Getting Published: an account of
writing, refereeing and editing practices (1996-8).
Interim Report 2
Margaret Scanlon, Angela Packwood, Gaby Weiner.
The Getting Published project was funded by the ESRC (Grant no R000236291)
1.INTRODUCTION: JOURNAL SURVEY
The main research base for this interim report is the project survey of 30 journals over a 10 year period (1986-1995) of which 20 were from Education, and five of which were from Sociology and Psychology. Two issues of each journal were randomly selected for each year though special issues were mainly excluded. (See appendix for journal titles and number of articles selected from each.)
The journal of the learned societies for each of the three disciplines involved in the study (Education, Psychology and Sociology) formed part of the sample. Other journals were selected on the basis of subscription rates and representativeness of journals within each discipline. The survey provided a data base of 4254 articles from 600 issues.
This report considers the findings concerning authorship characteristics: in particular, that of sex, status, institution, country of author, number of authors per article and patterns of self-citation. The data were entered into a relational database (Access) and were analyzed using this database and a compatible spreadsheet (Excel). The report is divided into sections dealing with specific author characteristics as above.
Research Themes
As we noted in Interim Report 1, academic journals are being increasingly used in the UK, the US and elsewhere, for confirming the quality of scholarship of individual academics. The aim has been to use academic journals to aid the ranking of research and scholarly work - for the distribution of research funds, and also to inform decisions concerning appointment and promotion. Publication counts are thus seen as important performance indicators of scholarship.
Performance indicators, properly understood, are thus measures that relate productivity (output) to resources (input). The most obvious output measures relevant to departmental research performance are simple publication counts and more elaborate publication-based measures designed to take quality into account. The most important input variables are the number of departmental staff members, the number of research assistants, the size of equipments and recurrent grants, and the amount of research income. (Colman, Garner and Jolly, 1992:97)
The original conception of the academic journal as a means of generating and debating epistemological and theoretical issues raised by scholars and scientists, seems to have diminished with the onset of pressures to publish deriving from various research assessment exercises. Thus, the sometimes inconsistent editorial practices of journals outlined in Interim Report 1 have become increasingly important, not only to individual writers and academics, but to their institutions which may or may not benefit financially from an academic's heightened publishing profile. The means of using articles in academic journals to evaluate scholarship include the frequency of citations, the level of productivity ie frequency of publication, the prestige of individual journals, and evaluating individual author characteristics. Thus the following suppositions apply:
the higher the number of citations, the greater the peer esteem and therefore the higher the quality of scholarship;
the higher the number of publications, the higher the quality of scholarship;
the more eminent and well respected the journal, the higher the quality of scholarship;
specific individual characteristics are predictive of high quality scholarship, in particular, aptitude and ability, gender, achievement orientation, competitiveness;
different disciplinary traditions have an influence on individual publishing profiles;
institutional factors such as the rating/prestige of the institution and/or department have an impact on publishing output.
(Becher, 1989; Stiver-Lie et al, 1994)
British concerns about academic publishing have focused, in particular, on the usefulness or otherwise of citations (Furnham 1990, Anderson, 1991; Field et al, 1992); the importance of peer review and academic refereeing (Smith and Gough, 1984; Anderson, 1991); institutional influences (Smedley, 1989); the advantages and disadvantages of high publishing productivity (Colman et al, 1992); and editorial requirements regarding submitted manuscripts (Noble, 1989). There are, however, disagreements about methodology. Some advocates of the use of citations over refereeing and peer review, cite the latter as an 'old boy' network which is unfair to outsiders and newcomers (Furnham, 1990). Others such as Field et al (1992) claim that citations too merely reflect the status quo, because of the frequency of self-citation and citation of friends.
Author Characteristics
The characteristics of authors of academic articles have been less discussed in the literature than issues of peer review and refereeing practices. Several sociologists of science have suggested that the status characteristics of authors, for example, where they were educated and where they are presently employed, influence the reviewers' recommendations and the editor's decision whether or not to publish. It has been further suggested that this influence is frequently greater than the merits of the manuscript under consideration (Bakanic, Simon and McPhail, 1987). Thus a 'big' name may well gain the advantage in the competition for journal space:
Judgement about the work of ranking scientists may be systematically skewed by deference, by less careful appraisals involving exacting criteria, by self-doubts of one's own sufficient competence to criticise a great [scholar] or by fear of affronting influential persons in the field (Zuckermann and Merton, 1971:82)
Zuckermann and Merton found in their study of scientists, that high status authors submitted more manuscripts, received quicker turnaround from reviewers and were more likely to be reviewed by peers of equivalent status. Other findings from this early study were that younger authors were more likely to have their manuscripts accepted for publication. However, Zuckermann and Merton (p. 95) concluded that overall, for their study, the relative status of reviewer and author 'has no perceptible influence on patterns of [manuscript] evaluation'. Bakanic et al's more recent study of manuscript submission to, and publication in American Sociological Review, found that prestigious institutions provided better environments, resources, time and encouragement for authors to publish, 'all of which increase the likelihood that more and higher quality manuscripts will be generated and submitted' (Bakanic et al, 1987:637). An author's experience as a referee or reviewer was also more likely to enhance publication and new academics were more likely to publish than those who had been in the profession longer.
New PhDs were more likely to publish than those who had been in the profession a long time; assistant professors were more likely to publish than full professors, associate professors, and graduate students in that order (p. 637)
Gender has only recently received attention as a key characteristic of authorship. The main factors emerging have been the generally higher level of productivity of men compared to women, and the lower levels of citation of women's work (Over, 1982; Helmreich et al, 1980). However Over (1982) found that article-for-article, women were as likely as men to be cited, but their proportion of citations was lower because of their lower overall publication rates. In contrast, Helmreich et al (1980) found that male authors were cited more and also were more likely to self-cite and Ward, Gast and Grant (1992), that male authors were more likely to be cited by men and that male articles had a higher 'citation range' (0-81) than women (0-73) Ward et al also found that women were more likely than men to cite women's work.
Wennerås and Wold (1997) argue that there are three main theories used to explain women's relatively low publishing profile compared with men: time lag ie women haven't yet quite caught up with men but they will, in time; low productivity of women due to lack of ambition and competitiveness, and domestic responsibilities; and sex discrimination. In their study of Swedish biochemists, Wennerås and Wold found that sex discrimination best explained the relatively poor performance of female scientists. In particular they found that peer reviewers were strongly biased against women researchers. Other studies have found women to be under-represented in citations. For example, Field, Lovell and Weller (1991) found that only one woman was among the top 10 cited authors in Studies in the Education of Adults and International Journal of Lifelong Education, and then, only in tenth position.
Other linked findings regarding women's and men academic careers are that women tend to gain their doctorates later than men (Helmreich et al, 1980), women's early work reaches peak visibility slower than men (Ward et al, 1992) and there are large differences between the sexes in academic rank and salary (Rodgers & Maranto, 1989)
This report seeks illuminate some of the above issues by focusing on the author characteristics of the articles surveyed for the project. Its aim is to establish whether the claims regarding academic publishing outlined above can be sustained, and whether it is ethical or indeed possible, to use academic journals as ranking or evaluation instruments.
2.AUTHOR CHARACTERISTICS: GENDER
One of the key issues emerging from the literature, as we have seen, has been a perceived bias against women as authors of academic articles except in women's or gender studies where the reverse is often the case. There are few journals in the three disciplinary areas involved in the study exclusively devoted to gender or women's studies and none in existence for the period covered by the survey (1986-1995). Investigation of the gender of authors of the articles surveyed for the project revealed similar male-dominated patterns to that found in other US and UK studies, though as the tables below show, different journals had a greater or lesser degree of authorship bias. Significantly, no journal in the survey had either equal proportions of males and females as authors, or a predominance of female authors.
Table 2.1:Gender Distribution of Journal Authors in Education, Psychology and Sociology
Sex of author / Education Journals % (n=2884) / Psychology Journals% (n=731) / Sociology Journals
% (n=731)
One male author / 44.5 / 17 / 52
One female author / 17 / 10 / 18
Two male authors / 8 / 13 / 8.5
Two female authors / 3 / 5 / 3
Male/female author / 4 / 5 / 3
Female/male author / 4 / 8 / 4
Other / 19.5 / 42* / 11.5
*Psychology journals have a higher proportion of multiple authored articles (see section 6) For example, 6% articles have 3 male authors and over 10% have 3 or more mixed sex authors.
Table 2.1 shows a higher number of male than female contributors to academic journals over the 10 year period of the study, and that this pattern is more or less constant across the three disciplines. However, for individual journals, clear differences emerge, for example, in the ratio of single male to female authors - see tables 2.2 and 2.3. For Sociology of Health and Illness, the proportion of single male, to single female authors is relatively equal (31:26), whereas for the Journal of the Philosophy of Education male authors are more prominent (74:14) and for Theory Culture and Society (TCS) , nearly 90% of authors are male. Comparative Education and Educational Review are included in tables 2.2 and 2.3 because they reflect rather more representative gender ratios exhibited by the project journals, ranging from approximately two thirds/three quarters male authors to a third/quarter female authors.
Table 2.2: Gender Distribution of Authors of Different Journals
Sex of Author / Sociology of Health & Illness % (n=100) / Comparative Education% (n=145) / Philosophy of Education
% (n=223)
One male author / 31 / 56 / 74
One female author / 26 / 14 / 14
Two male authors / 8 / 8 / >1
Two female authors / 5 / 3 / -
Male/female author / 1 / 2 / -
Female/male author / 7 / 2 / 1
Other / 22 / 15 / 11
Table 2.3: Gender Distribution of Authors of Different Journals (cont.)
Sex of Author / Educational Review % (n=151) / Theory, Culture and Society % (n=116)One male author / 37 / 82
One female author / 10 / 9
Two male authors / 15 / 4
Two female authors / 5 / 1
Male/female author / 3 / 2
Female/male author / 5 / 1
Other / 25
3.AUTHOR CHARACTERISTICS: OCCUPATIONAL STATUS
As we saw in the introduction, occupational status has been identified as a key factor for individuals in gaining a publishing record. However, the project team had problems with identifying author status as it was often not included with an article. We refer here to solo or first author status. As can be seen from table 3.1, there was a high level of non-specification of author status - just under 50% for Education authors, just over 50% for Sociology authors and nearly 70% for Psychology authors. Another difficulty was the use of the title of doctor, which could mean either a medical doctor or someone with a PhD.
Nevertheless, despite such problems of definition, certain differences in the occupational status of authors between the disciplines is evident. For example, table 3.1 shows rather more authors as lecturers and senior lecturers in Sociology (25%) and Education (18%) , than in Psychology (3%); and rather more professors as authors in Education (10.5%) and Psychology (6%), than in Sociology (1%). The reason for these discrepancies is not entirely clear. It could reflect the different constituencies of the three disciplines ie a possible higher number of professors overall in Education and Psychology, or a greater tendency in Education and Psychology, to use the title of professor in publications.
Table 3:1: Occupational Status of Authors of Journal Articles in Education, Psychology and Sociology
OccupationalCategory / Education %
n = 2884 / Psychology %
n = 731 / Sociology %
n = 639
Lecturer/Senior Lecturer / 18 / 3 / 25
Doctor / 14.5 / 16 / 11
Professor / 10.5 / 6 / 1
Researcher / 4 / 1 / 9
Practitioner / 7.5 / 7.5 / -
Unspecified / 45.5 / 66.5 / 54
If we focus on individual journals, however, it becomes clear that the individual cultures and practices of journals not only result in different patterns of occupational status, but also in different traditions of status disclosure. Thus some journals, for example, Sociology and Research Papers in Education encouraged authors to provide short biographical details, including current occupational status, but others did not, for example, British Journal of Sociology and Educational Research. (see table 3.2)
Although it is difficult to draw conclusions from such diverse data, and though there are substantial variations between journals, certain trends in authorship are discernible: for example, the relatively high proportions of professors and doctors as authors, and for some journals, the prominence of lecturers and senior lecturers, and to a lesser extent researchers.
Table 3.2: Occupational Status of Authors: Differences between Journals
OccupationalStatus / Soc [1]
n=154% / BJS
n=143% / QJEP
n=179% / BJSP
n=155
% / ER
n=162
% / RPE
n=86
%
Senior/
Lecturer / 33 / - / 2 / - / 8 / 23
Professor / 16 / 4 / 21 / - / 1 / 17
Doctor / 25 / 4 / 50 / 3 / 9 / 24
Practitioner / - / 2 / - / 10 / 5
Researcher / 15 / 2 / 2 / - / 11 / 19
Student / - / 3 / 3 / - / 1 / -
Other / 2 / 1 / - / - / 2 / 5
Unspecified / 9 / 87 / 19.5 / 97 / 59 / 7
4.AUTHOR CHARACTERISTICS: INSTITUTIONAL BASE
As was pointed out in the introduction to this report, one of the key factors in academic publishing and in achieving the status of high ranking scholar, is institutional base. It has been suggested that an institutional base with a good research reputation is likely to be more of an asset when seeking to publish, than one with less of a reputation. This is borne out in findings from the scrutiny of institutional bases of the authors of the articles in the survey. Again we refer here to solo or first author status. As we see from table 4.1, universities were by far the most common institutional bases for the authors of academic articles. The proportion of polytechnics involved, though low, may seem misleadingly so, as all UK polytechnics changed their status to university in 1992. Therefore for the period 1992-5, previously polytechnic-based authors would be described as university-based.
From table 4.1, we can see that just under three quarters of authors in Education journals and just over three quarters of authors in Psychology and Sociology, were based in the university sector. Differences occur between disciplines, however, in some of the more vocationally-orientated areas of research. For example, 7% of Education authors were based in schools and Local Education Authorities (LEAs), and a slightly higher proportion of Psychology authors, in hospitals and medical centres (9.5%). Interestingly, Psychology authors were more likely to be based in research centres (7%) rather than polytechnics or colleges, whereas for both Education and Sociology, 10% of authors were based either in colleges or polytechnics.
Table 4.1Institutional Base of Journal Authors in Education, Psychology and Sociology
nstitutional Base of Author(s) / Education %n = 2884 / Psychology %
n = 731 / Sociology %
n = 639
University / 72.5 / 77 / 78
Polytechnic (until 1992) / 4 / 1.5 / 5
College / 6 / - / 5
Research Centre / 2 / 7 / 3
Medical Centre / - / 1.5 / -
Hospital / - / 8 / -
LEA / 3 / - / -
School / 4 / - / -
Other / 8.5 / 5 / 9
When we consider individual journal patterns of authors' institutions, we find similar concentrations of authors in universities (70%-90%) for Sociology and Psychology, but different trends for Education, as can be seen in table 4.2 below.
Table 4.2Institutional Base of Authors in Education Journals
Institutional Base of Author(s) / ER [2]n=162 % / BERJ
n=148 % / BJSpE
n=182 % / CE
n=145 %
University / 59 / 78 / 33.5 / 86
Polytechnic (until 1992) / 7 / 5 / 2 / 1
College / 4 / 6 / 2 / 3
Research Centre / 10.5 / 2 / 2 / 1
Medical Centre/Hospital / 2 / - / 2.5 / -
LEA / 6 / 2 / 21 / -
School / 4 / 2 / 20 / -
Business / 1 / 1 / 1 / -
Other / 4 / 4 / 11.5 / 5.5
Unspecified / 3 / 3 / 5 / 1
While some Education journals showed similar patterns to Sociology and Psychology, namely that most authors (over 70%) worked in universities, others have a lower proportion of authors from universities and a higher proportion from other professionally related institutions such as schools and LEAs. A good example is British Journal of Special Education which had only a third of its authors based in universities, and approximately a fifth each coming from schools and LEAs, with another 11% from other institutions such as charities and voluntary societies. This reflects both the vocational and academic nature of BJSpE's content and readership, and suggests that academic journals with strong vocational orientations are more likely to attract authors from outside the university.
5.AUTHOR CHARACTERISTICS:COUNTRY OF AUTHORS
As table 5.1 shows, project articles tended to be written by UK-based academics, although rather more Psychology articles came from the US than for either Education or Sociology. As might be expected, English-speaking countries provided the bulk of the authors, although there was a smaller number of contributions from Holland, Germany, the Scandinavian countries and Israel.