . . . . . . . .      
Q4. . . . . . . . . . . . . . . . . . . . . . . . . . . .      
Outside citation quartiles:      
Q1. . . . . . . . . . . . . . . . . , . . . . . . . . . . .      
Q2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . .      
Q3. . . . . . . . . . . . . . . . . . . . . . . . . . . .      
Q4. . . . . . . . . . . . . . . . . . . . . . . . . . . . .      

APPENDIX C

10 Most Cited Articles

  1. Sara McClanahan and Larry Bumpass, "Intergenerational Conse- quences of Family Disruption," AJS July 1988 (55 citations).
  2. Robert J. Sampson, "Urban Black Violence: The Effect of Male Joblessness and Family Disruption," AJS September 1987 (3 7 cita- tions).
  3. Alejandro Portes and Saskia Sassen, "Making It Underground: Comparative Material on the Informal Sector in Western Market Economies," AJS July 1987 (32 citations).
  4. Stephen Hilgartner and Charles L. Bosk, "The Rise and Fall of Social Problems: A Public Arenas Model," AJS July 1988 (30 cita- tions).
  5. D. Wayne Osgood, Lloyd D. Johnston, Patrick M. O'Malley, and Jerald G. Bachman, "The Generality of Deviance in Late Adoles- cence and Early Adulthood," ASR February 1988 (29 citations).
  6. Erik Olin Wright and Bill Martin, "The Transformation of the American Class Structure, 1960-1980," AJS July 1987 (26 citations).

6. Paula England, George Farkas, Barbara Kilbourne, and Thomas Dou, "Explaining Occupational Sex Segregation and Wages: Find- ings from a Model with Fixed Effects,"ASR August 1988 (26 cita- tions).

8. Susan P. Shapiro, "The Social Control of Impersonal Trust," AJS November 1987 (24 citations).

8. Michael T. Hannan and John Freeman, "The Ecology of Organiza- tional Mortality: American Labor Unions, 1836- 1985," AJS July 1988 (24 citations).

10. J. Richard Udry, "Biological Predispositions and Social Control in Adolescent Sexual Behavior," ASR October 1988 (22 citations).

APPENDIX D

10 Most Cited Books

  1. William J. Wilson, The Truly Disadvantaged (512 citations).
  2. Bruno Latour, Science in Action (238 citations).
  3. Amitai Etzioni, The Moral Dimension (133 citations).
  4. John R. Logan and Harvey L. Molotch, Urban Fortunes (115 cita- tions).
  5. Shoshana Zuboff, In the Age of the Smart Machine (114 citations).
  6. Leslie Hayduk, Structural Equation Modeling with LZSREL (102 citations).
  7. Dorothy Smith, The Everyday World as Problematic (85 citations).
  8. Frank Furstenberg, Jeanne Brooks-Gunn, and S. Philip Morgan, Adolescent Mothers in Later Lge (75 citations).
  9. Roberta Simmons and Dale Blyth, Moving into Adolescence (69 cita- tions).
  10. Andrew Abbott, The System of Professions (54 citations).

10. James S. Coleman and Thomas Hoffer, Public and Private High Schools (54 citations).

APPENDIX E

Short Titles for Universities

To save space in the text and the tables, we have used abbreviated names to refer to universities. Following is a list of the abbreviated names followed by their complete names.

Arizona = University of Arizona

Berkeley = University of California, Berkeley

Chicago = University of Chicago

Colorado = University of Colorado

Columbia = Columbia University

Cornell = Cornell University

CUNY = City University of New York

Harvard = Harvard University

Illinois = University of Illinois at Urbana-Champaign

IL-Chicago = University of Illinois at Chicago

Indiana = University of Indiana at Bloomington

Kansas = University of Kansas

Maryland = University of Maryland College Park

Michigan = University of Michigan

Michigan St. = Michigan State University

Minnesota = University of Minnesota at Minneapolis St. Paul

MIT = Massachusetts Institute of Technology

N. Carolina = University of North Carolina at Chapel Hill
Northwestern = Northwestern University
NYU = New York University
Ohio St. = Ohio State University
Pennsylvania = University of Pennsylvania
Portland St. = Portland State University
Princeton = Princeton University
Riverside = University of California, Riverside
Rutgers = Rutgers University

S. Illinois = University of Southern Illinois at Carbondale Southern Cal. = University of Southern California

Stanford = Stanford University

Texas = University of Texas at Austin

UCLA = University of California, Los Angeles

Washington = University of Washington

Wisconsin = University of Wisconsin-Madison

Yale = Yale University

REFERENCES

Allison, Paul D., and J. Scott Long. 1990. "Departmental Effects on Scientific Pro- ductivity." American Sociological Review 55 (August): 469-78.

Anheier, Helmut K., and Jiirgen Gerhards. 1991. "The Acknowledgement of Literary Influence: A Structural Analysis of a German Literary Network." Sociological Forum 6 (1): 137-56.

Appel, Carole S. 1994. "University Press Editing and Publishing." In Editors as Gatekeepers, edited by Rita Simon and James Fyfe. Savage, Md.: Rowman & Littlefield, in press.

Ashenfelter, Orley. 1994. "Report of the Editor." American Economic Review 84 (2): 474-81. Bairam, Erkin I. 1994. "Institutional Affiliation of Contributors to Top Economic Journals." Journal of Economic Literature 32 (June): 674-79. Baughman, James C. 1974. "A Structural Analysis of the Literature of Sociology." Library Quarterly 44 (4): 293-308. Bazerman, Charles. 1988. Shaping Written Knowledge: The Genre and Activity of the Experimental Article in Science. Madison: University of Wisconsin Press.

Blank, Rebecca M. 1991. "The Effects of Double-Blind versus Single-Blind Re- viewing: Experimental Evidence from The American Economic Review." American Economic Review 81 (5): 1041-67.

Broadus, Robert N. 1952. "An Analysis of Literature Cited in the American Sociologi- cal Review." American Sociological Review 17 (3): 355-57. . 1967. "A Citation Study for Sociology." American Sociologist 2 (1): 19-20. Brown, Julia S., and Brian G. Gilmartin. 1969. "Sociology Today: Lacunae, Empha- ses, and Surfeits." American Sociologist 4 (4): 283-91.

Campanario, Juan Miguel. 1993. "Consolation for the Scientist: Sometimes It Is Hard to Publish Papers That Are Later Highly-Cited." Social Studies of Science 23:342-62.

Chubin, Daryl E. 1973. "On the Use of the Science Citation Index in Sociology." American Sociologist 8 (November): 187-91.

Chubin, Daryl E.. and Soumyo D. Moitra. 1975. "Content Analysis of References:
Adjunct or Alternative to Citation Counting?" Social Studies of Science 5:423-41.
Clemens, Elisabeth S. 1986. "Of Asteroids and Dinosaurs: The Role of the Press in

the Shaping of Scientific Debate." Social Studies of Science 16:421-56. Cole, Jonathan R. 1970. "Patterns of Intellectual Influence in Scientific Research." Sociology of Education 43 (Fall): 377-403.

Cole, Jonathan R., and Stephen Cole. 1971. "Measuring the Quality of Sociological Research: Problems in the Use of the Science Citation Index." American Sociologist 6 (1): 23-29. . 1973. Social StratiJication in Science. Chicago: University of Chicago Press.

Cole, Stephen. 1993. "The Editor's View: Why Most of the Articles in Sociology

Journals Are Boring and What Could Be Done about It." Sociological Forum 8 (3): 337-39. Coser, Lewis A., Charles Kadushin, and Walter W. Powell. 1982. Books: The Culture and Commerce of Publishing. New York: Basic.

Cozzens, Susan E. 1985. "Comparing the Sciences: Citation Context Analysis of
Papers from Neuropharmacology and the Sociology of Science." Social Studies of
Science 15:127-53.

Crane, Diana. 1965. "Scientists at Major and Minor Universities: A Study of Produc- tivity and Recognition." American Sociological Review 30 (5): 699-714. . 1972. Invisible Colleges: Diffusion of Knowledge in Scienti5c Communities. Chicago: University of Chicago Press.

Crane, Diana, and Henry Small. 1992. "American Sociology since the Seventies: The Emerging Identity Crisis in the Discipline." Pp. 197-234 in Sociology and Its Publics: The Forms and Fates of Disciplinary Organization, edited by Terence Halliday and Morris Janowitz. Chicago: University of Chicago Press.

Crewe, Ivor, and Pippa Norris. 1991. "British and American Journal Evaluation:
Divergence or Convergence?" PS: Political Science and Politics 24 (3): 524-31.
Darnton, Robert. 1983. "A Survival Strategy for Academic Authors." American

Scholar 52533-37. DiMaggio, Paul. 1987. "Classification in Art." American Sociological Review 52 (August): 440-5 5. England, Paula. 1994. "Editor's Report: American Sociological Review." Footnotes 22 (5): 9-10.

Ennis, James G. 1992. "The Social Organization of Sociological Knowledge: Modeling the Intersection of Specialties." American Sociological Review 57 (April): 259-65. Glenn, N. D., and W. Villemez. 1970. "The Productivity of Sociologists at 45 Ameri- can Universities." American Sociologist 5 (August): 244-252.

Gouldner, Alvin W. 1970. The Coming Crisis of Western Sociology. New York: Basic.

Hamilton, D. P. 1990. "Publishing by-and For?-the Numbers." Science 250 (7 December): 113 1-32. Hankins, F. H. 1937. "Editorial Notes." American Sociological Review 2 (3): 413. Hargens, Lowell L. 1988. "Scholarly Consensus and Journal Rejection Rates." Amer

ican Sociological Review 53 (1): 139-51. . 1991. "Impressions and Misimpressions about Sociology Journals." Contemporary Sociology 20 (May): 343-49. Heinzkill, Richard. 1980. "Characteristics of References in Selected Scholarly English Literary Journals." Librayr Quarterly 50:352-65. Hirsch, E. D., Jr. 1967. Validity in Interpretation. New Haven, Conn.: Yale Univer- sity Press.

Kanter, Rosabeth Moss. 1977. Men and Women of the Corporation. New York: Basic.

Kuhn, Thomas S. (1962) 1970. The Structure of Scientific Revolutions, 2d ed. Chi- cago: University of Chicago Press. Lang, Gladys Engel, and Kurt Lang. 1988. "Recognition and Renown: The Survival of Artistic Reputation." American Journal of Sociology 94 (1): 79-109. Latour, Bruno. 1987. Science in Action: How to Follow Scientists and Engineers through Society. Cambridge, Mass.: Harvard University Press. Lightfield, E. Timothy. 1971. "Output and Recognition of Sociologists." American Sociologist 6 (2): 128-33. Lin, Nan, and Carnot E. Nelson. 1969. "Bibliographic Reference Patterns in Core Sociological Journals, 1965-1966." American Sociologist 4 (1): 47-50. MacRae, D. 1969. "Growth and Decay Curves in Scientific Citations." American Sociological Review 34:63 1-65. Marwell, Gerald. 1992. "Let's Train Reviewers." American Sociological Review 57 (5): iii-iv.

Meltzer, Bernard N. 1949. "The Productivity of Social Scientists." American Journal of Sociology 55 (1): 25-29. Merton, Robert K. (1942) 1973. "The Normative Structure of Science." Pp. 267-78 in The Sociology of Science. Chicago: University of Chicago Press. . (1968) 1973. "The Matthew Effect in Science." In The Sociology of Science. Chicago: University of Chicago Press. Moravcsik, Michael J., and Poovanalingam Murugesan. 1975. "Some Results on the Function and Quality of Citations." Social Studies of Science 5:86-92. Morton, Herbert C., and Anne J. Price. 1989. The ACLS Survey of Scholars. Washington, D.C.: American Council of Learned Societies. Neal, Larry. 1994. "Explorations in Economic History." In Editors as Gatekeepers, edited by Rita Simon and James Fyfe. Savage, Md.: Rowman & Littlefield.

Patterson, Samuel C. 1994. "The Itch to Publish in Political Science." In Editors as Gatekeepers, edited by Rita Simon and James Fyfe. Savage, Md.: Rowman & Littlefield.

Patterson, Samuel C., Michael S. Bailey, Valerie J. Martinez, and Susan C. Angel. 1987. "Report of the Managing Editor of APSR." PS: Political Science and Poli- tics 20 (Fall): 1006-16.

Patterson, Samuel C., and Shannon Smithey. 1990. "Monitoring Scholarly Journal Publication in Political Science: The Role of APSR." PS: Political Science and Politics 23:647-56.

Pertiz, Bluma C. 1983. "Are Methodological Papers More Cited than Theoretical or Empirical Ones? The Case of Sociology." Scientometrics 5 (4): 211-18.

Peters, Douglas P., and Stephen J. Ceci. 1982. "Peer-Review Practices of Psychologi- cal Journals: The Fate of Published Articles, Submitted Again." Behavioral and Brain Sciences 5:187-255.

Pfeffer, Jeffrey. 1993. "Barriers to the Advance of Organizational Science." Academy of Management Review 18 (4): 599-620. Powell, Walter W. 1985. Getting into Print: The Decision Making Process in Schol- arly Publishing. Chicago: University of Chicago Press. Raup, David M. 1986. The Nemesis Affair: A Story of the Death of Dinosaurs and the Ways of Science. New York: Norton. Schudson, Michael. 1989. "How Culture Works: Perspectives from Media Studies on the Efficacy of Symbols." Theory and Society 18:153-80. Schwartz, Barry. 1987. George Washington: The Making of An American Symbol. New York: Free Press.

Simon, Rita. 1994. "An Effective Journal Editor: Insights Gained from Editing ASR." In Editors as Gatekeepers, edited by Rita Simon and James Fyfe. Savage, Md.: Rowman & Littlefield.

Simpson, Ida Harper. 1991. Editorial. Contemporary Sociology 20 (6): v-vi. Small, Albion W. 1895. "The Era of Sociology." American Journal of Sociology 1 (1): 1-15.

Snow, David A,, and Calvin Morrill. 1994. "Linking Ethnography and Theoretical Development: Systematizing the Analytic Process." Manuscript. University of Ar- izona.

Stinchcombe, Arthur L. 1994. "Disintegrated Disciplines and the Future of Sociol- ogy." Sociological Forum 9 (2): 279-91. Sturgis, Richard B., and Frank Clemente. 1973. "The Productivity of Graduates of 50 Sociology Departments." American Sociologist 8 (November): 169-80.

Sullivan, Teresa A. 1994. "Genre in Sociology: The Case for the Monograph." In Editors as Gatekeepers, edited by Rita Simon and James Fyfe. Savage, Md.: Row- man & Littlefield.

Thatcher, Sanford. 1994. "Listbuilding at University Presses." In Editors as Gatekeep- ers, edited by Rita Simon and James Fyfe. Savage, Md.: Rowman & Littlefield.

Thompson, James. 1967. Organizations in Action. New York: McGraw-Hill. Topkis, Gladys S. 1985. "Book Publishing: An Editor's-Eye View." In Scholarly Writing and Publishing, edited by M. Frank Fox. Boulder, Colo.: Westview. Turner, Ralph H. 1960. "Sponsored and Contest Mobility and the School System." American Sociological Review 25:855-67. Whitley, Richard. 1984. The Intellectual and Social Organization of the Sciences. New York: Oxford University Press, Clarendon Press. Wiley, Norbert. 1979. "Recent Journal Sociology: The Substitution of Method for Theory." Contemporary Sociology 8:793-99. Wilson, Everett K. 1979. "Comments from a Servant of the Scattered Family." Contemporary Sociology 82304-8. Wolfe, Alan. 1990. "Books vs. Articles: Two Ways of Publishing Sociology." Sociological Forum 5 (3): 466-89.

Yoels, William C. 1971. "Destiny or Dynasty: Doctoral Origins and Appointment Patterns of Editors of the American Sociological Review." American Sociologist 6 (2): 134-39.

Zelditch, Morris, Jr. 1979. "Why Was the ASR So Atheoretical?" Contemporary Sociology 8:808-13. Zuckerman, Harriet. 1988. "The Sociology of Science." In Handbook of Sociology, edited by Neil Smelser. Beverly Hills, Calif.: Sage.

Zuckerman, Harriet, Jonathan R. Cole, and John T. Bruer, eds. 1992. The Outer Circle: Women in the Scientific Community. New Haven, Conn.: Yale University Press.

Zuckerman, Harriet, and Robert K. Merton. (1971) 1973. "Institutionalized Patterns of Evaluation in Science." In The Sociology of Science. Chicago: University of Chicago Press.

Comments
  • Recommend Us