Abstract
This research investigates changes in scholarly influence by identifying the most-cited scholars and their most-cited works in 20 journals: Five American criminology journals, Five American criminal justice journals, Five international criminology journals, and Five international criminal justice journals. Results obtained in 2005 were compared with previous findings in 2000, 1995, and 1990. Exactly the same methods and journals were used in each year. In 2005, the most-cited scholars were Robert J. Sampson in American criminology journals, American criminal justice journals, and international criminal justice journals, and David P. Farrington in international criminology journals. Overall, Robert J. Sampson was the most-cited scholar in these 20 journals in 2005. He was also the most-cited scholar in these journals in 2000, compared with Lawrence W. Sherman in 1995 and Marvin E. Wolfgang in 1990. The most-cited works of the most-cited scholars included the theories of Sampson and Laub, Gottfredson and Hirschi, and Moffitt, as well as the criminal career paradigm, the effectiveness of correctional treatment, and evidence-based crime prevention. The authors conclude that these analyses reveal changes over time in theoretical concerns and policy issues.
Introduction
The main aim of this study is to investigate scholarly influence by determining the most-cited scholars in 20 Criminology and criminal justice (CCJ) journals of the United States and other major countries of the English-speaking world (Great Britain, Canada, Australia, and New Zealand). The research focused on one year (2005), but comparisons were made with previous analyses of the same 20 journals in 1990 (Cohn, Farrington, & Wright, 1998), 1995 (Cohn & Farrington, 1999), and 2000 (Cohn & Farrington, 2008). Citations in five American criminology journals, five American criminal justice journals, five international criminology journals, and five international criminal justice journals were studied. Identifying the most-cited authors helps identify the most influential scholars and topics during a particular time period, and thus helps document the most important fundamental and policy issues in CCJ in different time periods.
Scholarly influence refers to the importance of a given scholar to the discipline, as well as to the impact that a given work has on subsequently published research in the field. According to Meadows (1974), one of the main assumptions in citation analysis is that citations indicate scholarly influence. Essentially, if multiple researchers working independently on the same problem cite the same material, that material is considered to have scholarly influence. If a scholar’s work is highly cited, it suggests that others in the field find that scholar’s work important and valuable. Impact factors, which are widely used to rank journals and evaluate scholars and institutions (Chew, Villanueva, & Van Der Weyden, 2007), as well as to evaluate promotion applications and allocate research funding (van Driel, Maier, & Maesener, 2007), are based on citations.
It has been known for many years that citation counts are strongly correlated with other measures of scholarly prestige, such as peer ratings of professional eminence, peer ratings of departmental prestige, the receipt of the scholarly prizes including Nobel prizes, election to the presidency of professional associations and elite bodies such as the National Academy of Sciences, numbers of publications, and even earnings (Cohn et al., 1998). The use of citation analysis to measure the prestige and influence of individual scholars has become an accepted technique, and there have been a large number of studies employing citation analysis that were designed to establish the most influential scholars and works in CCJ (see e.g., Cohn et al., 1998; Wolfgang, Figlio, & Thornberry, 1978).
It is important to realize that the distribution of citations is highly skewed; researchers have found that most scholarly articles are never or only rarely cited (Hamilton, 1990; 1991; Laband & Piette, 1994), while only a very small number are highly cited. Additionally, not all citations are of equal importance to the citing scholar; as Chapman (1989, p. 341) stated, “Citation does not necessarily denote approval.” While it is difficult to directly measure the relative impact of the individual works cited by a given researcher, research does suggest that the vast majority of citations are positive or neutral rather than negative (see e.g., Cole, 1975; Cohn & Farrington, 1994a; Garfield, 1979).
Previous Citation Research
This article builds on previous citation research investigating the most-cited scholars in a small number of major American and international CCJ journals. For 1986–1990, the most-cited scholars in three major American criminology journals (Criminology—CRIM, Journal of Quantitative Criminology—JQC, and Journal of Research in Crime and Delinquency—JRCD) and three major criminal justice journals (Justice Quarterly—JQ, Journal of Criminal Justice—JCJ, and Criminal Justice and Behavior—CJB) were determined (Cohn & Farrington, 1994b). The choice of these journals as the major ones in CCJ at that time was discussed and justified in detail. The most-cited scholars were Marvin E. Wolfgang, Michael J. Hindelang, Travis Hirschi, and Alfred Blumstein.
In order to investigate changes in influence over time, this research was repeated for 1991–1995, using the same six journals (Cohn & Farrington, 1998a). The most-cited scholars during the second time period were Travis Hirschi, Michael R. Gottfredson, Robert J. Sampson, and Alfred Blumstein. It was then repeated for 1996–2000 (Cohn & Farrington, 2007a). The most-cited scholars in these six journals during the third time period were Travis Hirschi, Michael R. Gottfredson, David P. Farrington, and Robert J. Sampson. Most recently, this research has been extended to 2001–2005, revealing that the most-cited scholars in these journals during the fourth 5-year time period were David P. Farrington, Robert J. Sampson, Travis Hirschi, and Francis T. Cullen (Cohn, 2011a).
Citations in four international criminology journals were also examined. For 1986–1990, the most-cited scholars in the major criminology journals of the major English-speaking countries (CRIM, Australian and New Zealand Journal of Criminology—ANZ, British Journal of Criminology—BJC, and Canadian Journal of Criminology and Criminal Justice—CJC) were determined (Cohn & Farrington, 1994a). The most-cited scholars were Marvin E. Wolfgang, Alfred Blumstein, David P. Farrington, James Q. Wilson, and Stanley Cohen.
In order to investigate changes in influence over time, this research was repeated for 1991–1995, using the same four journals (Cohn & Farrington, 1998b). The most-cited scholars during the second time period were Travis Hirschi, David P. Farrington, Michael R. Gottfredson, Alfred Blumstein, and John Braithwaite. It was then repeated for 1996–2000 (Cohn & Farrington, 2007b). The most-cited scholars in these four journals during the third time period were John Braithwaite, David Garland, David P. Farrington, Richard V. Ericson, and Ken Pease. Most recently, this research has been extended to 2001–2005, revealing that the most-cited scholars in these journals during the fourth 5-year time period were David P. Farrington, Robert J. Sampson, Travis Hirschi, Michael R. Gottfredson, and Lawrence W. Sherman (Cohn, 2011b).
Increasing the Number of Journals
Despite the acknowledged prestige of the six American and three international journals that were examined, the previous analyses are vulnerable to the criticism that only a small number of mainstream CCJ journals were covered. Arguably, this creates a bias against scholars who are working in more specialized areas that are less well represented in mainstream journals. There are at least 326 journals within CCJ, most of which are very specialized (Vaughn, del Carmen, Perfecto, & Charand, 2004). It could also be argued that the most-cited scholars are, at least to some extent, specific to particular journals, and that the results would differ if other journals were analyzed. One method of investigating these concerns is to study the most-cited scholars in specific sub-areas of CCJ, such as critical criminology (Wright & Friedrichs, 1998), law enforcement (Wright & Miller, 1998), or women and crime (Wright & Sheridan, 1997).
Another alternative is to expand the number of CCJ journals that are analyzed. One way to achieve this is to use the Web of Science databases, which includes the Social Sciences Citation Index (SSCI). SSCI was used previously to identify the most-cited CCJ works (Cohn & Farrington, 1996), but it has some severe disadvantages for identifying the most-cited CCJ scholars. First, SSCI includes citations to many journals that are not directly or centrally concerned with CCJ. Second, the list of journals included in SSCI is not stable, making longitudinal research extremely difficult or impossible. Third, SSCI includes self-citations, which should be excluded in assessing the influence of one scholar’s work on another. Self-citations have been excluded in all the present analyses of journals. Fourth, errors in journal reference lists, such as misspelling of names or incorrect initials, are carried over into SSCI. Fifth, authors are identified only by initials, not by first names, creating confusion if two or more scholars share the same initials. In the present searches of CCJ journals, a great deal of time was spent distinguishing between authors with the same surname and first initial (e.g., Jacqueline, Jacob, and Joseph Cohen), and in resolving errors in citations (e.g., whether a cited R. A. Sampson was really Robert J. Sampson or a different person) by consulting original sources. This sort of extensive checking would take a prohibitively long time with the SSCI.
Similar problems apply to other electronic sources of citation data, such as Scopus and GoogleScholar. Scopus was released by Elsevier in 2004, and its citation backfile includes materials published from 1996 onward (Dess, 2006). It has good coverage of the “hard” (e.g., physical or medical) sciences but much more limited coverage of the social sciences (Fingerman, 2006). As of 2009, the Web of Science had about four times as many social science records as Scopus (Jascó, 2009c).
Unlike Scopus and the Web of Science, GoogleScholar is free. It is an online scientific archive that began in 2004 (Tenopir, 2007). It includes citation data from journals, books, theses, technical reports, and selected “scholarly” web pages, and tends to produce more citations than the Web of Science (Bauer & Bakkalbasi, 2005; Meho & Yang, 2007). However, there are many problems with GoogleScholar and its software for searching citations (see e.g. Jascó, 2008a, 2008b, 2009a, 2009b). A particular worry is that Google Inc. provides no information about the database coverage; users do not know which materials are indexed, which journals are included, which publishers are included, which time periods are covered, or how often GoogleScholar is updated. And, of course, all these electronic sources of citation data include self-citations. It is clear that no existing electronic source of citation data could provide reliable information about citation trends in CCJ over a 20-year period.
The strategy adopted was to expand the number of CCJ journals to include some that were slightly less mainstream. Nine journals were studied previously: CRIM, JQC, JRCD, JQ, JCJ, CJB, BJC, CJC, and ANZ. In addition, the most-cited scholars and works in the serial publication Crime and Justice (CAJ) were studied for 1979–1993 (Cohn & Farrington, 1996). The selection of these ten publications as the major CCJ journals in the English-speaking world was justified (Cohn & Farrington, 1994b, 1998c) on the basis of prior research on the rated prestige of journals (Regoli, Poole, & Miracle, 1982; Shichor, O’Brien, & Decker, 1981; Sorensen, Patterson, & Widmayer, 1992; Williams, McShane, & Wagoner, 1995). One prestigious CCJ journal, Journal of Criminal Law and Criminology, could not be analyzed because the legal style of footnotes used at the time omitted the initials of authors of cited articles and thus made it virtually impossible in practice to distinguish between different authors with the same surnames. While today the journal does include first names or initials, the continued use of footnotes rather than a reference list still makes it difficult to easily access the information needed for citation analysis.
Cohn et al. (1998) doubled the number of CCJ journals covered to 20, by adding ten more CCJ journals. Because of the number of citations and the amount of work involved, the 20 journals were only analyzed for 1 year—1990. Their aim was to compare American and international journals, and criminology and criminal justice journals, and to analyze five journals in each of the four possible categories (American criminology, American criminal justice, International criminology, International criminal justice). Achieving this planned design, however, was not easy.
First, the following eight CCJ journals were added: Crime and Delinquency (CAD), Criminal Justice Review (CJR), Federal Probation (FP), Criminologie (CRGE), Crime, Law, and Social Change (CLSC; formerly Contemporary Crises), International Journal of Comparative and Applied Criminal Justice (IJCA), International Journal of Offender Therapy and Comparative Criminology (IJOT), and Social Justice (SJ). Journal of Crime and Justice (JCAJ) and Howard Journal of Criminal Justice (HJCJ) were considered but not included, because they contained too few citations; the most-cited scholar in JCAJ in 1990 had only four citations, and only three scholars had four or more citations in HJCJ in 1990. As there were too few American criminology journals, two more specialized journals focusing on violence were added: Journal of Interpersonal Violence (JIV) and Violence and Victims (VAV).
The analysis of 20 journals in 1990 largely confirmed and amplified the previous analyses of nine major journals in 1986–1990. Marvin E. Wolfgang, Alfred Blumstein, and James Q. Wilson were still among the most-cited authors. Generally, the most-cited works of the most-cited authors (e.g. Blumstein, Cohen, Roth, & Visher, 1986; Wolfgang, Figlio, & Sellin, 1972) were concerned with criminal career research. The expanded number of journals, however, revealed other authors and other issues that were also influential. In particular, Francis T. Cullen’s work on rehabilitation (Cullen & Gilbert, 1982) and Ronald V.G. Clarke’s work on rational choice theory and situational crime prevention (Cornish & Clarke, 1986) were relatively more cited in the 20-journal analysis.
Cohn and Farrington (1999) then repeated their analysis of 20 journals for the year 1995. The most-cited scholars in 1995 were Lawrence W. Sherman, Travis Hirschi, Michael R. Gottfredson, David P. Farrington, and Robert J. Sampson. Both Sherman and Sampson became increasingly highly cited between 1990 and 1995. The most-cited works of the most-cited authors were Policing Domestic Violence (Sherman, 1992), A General Theory of Crime (Gottfredson & Hirschi, 1990), Understanding and Controlling Crime (Farrington, Ohlin, & Wilson, 1986), and Crime in the Making (Sampson & Laub, 1993).
Cohn and Farrington (2008) repeated the analysis again for the year 2000. The most-cited scholars in 2000 were Robert J. Sampson, David P. Farrington, Francis T. Cullen, Travis Hirschi, and Terrie E. Moffitt. The most-cited works of the most-cited authors were Crime in the Making (Sampson & Laub, 1993), “The development of offending and antisocial behavior from childhood” (Farrington, 1995), “Does correctional treatment work?” (Andrews et al., 1990), A General Theory of Crime (Gottfredson & Hirschi, 1990), and “Adolescence-limited and life-course-persistent antisocial behavior” (Moffitt, 1993). In contrast to previous years, these most-cited works included a number of articles as well as books.
The Present Research
The present research repeated and extended the previous analyses of Cohn et al. (1998), Cohn and Farrington (1999), and Cohn and Farrington (2008) by analyzing the most-cited scholars in the same 20 journals in 2005 (see Table 1 for a list of the journals). These were compared not only with the most-cited scholars in 20 journals in 1990, 1995, and 2000 but also with the most-cited scholars in nine journals in 2001–2005. Exactly the same methods were used as in previous analyses; the previous publications should be consulted for detailed comparative data.
Articles, Authors, and Cited Scholars in 2005
Note. % U.S. shows the percent of scholars located in the United States.
As in the earlier research, the reference lists of all articles were analyzed. All scholars cited in all articles in these 20 journals in 2005 were counted. “Articles” included research notes, comments, and rejoinders, but excluded book reviews, editorials, introductions, letters, and obituaries. Every cited author in every eligible article was counted (not just first authors), with the exception of institutional authors (e.g., Federal Bureau of Investigation, Bureau of Justice Statistics, New York Times). It was not practical to restrict the count only to published books and articles; unpublished reports and conference papers were included if they were cited. All self-citations were excluded, because the aim of the research was to study the influence of one scholar upon other scholars.
For each journal, the reference pages were either downloaded from online copies of journals or entered into a computer from printed copies using an optical scanner. The information was edited to correct scanner-induced typographical errors, to eliminate institutional authors and self-citations, and to create duplicate listings for references with multiple authors (with each coauthor listed first). Extensive checking was carried out to ensure that no references were omitted, to minimize the possibility of typographical errors, and to detect mistakes in reference lists. When the references for each journal had been entered into the computer file, they were sorted into alphabetical order by author and examined to determine the number of times that each name occurred. It is of course possible for a scholar’s surname to change over the course of a career; this most commonly occurs among female scholars who marry after beginning a publication career. Therefore, citations to scholars with multiple names (e.g., Kimberly Kempf/Leonard) were amalgamated, where these were known.
Where references did not include first names or middle initials, it was necessary to spend a considerable amount of time checking them against the original publications to distinguish between, for example, the various D. Smiths (Douglas A., David D., David E., David J., Dwight, etc.) and J. Cohens (Jacqueline, Joseph, Jacob, etc.). It was also necessary to distinguish between different people with the same name (e.g. David Brown, Richard Sparks, Patrick O’Malley, Richard Wright). Where references did not list all the authors of a work, but merely specified “et al.,” the original publication was consulted and the missing names were inserted, so that every author of a work was counted. A good deal of time was also spent checking and correcting errors in the original reference lists, which were depressingly common. Because of the frequency of errors, a detailed knowledge of CCJ scholars and works was essential in maximizing the accuracy of the data; this was not merely a clerical task by any means. It is doubtful that every error in the reference lists was detected, but the vast majority of them were located and corrected, especially those involving the most-cited scholars in each journal.
One advantage of citation analysis, apart from the fact that it is quantitative and objective, is that the raw data are readily available to anyone who would like to try to replicate the present analyses. Researchers should be warned, however, that the present results required the analysis of 53,263 cited scholars in 644 separate articles in the 20 journals in 2005 (compared with 41,742 cited scholars in 600 articles in 2000, 27,536 scholars in 539 articles in 1995, and 23,697 scholars in 503 articles in 1990). These were not all different persons; the same person could be counted more than once if he or she was cited several times. The number of cited authors more than doubled between 1990 and 2005. Part of this increase might be attributable to an increase in the average number of authors per article (Tewksbury & Mustaine, 2011).
Cohn and Farrington (1996) argued that concepts developed in criminal career research could be used to enrich citation analysis. In particular, they distinguished between the prevalence and frequency of citations. One problem with using the total number of citations as a measure of influence is that a large number of citations may be obtained if a scholar is cited a few times in many different articles (a high prevalence) or cited many times in a few articles (a high frequency per article). A high prevalence of citations may be a better measure of influence on a large number of other scholars than a high frequency (which may reflect a great influence on only a small number of other scholars).
A distinction may also be drawn between specialization and versatility. Some highly cited authors may be specialized because they have one or two highly cited works, often books and often presenting a major theory. Other highly cited authors may be versatile because they have many different works cited, with no single work standing out for its large number of citations. For example, in Cohn and Farrington’s (2008) study of the most cited scholars in 20 journals in 2000, Travis Hirschi was found to be very specialized because his two most-cited works (Gottfredson and Hirschi, 1990; Hirschi, 1969) accounted for 64% of all his citations. Conversely, David P. Farrington was found to be extremely versatile because his two most-cited works (Farrington, 1995; Loeber & Farrington, 1998) accounted for only 10% of all his citations. Versatile authors tend to have written many articles rather than a single seminal book, although some scholars with one or two highly cited works may also have many different works cited. Logically, a high frequency of citations per article has to be associated with versatility because, for a scholar to be highly cited in a small number of articles, the scholar must have many different works cited (as, obviously, any individual work only appears once per article). A high prevalence may be associated with specialization, if the same work or small number of works are cited in many different articles. However, a high prevalence may also be associated with versatility, if those many citations were to a wide variety of works by the scholar. This would also indicate that the scholar's works were influencing a large number of other researchers.
A number of other recent studies have also applied criminal career concepts and dimensions to the study of the publication productivity and citation frequency of CCJ scholars. For example, Rice, Cohn, and Farrington (2005) used criminal career concepts when studying divergent pathways of criminal justice scholars, pointing out that criminal career dimensions have been found to be useful for visualizing lives (Maltz & Mullany, 2000). They equated the dimensions of seriousness and specialization to a scholar’s publication output and logical consistency within a research agenda. They also applied career concepts such as frequency, specialization, seriousness, and co-offending (multiple authorship) to the curriculum vitae of the scholars under study. Rice, Terry, Miller, and Ackerman (2007) also applied criminal career concepts to their study of female CCJ scholars, and their attempts to identify gender differences in scholarship patterns. Jennings, Gibson, Ward, and Beaver (2008) also compared publishing trajectories to criminal career trajectories, focusing on the dimensions of frequency (operalizationed as publication rates) and seriousness (journal quality).
Citations in 20 Journals
Table 1 summarizes key statistics for the 20 journals, including the number of articles published in 2005, the number of authors of these articles, the percentage of these authors who were located in the United States, and the total number of eligible cited scholars in the journal (excluding self-citations and institutional authors). In the five American criminology journals, there were a total of 235 articles, with 620 authors (86% of whom were American) and a total of 21,105 cited scholars (average 90 per article). The five American criminal justice journals contained a total of 123 articles written by 275 authors (98% of whom were American) and a total of 10,448 cited scholars (average 85 per article). In the five international criminology journals, there were 153 articles, with 255 authors (11% American) and a total of 11,129 cited authors (average 73 per article). Finally, in the five international criminal justice journals, there were 133 articles written by 291 authors (62% American) and a total of 10,581 cited scholars (average 80 per article). Overall, this research examined 644 articles in these 20 journals (an average of 32 articles per journal) with 1,441 authors (64% American) and 53,263 citations (an average of 2,663 cited scholars per journal, or 83 per article).
Questions might be raised about the classification of journals as international versus American. This classification was determined using data from 1990, and the aim was to repeat exactly the same analyses in 1995, 2000, and 2005, so that trends over time could be established. In regard to the international criminology journals, four (ANZ, BJC, CJC, and CRGE) were published outside the United States and contained very few American authors. The fifth, CLSC, was originally subtitled “An International Journal,” is published in The Netherlands, contained only 18% American authors in 2005, and included fourteen non-Americans out of 27 editors, senior editors, and associate editors in 2005.
In regard to the international criminal justice journals, IJCA is the official journal of the American Society of Criminology’s Division of International Criminology and is explicitly international in its focus. IJOT, “a journal of international cooperation,” is also explicitly international in its focus, and included 29 non-Americans out of 56 consulting and associate editors in 2005. SJ, a project of “Global Options,” included an international editorial advisory board drawn from 13 countries in addition to the United States in 2005. CJB, the official publication of the International Association for Correctional and Forensic Psychology, is subtitled “An International Journal,” and the journal information explicitly states that “articles. . . are welcomed from throughout the world.” CAJ, according to its promotional material, “since 1979 has presented a review of the latest international research.” Its editor, Michael H. Tonry, was located in the United Kingdom in 1999–2004. Its content varies considerably from year to year. In contrast to most journals, most articles in CAJ (state-of-the-art literature reviews) are solicited or commissioned rather than unsolicited. These essays are, however, rigorously peer reviewed, and the present analysis is of citations, not publications.
Table 2 shows the five most-cited scholars in each of the 20 journals in 2005. In CRIM, for example, the most-cited scholar was Robert J. Sampson, with 66 citations, followed by Terrie E. Moffitt (51), John H. Laub (42), Raymond Paternoster (33), and Alex R. Piquero (33). Robert J. Sampson was the most-cited scholar in 6 of the 20 journals, while Murray A. Straus and Lawrence W. Sherman were each most-cited in two journals. Robert J. Sampson was among the five most-cited scholars in eight of these journals, while Don A. Andrews, Francis T. Cullen, James L. Bonta, John H. Laub, and Terrie E. Moffitt each appeared three times in Table 2. The most-cited scholars in each journal reflect the different interests of the authors of each journal; for example the most-cited scholars in SJ were rarely cited in CRIM, and vice versa.
Five Most-Cited Scholars in Each Journal in 2005
Note. For abbreviations of journals, see Table 1. The most-cited five scholars are shown in general, but there can be more than five in cases of ties.
It is clear from Table 2 that even the most-cited scholars in a journal or group of journals account for a fairly small fraction of the total number of citations within that journal or group of journals. For example, Robert J. Sampson, the most highly cited scholar in CRIM, accounted for approximately 1.5% of all citations in the journal in 2005. However, as noted above, citations are highly skewed (Hamilton, 1990, 1991; Laband & Piette, 1994) and most scholars and works are never cited, or are cited only once or twice. There is a clear difference between being cited once and being cited 66 times, as Sampson was in CRIM. In fact, Sampson was cited in 56% of all the articles in CRIM in 2005 (20 of the 36 articles; he was a coauthor of one other article and his self-citations in that were not counted). Therefore, while the most-cited scholar in a journal may account for a relatively small fraction of citations overall, a large number of citations is nevertheless an objective indicator of influence.
Most-Cited Scholars
In order to compare the citations of scholars in each of the 20 journals, the most-cited scholars in each journal were ranked and given a score of 26 minus their ranking. Generally, this meant that those scholars who were ranked from 1 to 25 in each journal (according to their citations) received scores ranging from 25 to 1 (respectively). However, more (or fewer) than 25 scholars could score in a given journal if there were ties. For example, in CJC, Lawrence W. Sherman, the most-cited scholar, was given a score of 25 and the next-most cited scholar, Philip C. Stenning, was given a score of 24. The next two scholars, Doris L. MacKenzie and Brandon C. Welsh, who were each ranked 3.5 because they were tied, were each given a score of 22.5. The ranking continued up to the nine scholars with six citations, who were each ranked 20. A total of 24 scholars were given a score in CJC; all other scholars cited in this journal were scored 0. This procedure was originally developed as a way of equally weighting all the journals (Cohn & Farrington, 1994b). If the raw numbers of citations of each scholar in each journal had simply been added together, journals with a large number of citations (e.g., CRIM) would have contributed disproportionately to the total score.
Table 3 shows the ten most-cited scholars in each group of five journals in 2005, and their comparative rankings in 2000, 1995, and 1990. The total score in 2005, which is computed by summing a scholar’s scores in each set of five journals, is shown in the right-hand column. Robert J. Sampson was the most-cited scholar in American criminology journals in 2005 and 2000, compared with Travis Hirschi in 1995, and Marvin E. Wolfgang in 1990. Sampson was also the most-cited in American criminal justice journals in 2005, compared with Francis T. Cullen in 2000, Lawrence W. Sherman in 1995, and Joan Petersilia in 1990. David P. Farrington was the most-cited in international criminology journals in 2005, compared with John Braithwaite in 2000, 1995, and 1990. Robert J. Sampson was the most-cited scholar in international criminal justice journals in 2005, compared with Robert D. Hare in 2000, Steven F. Messner in 1995, and Travis Hirschi in 1990.
Most-Cited Scholars in Groups of Five Journals in 2005
Note. More than 10 scholars are shown where there are ties. Ranks upto 30 in prior years are shown.
In American criminology journals, 6 of the top 11 scholars were highly cited (among the top 25) in CRIM, JQC, and JRCD; a seventh (Daniel S. Nagin) was highly cited in CRIM and JQC but not in JRCD. The other four (Murray A. Straus, Jacqueline C. Campbell, Kent D. O’Leary, and Richard J. Gelles) were highly cited in JIV and VAV. This shows the effect of including these two more specialized journals. In American criminal justice journals, 4 of the 10 most-cited scholars were highly cited in 3 of the 5 journals. In international criminal justice journals, the top four scholars were highly cited in three of the five journals and the other six were highly cited in only two journals. The international criminology journals showed the least agreement; David P. Farrington was highly cited in four journals, Clifford D. Shearing was highly cited in three journals, five other scholars were highly cited in two journals, and three (Philip C. Stenning, Terrie E. Moffitt, and Marc LeBlanc) were highly cited in only one journal.
In American criminology journals, 10 of the top 11 scholars in 2005 had also been in the top 14 in 2000, 6 had been in the top 15 in 1995, and 6 had been in the top 11 in 1990. All of the top 6 scholars in 2000 and four of the top 5 scholars in 1995 were still in the top 10 in 2005. Five scholars were among the top 11 in all 4 years. However, two of the top five scholars in 1990 (Marvin E. Wolfgang and Jacqueline Cohen) were not among the 30 most-cited scholars in 2005. In American criminal justice journals, only 2 of the top 10 scholars in 2005 had been in the top 30 in 2000, 2 had been in the top 30 in 1995, and only 1 had been in the top 30 in 1990. Francis T. Cullen was among the top 14 scholars in all 4 years. The highest-ranked scholars in 2000 who were not in the top 30 in 2005 were Lawrence W. Sherman (3), Robert Agnew (5), and Travis Hirschi (6).
In international criminology journals, 4 of the top 10 scholars in 2005 had been in the top 30 in 2000 (3 in the top 10), 3 had been in the top 30 in 1995, and only 2 had been in the top 30 in 1990. John Braithwaite was among the top 5 scholars in all 4 years. The highest-ranked scholars in 2000 who were not in the top 30 in 2005 were Michel Foucault (2) and Richard V. Ericson (3). In international criminal justice journals, 4 of the top 10 scholars in 2005 had been in the top 30 in 2000, 2 had been in the top 30 in 1995, and 2 had been in the top 30 in 1990. Robert D. Hare was among the top 19 scholars in all 4 years.
Table 4 shows the 10 most-cited scholars in each group of 10 journals (Criminology journals, Criminal justice journals, American journals, and International journals) in 2005, and their comparative rankings (up to 30) in 2000, 1995, and 1990. The total score in 2005 is shown in the right-hand column. Robert J. Sampson was most-cited in American journals, as he was in 2000, compared with Travis Hirschi in 1995, and Marvin E. Wolfgang in 1990. Sampson was also most-cited in criminal justice journals in 2005, compared with Francis T. Cullen in 2000, Lawrence W. Sherman in 1995, and Travis Hirschi in 1990. David P. Farrington was most-cited in international journals in 2005, compared with John Braithwaite in 2000, Lawrence W. Sherman in 1995, and Francis T. Cullen in 1990. Farrington was also most-cited in criminology journals in 2005, compared with Robert J. Sampson in 2000, Travis Hirschi in 1995, and Marvin E. Wolfgang in 1990.
Ten Most-Cited Scholars in Groups of 10 Journals in 2005
Note. Ranks upto 30 in prior years are shown.
In American journals, 9 of the top 10 scholars in 2005 had been in the top 30 in 2000, 6 were in the top 30 in 1995, and 5 were in the top 30 in 1990. The highest ranked scholars in 2000 who were not in the top 30 in 2005 were Lawrence W. Sherman (5) and Robert Agnew (7). In international journals, 8 of the top 10 scholars in 2005 had been in the top 30 in 2000, compared with three in 1995 and three in 1990. The highest-ranked scholars in 2000 who were not in the top 30 in 2005 were Michel Foucault (2) and Richard V. Ericson (4).
In criminology journals, 9 of the top 11 scholars in 2005 had been in the top 30 in 2000, compared with 7 in 1995 and 5 in 1990. The highest-ranked scholars in 2000 who were not in the top 30 in 2005 were Michel Foucault (5) and Richard V. Ericson (6). In criminal justice journals, 5 of the top 10 scholars in 2005 had been in the top 30 in 2000, compared with four in 1995 and three in 1990. The highest-ranked scholars in 2000 who were not in the top 30 in 2005 were Lawrence W. Sherman (2) and Terrie E. Moffitt (6.5).
Table 5 shows the forty most-cited scholars in all 20 journals in 2005 and their comparative rankings (upto 40) in 2000, 1995, and 1990. The most-cited scholar in 2005, as in 2000, was Robert J. Sampson, compared with Lawrence W. Sherman in 1995 and Marvin E. Wolfgang in 1990. Sampson advanced from 38 in 1990 to 5 in 1995 and first in 2000. The comparison of citation rankings over the years shows the advance of a new generation of scholars such as Sampson and John H. Laub (and even younger scholars such as Alex R. Piquero), and the corresponding decline of older (and deceased) scholars such as Wolfgang and Michael J. Hindelang. David P. Farrington is notable for the unusual stability of his rankings over the 4-time periods (ranked third in 1990, fourth in 1995, second in 2000, and third again in 2005). Travis Hirschi is also fairly stable, although there appears to be a downward trend (he was ranked second in 1990 and 1995, fourth in 2000, and sixth in 2005). Nine of the top 10 scholars in 2005 had been in the top 40 in 2000; the new entry was Daniel S. Nagin. Conversely, 8 of the top 10 scholars in 2000 were still in the top 40 in 2005. The highest ranked scholars in 2000 who were not in the top 40 in 2005 were Michel Foucault (8) and David Garland (9).
Forty Most-Cited Scholars in all 20 Journals in 2005
Note. Ranks upto 40 in prior years are shown. CRIM = 10 criminology journals. CJ = 10 criminal justice journals.
Five of the top 10 most-cited scholars in 2005 (David P. Farrington, Travis Hirschi, Terrie E. Moffitt, Daniel S. Nagin, and Lawrence W. Sherman) were more highly cited in criminology journals than in criminal justice journals. Four (Robert J. Sampson, Francis T. Cullen, James L. Bonta, and Don A. Andrews) were more highly cited in criminal justice journals. John H. Laub was about equally cited in both categories.
As already indicated, citation counts are correlated with other indicators of scholarly prestige. For example, many of the scholars listed in Table 5 have been honored by the American Society of Criminology. Robert J. Bursik, Francis T. Cullen, David P. Farrington, Travis Hirschi, John H. Laub, Joan Petersilia, and Lawrence W. Sherman are past Presidents of the Society, and Robert J. Sampson has been elected President for 2011–2012. Twenty of the most-cited scholars are Fellows of the ASC and 15 have received awards from the Society. John Braithwaite, David P. Farrington, David Garland, and Jock Young have all received the Sellin-Glueck Award for international contributions to criminology, while Robert J. Sampson, John H. Laub, Terrie E. Moffitt, and John Braithwaite have all received the Stockholm Prize.
Table 6 compares results obtained with 20 journals in 1 year and with 9 journals in 5 years, showing that there was a considerable overlap of the most-cited scholars. Twelve of the 20 most-cited scholars in 20 journals in 2005 were among the 20 most-cited scholars in nine journals in 2001–2005. However, it was noticeable that six international scholars on the 20-journal list (Don A. Andrews, James L. Bonta, John Braithwaite, Paul Gendreau, Robert D. Hare, and Clifford D. Shearing) were missing from the top 20 on the nine-journal list, most likely because the nine-journal list gave less weight to international journals. Sixteen of the top 20 scholars on the 20-journal list (all except David Garland, Paul Gendreau, David Finkelhor, and Clifford D. Shearing) were in the top 50 on the nine-journal list, an impressive level of agreement. Nevertheless, it is clear that international scholars such as John Braithwaite, and scholars in less mainstream areas such as David Finkelhor, tend to be given higher rankings in the 20-journal analysis than in the nine-journal analysis.
Comparison of 20 Journals in 2005 and 9 Journals in 2001–2005
Note. The nine journals were: CRIM, JQC, JRCD, JQ, JCJ, CJB, BJC, CJC, and ANZ.
Further Analyses
Table 7 shows the two most-cited works of the 10 most-cited scholars in all 20 journals in 2005. The most-cited work of the most-cited scholar, Robert J. Sampson, was Crime in the making (Sampson & Laub, 1993) with 30 citations. The number of citations of this work differed from those of John H. Laub because of the exclusion of self-citations. As in prior years, the majority of the most-cited works of the 10 most-cited scholars in 2005 were books rather than journal articles. Five works from the 2000 table of the most-cited works of the most-cited scholars appear in the 2005 table: the most-cited works of Sampson, Laub, Cullen, Hirschi, and Moffitt.
Most-Cited Works of the Most-Cited Scholars in 2005
Note. *Including earlier editions.
Table 8 shows the total number of citations of the 10 most-cited scholars in 2005, together with the number of different articles in which they were cited (prevalence), and the average number of citations per article (frequency). Because of the scoring system which gave equal weight to each journal, some authors with high scores (e.g. James L. Bonta and Don A. Andrews) had fewer citations than others; they were highly cited in journals that contained relatively few citations.
Prevalence and Frequency in 2005
Note. This shows the number of different articles in the 20 journals (of a total of 644) in which the author was cited.
As an example, John H. Laub had 147 total citations and was cited in 67 different articles. His average number of citations per article was therefore 2.2. As mentioned above, a high frequency of citations per article, or being cited many times in a few articles, may be a poorer measure of influence on a large number of other scholars than a high prevalence of citations. James L. Bonta and Don A. Andrews had a relatively high frequency of citations combined with a relatively low prevalence. Counting the total number of citations may overestimate the influence of these scholars and underestimate the influence of scholars with a high prevalence and a low frequency, such as Francis T. Cullen and Travis Hirschi. The total number of citations, however, is the most widely used measure. Robert J. Sampson had both a high prevalence and a high frequency.
The total number of citations can also be disaggregated into the number of different works cited and the average number of citations per work. Some scholars (defined above as versatile) may be highly cited primarily because they have a large number of different works cited, while others (defined above as specialized) may be highly cited because a small number of works are each cited many times. Table 9 shows three measures of specialization or versatility: the number of different works cited, the average number of citations per work, and the percentage of total citations accounted for by the two most-cited works. For example, Robert J. Sampson had a total of 239 citations of 58 different works, with an average of 4.1 citations per work. His most-cited work, Crime in the making (Sampson & Laub, 1993) was cited thirty times and therefore accounted for 13% of his citations. His next most-cited work was “Neighborhoods and violent crime” (Sampson, Raudenbush, & Earls, 1997), with 20 citations. His two most-cited works therefore accounted for 21% of all his citations (50 of the 239).
Specialization and Versatility in 2005
Note. This shows the number of different works by an author that were cited in the 20 journals, plus the percentage of total cites that were accounted for by the two most-cited works of that author.
On the criterion of the largest number of different works cited, David P. Farrington (86), Terrie E. Moffitt (61), Francis T. Cullen (60), and Robert J. Sampson (58) were the most versatile, whereas Travis Hirschi (21), James L. Bonta (26), Don A. Andrews (35), and John H. Laub (36) were the most specialized. On the criterion of the average number of citations per work, Travis Hirschi (6.0), Robert J. Sampson (4.1), and John H. Laub (4.1) were the most specialized. Those with few citations per work were not necessarily versatile, because versatility depends also on the number of different works cited. Scholars with a low frequency of citation per work and a large number of different works cited (e.g., David P. Farrington, Francis T. Cullen, Terrie E. Moffitt) were the most versatile.
Travis Hirschi was the most specialized in that two of his works accounted for a large fraction of all his citations. His two most-cited works, A General Theory of Crime (Gottfredson & Hirschi, 1990) and Causes of Delinquency (Hirschi, 1969), accounted for 54% of all his citations. On this criterion, James L. Bonta (45%), and Don A. Andrews (42%) were also specialized, whereas David P. Farrington (10%), Daniel S. Nagin (15%) and Lawrence W. Sherman (19%) were the most versatile.
Policy implications
The results of this study may have important public policy implications. An examination of the most-cited scholars and works indicates those topics that criminological researchers consider to be of most importance in different time periods. The most-cited works may also reflect current policy concerns, and an awareness of the topics that criminologists consider to be important and influential should guide policy makers and legislators in the development of new public policies.
The most-cited works of five of the six most-cited scholars in 1990 focused on criminal career research (Cohn et al., 1998). This research should inform criminal justice decision making because it can provide useful information about the likely future course of criminal careers. This research greatly impressed James K. Stewart, the director of the National Institute of Justice in the 1980s. As a result, he committed a large amount of the Institute’s funding to research in this area, which led to considerable controversy among the criminological community. Gottfredson and Hirschi (1986) strongly criticized this policy, arguing that the focus of criminological research on criminal careers and “career criminals” was unjustified, while Blumstein, Cohen, and Farrington (1988) defended the value of the criminal career approach in criminology.
The other most-cited scholar in 1990 was Travis Hirschi, and his most-cited work was the theoretical book Causes of delinquency (Hirschi, 1969). According to Laub (2004, p.18), “successful theories. . . provide influential guides to public policy”. The main policy implication of this theory is that efforts should be made to increase individuals’ bonding to society. The seventh most-cited scholar was Francis T. Cullen, and his most-cited work was Reaffirming Rehabilitation (Cullen & Gilbert, 1982), which has clear implications for effective correctional treatment. Along with a number of other researchers, Cullen’s works helped to overcome the belief, fostered by Robert Martinson’s “nothing works” doctrine (Lipton, Martinson, & Wilks, 1975), that “rehabilitation is dead” (Cullen, 2005).
The eighth most-cited scholar was Ronald V. G. Clarke, and his most-cited work was The Reasoning Criminal (Cornish & Clarke, 1986). This book propounded a rational choice theory of offending that has important implications for situational crime prevention. This idea had enormous influence on the British Home Office in the 1980s and 1990s; most government crime prevention initiatives during this time period focused on situational crime prevention.
The most-cited works of the most-cited scholars in 1995 also had clear policy implications (Cohn & Farrington, 1999). Policing Domestic Violence (Sherman, 1992) was the most-cited work of the most-cited scholar. This book described the Minneapolis Domestic Violence Experiment and several attempts to replicate this experiment. The results of the Minneapolis experiment greatly impressed the then Assistant Attorney General Lois Herrington, leading her to encourage police officers to arrest perpetrators whenever possible in cases of domestic assault. A 1986 survey of big-city police departments found that 30% stated that their policies or practices on domestic violence had changed as a result of the Minneapolis Experiment (Sherman & Cohn, 1989).
The most-cited work of the next two scholars was A General Theory of Crime (Gottfredson & Hirschi, 1990), which proposed that offending depended on self-control. A clear implication of this theory is that policy makers should develop early intervention programs to improve juveniles’ self-control at a young age, perhaps focusing on parent training (Piquero, Jennings, & Farrington, 2010). The next most-cited work, Understanding and Controlling Crime (Farrington, Ohlin, & Wilson, 1986) set out methods of advancing knowledge about the development of criminal careers and how this knowledge might be used to reduce offending.
Similar policy implications can be drawn from the most-cited works of the most-cited scholars in 2000. The most-cited work of the most-cited scholar was Crime in the Making (Sampson & Laub, 1993), which reported on the development of offending and proposed that the most important theoretical construct was informal social control. The main implication of this theory is that bonding to the family, the school and the community should be increased, through programs such as those providing job training and structured routine activities in adulthood. It also suggests that desistance can be encouraged by fostering bonding to adult institutions such as employment and marriage. Another suggestion is that informal social control in communities could be improved by increasing community cohesiveness or “collective efficacy” (Sampson, Raudenbush, & Earls., 1997). The theory also suggests that it is important to minimize labelling or stigmatization of offenders by reducing the use of incarceration.
Some of the other most-cited scholars in 2000 were the same as in earlier years: David P. Farrington for the development of offending, Francis T. Cullen for the effectiveness of correctional treatment, Travis Hirschi for the general theory of crime, and Lawrence W. Sherman for policing domestic violence. However, a new entry was Crime, Shame, and Reintegration by John Braithwaite (1989), with the policy implication of restorative justice programs for offenders. These have become increasingly important in different countries in recent years (Sherman & Strang, 2007).
Another new entry in 2000 was Terrie E. Moffitt’s (1993) theory, which implies that different types of programs are needed for adolescence-limited and life-course-persistent offenders. For adolescence-limited offenders, it is especially important to limit contact with delinquent peers. Research on co-offending (Reiss & Farrington, 1991) suggests that it is essential to identify and target “recruiters,” offenders who repeatedly commit crimes with younger, less experienced offenders, and who seem to be dragging increasing numbers of young people into crime. Programs that put antisocial peers together may have harmful effects (Dishion, McCord, & Poulin, 1999). Moffitt also suggests that, in order to target the “maturity gap” of adolescence-limited offenders, it is important to provide opportunities for them to achieve status and material goods by legitimate means.
The most-cited works of the most-cited scholars in 2005 show both continuity and change in their theoretical concerns and policy implications. The most-cited work of the most-cited scholar is still Crime in the Making (Sampson & Laub, 1993). The criminal career paradigm continues to be seen as important, significantly impacting research in the field.
Other areas that retain importance include rational choice theory, the effectiveness of correctional treatment, and the theories of Moffitt (1993) and of Gottfredson and Hirschi (1990). A newly identified work with clear policy implications for reducing crime is Preventing Crime by Lawrence W. Sherman and his colleagues (1998); this report has helped policy makers and criminal justice officials better understand what types of programs do (and don’t!) work to prevent crime. Sherman’s second most-cited work, “The specific deterrent effects of arrest for domestic assault” (Sherman & Berk, 1984), which reported the results of the Minneapolis Domestic Violence Experiment, has also had an enormous impact on criminal justice policy (see above).
Rehabilitation is another area with continuing influence. Francis T. Cullen’s most-cited work, “Does correctional treatment work?” (Andrews et al., 1990), focuses on the effectiveness of correctional treatment, as does the most-cited work of both Andrews and Bonta, The Psychology of Criminal Conduct (Andrews & Bonta, 2003), which emphasizes the policy issues of risk assessment and correctional effectiveness.
Conclusions
This study employs a longitudinal design in which citations are studied in the same set of 20 journals at 5-year intervals. Exactly the same methods and journals are used in each year. The nature of a longitudinal design results in several limitations. Because we are comparing results across years, we are unable to add new journals to the study but are limited to long-established journals that were being published at the time we began the research. Additionally, we cannot take advantage of new sources of citation data that have been developed but must use the same data collection method that was developed at the start of this research. However, the longitudinal design of this study is also a major strength. Studying the same set of journals at five-year intervals allows us to examine citation trends over time and to consider how the influence of both scholars and works waxes and wanes.
Expanding the number of CCJ journals from 9 to 20 identified many of the same most-cited scholars found in other research that involved fewer journals. However, the rankings of some international scholars and some scholars working in less mainstream CCJ areas were improved when the number of journals was increased. The use of additional journals had both advantages and disadvantages. The obvious advantages of increased coverage were to some extent counteracted by the disadvantages of including progressively less mainstream CCJ journals. Continuing to expand the analysis to even more journals would require the inclusion of more peripheral or specialized journals, thus further diluting the importance of mainstream CCJ topics. It would, however, increase the visibility of scholars who publish in more specialized fields. The present results depend to a considerable extent on the choice of journals to be analyzed.
The analysis of citations in 20 journals was restricted to data from only 1 year because this research was carried out with limited resources and no external funding. The use of only one year of citations in each journal inevitably caused more variability in the results over time than would occur in analyses based on five years. However, while the most-cited scholar in one journal could possibly be affected by one article that extensively cited a single author, the main focus of this paper is on identifying the most-cited scholars in groups of 5, 10, or 20 journals. Studying groups of journals, which are less vulnerable to such distortion, reduces the variability of the results and increases both the validity and reliability of the findings. Against the variability argument, there was considerable agreement between the results of this research and results obtained in earlier studies using the same set of journals, as well as results obtained using more years of data from smaller numbers of journals.
In order to make comparisons over time, this study focused primarily on the total number of citations as a method of determining scholarly influence, but also reported prevalence (the number of different articles in which a scholar was cited) for the most-cited scholars. Prevalence may be a better measure of the influence of one scholar on others than raw citation counts. In addition, this research has shown the increasing citations of younger scholars and the decreasing citations of older scholars.
This research has identified some scholars (e.g., Robert J. Sampson, John H. Laub, James L. Bonta), as specialized, because their influence was based primarily on one or two highly cited works, frequently books and usually of a theoretical nature. Other scholars, (e.g., David P. Farrington, Francis T. Cullen) were found to be versatile, because they had many different works, usually articles, which were cited only a few times each. Thus, it appears that there are two different ways in which scholarly influence operates in CCJ. One way a scholar influences his or her colleagues is to write a scholarly book that proposes a new theory, such as Crime in the Making (Sampson & Laub, 1993) or A General Theory of Crime (Gottfredson & Hirschi, 1990); inevitably, the book is cited by later authors who discuss the theory. The second is to write a large number of different works on a variety of topics so that, while no individual work is a highly cited seminal work, the many articles that are “at risk” of being cited lead to a large number of total citations for the author.
This research also illustrates the possible importance of working with influential coauthors. For example, the most-cited scholar in the 20 journals in 2005 is Robert J. Sampson; his most-cited work is coauthored by John Laub, who is ranked second. In almost every table in which Sampson is highly cited, Laub is also one of the most-cited scholars. In fact, Sampson is the one of the top two most-cited scholars in each of the four groups of 20 journals and Laub is also one of the most-cited scholars in each group. Similarly, Francis T. Cullen is the fourth most-cited scholar in the 20 journals; his most-cited work is co-authored by James L. Bonta, the fifth most-cited scholar, and Don A. Andrews, who is ranked ninth in the 20 journal list. This suggests that a scholar’s influence may be related to the influence of his or her coauthors; publishing jointly with a scholar who publishes high quality research and who publishes frequently may increase one’s own prestige in the field.
While the study of 20 journals has both advantages and disadvantages in comparison with other analyses, these results add to the growing body of knowledge concerning changes in the most-cited and most influential scholars and works in CCJ over time. Future research into citation analysis should continue this longitudinal series of studies and should focus more on changes in the most-cited works over time, perhaps as a way to eventually predict important research topics and provide greater assistance to policy makers. We conclude that these analyses reveal changes over time in scholarly influence and in theoretical concerns and policy issues.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research and/or authorship of this article.