Representation in <i>BioScience</i> Authorship

The journal BioScience, which first appeared in 1951 as the AIBS Bulletin, is published by the American Institute of Biological Sciences (AIBS). The journal seeks to provide a specialized niche in scientific publishing in that it is one of the few journals dedicated to review and synthesis papers broadly related to ecology and conservation biology, rather than primary research articles on these topics. The only exception to this rule is with primary research on secondary and postsecondary education and on citizen science. Peer-reviewed articles are submitted to the journal under a variety of categories, with Overview, Forum, and Education articles being the most common. Other peer-reviewed categories include Thinking of Biology and Biologist’s Toolbox. Our distinct focus on review and synthesis articles, for the most part, likely influences the career stage and diversity of authors interested in publishing in BioScience. The life sciences have long suffered from the underrepresentation of minority communities, a well-known problem that is, unfortunately, more frequently commented on than substantively addressed. In 2021, the AIBS introduced its Diversity Plan (Croslan et al. 2021) to outline concrete, measurable actions to improve the state of biology. In the plan, the authors expressed their commitment to “developing programs that center on our core activities of assessment, training, communication, and advocacy.” Consistent with the outlined goal of assessment, BioScience staff have sought to understand more about the diversity of its authorship, with the aim of creating a baseline of knowledge from which to measure progress and focus future efforts to infuse more diverse perspectives into BioScience. To this end, a survey was disseminated to BioScience authors to gather demographic, authorship, and publication information. We have briefly described the results already (https://doi.org/10.1093/biosci/biab118). In the present article, we analyze the results, with the intention of promoting transparency in the first step of our process to guide us “toward a future in which a diversity of voices and perspectives, reflecting that of society at large, is presented within the pages of BioScience” (Collins et al. 2021). We developed a survey using Google Forms to better understand the baseline demographics of Bioscience authors, asking standard identity questions about gender, sexuality, race and ethnicity, institution type, country, and years since the last degree was awarded (age). We also collected information about the respondents’ activities with Bioscience—namely, their most recent authorship role (first, last, middle), how many of their articles had been accepted by BioScience in the last 3 years, and the type of article (e.g., Overview, Education, Viewpoint). We provided predetermined multiple-choice answers for all questions (for race or ethnicity and article type, the respondents could mark all that apply). For demographic questions, the respondents were also given a “prefer not to answer” option and for sexuality, race, and institution; an “other” option was provided with an open text field to submit an answer not represented in the choices. A total of nine questions were developed, including an optional open text field at the end for any general comments. We estimated that the survey would take five minutes to complete. In the survey's introduction, we outlined the purpose of the survey, the anonymity and privacy of the data collection, the potential benefits in participating, and a consent to the analysis of the anonymized data collected. The respondents were also given an email to contact if they had any questions. We compiled a list of 2028 email addresses of potential respondents from our ScholarOne peer-review system. The list of potential respondents included authors of BioScience articles accepted between 1 January 2018 and 31 December 2020. Because some authors had multiple email addresses and because it was not feasible to limit one email address per individual, the respondents were given instructions in an invitation email to fill out the form only once. Our survey was sent out 11 August 2021, with 96% of the emails successfully delivered. One reminder email was sent out to the list 2 weeks later, and the survey was closed in early October 2021. A total of 283 responses were recorded from the survey (13.9% response rate), which compares favorably with similar surveys conducted by AIBS. Although all of those respondents completed the survey, only 35 made a general comment in the optional text box, with their responses ranging from those about how the survey was implemented to comments on the BioScience journal itself. A detailed description of our data analysis can be found in the present article's supplemental materials. The geographic distribution of BioScience authors is displayed in ­figure 1; whereas a total of 33 countries were represented, this distribution was heavily skewed, with 62% of the authors from the United States, 5% from Canada, and all other countries representing less than 5%. The proportion of male and female authors approached parity (figure 2a), with five respondents preferring not to answer. The overwhelming majority of the authors identified as heterosexual (figure 2b), with eight respondents preferring not to answer. Similarly, most of the authors identified as Caucasian (about 7% identified with multiple categories); no other category represented more than 10% (figure 2c), with three respondents preferring not to answer. A total of 66% of the authors had earned their degree in the last 20 years (figure 2d), with three respondents preferring not to answer. The most frequently indicated author institution was, unsurprisingly, academia (75%), with 10% of the respondents indicating government, 8% indicating a nonprofit, 6% indicating other, and 1% from the industrial sector, with one respondent preferring not to answer. The respondents’ demographics correlated with age (years since degree) and fewer older respondents identified as female, non-Caucasian or White, or not heterosexual (figure 3). Geographic diversity of authorship. Author demographics: (a) gender, (b) sexuality, (c) race or ethnicity, and (d) Years Since Degree. The intersection of age and gender, sexuality, and race or ethnicity. The respondents were asked to indicate the type of article they had most recently published in BioScience (they could choose all that applied). Overviews and Forums made up the majority of the article types (figure 4a). A total of 11% (n = 32) indicated more than one article type. In terms of their recent authorship role, only 16% indicated that they had been the last author (figure 4b), with three respondents preferring not to answer. The respondents also indicated that they appeared as an author on an average of 1.4 (standard error = 0.1) articles accepted by BioScience in the last 3 years; the majority of the respondents had only one accepted in this timeframe ­(figure 4c). The maximum number of articles reported by a respondent was 10. Authorship characteristics as a function of (a) article type, (b) authorship type, and (c) the number of articles. To examine how the demographic characteristics of authors may interact with their authorship role, we created a series of regression models. Because there was a correlation between age and other demographic variables, to avoid multicollinearity, we separated the data into two groups: those who had earned their degree in the last 20 years (n = 186) and those who received their degree 20 years ago or more (n = 94). The respondents who indicated that they preferred not to answer any of the demographic questions were not included in these analyses. The regression analyses were conducted separately for each group, examining the effects of gender, sexuality, and race or ethnicity on all three authorship characteristics. The results are listed in supplemental tables S1–S6. None of these models were statistically discernible, although the regression of article type of the older data set (earned degrees 20 years or more) approached discernability, and gender was observed as a statistically discernible predictor in this model (supplemental table S2), but not in the model of the younger cohort (supplemental table S1). To explore this further, we plotted the percentage of the respondents with an Overview or Forum article as a function of gender for both cohorts and found even levels of article-type publication across genders for the younger cohort but somewhat disparate levels of for the older cohort (figure 5). Overview or Forum as a function of gender by years since degree. Using the full data set, we conducted separate regression models to explore only the effect of age on all three authorship characteristics. The results are listed in supplemental tables S7–S9. All three models were statistically discernible, although the amount of variance they explained was small (3%–5%). Nevertheless, age was a discernible predictor of all three authorship characteristics. We then plotted the distributions of the number of years since their degree alongside first and middle authorship compared with last authorship, which suggested an older age for last authors (figure 6a). Examining the number of years since their degree as an interaction with article type suggests that the authors that published article types other than Overviews or Forums are slightly older (figure 6b). A plot of the distribution of the number of years since their degree as a function of the number of articles suggests that the authors of multiple articles are slightly older than those of only one article (figure 6c). It should be noted that the authorship characteristics data for all three plots were somewhat skewed (most of the authors were not last authors). Nevertheless, it seems likely that older authors in our sample were more likely to have published more than one article in BioScience over this period, were more likely to be a last author of an article, and were more likely to publish article types that were not Overviews or Forums. Age distribution (years since degree) as a function of (a) authorship, (b) article type, and (c) the number of articles published. Our respondents were younger, were not often the last author, and often published only once. Given the small size of this sample and the skew toward younger authors, there may be limits to the generalizability of these results. It may be that younger authors were more likely to respond to our survey. It is clear that age affects the demographic distributions of authors, in terms of race or ethnicity, sexuality, and gender. Future inquiries should make sure to consider this interaction, which, based on our data, suggests an increasing level of diversity in younger generations. Age was also the major demographic factor suggested by our analysis to influence authorship characteristics; other factors, such as gender, sexuality, and race, were generally not observed to be discernible predictors. These results make sense, given the lifecycle of science researchers; most senior scientists have funding for their own lab, allowing them to publish more papers in general and for their role to be the last author. One interesting observation is that the respondents publishing article types that were not Overviews or Forums tended to be older. Also interesting is that older female respondents were less likely to publish Overview or Forum articles. The cause of these observed tendencies is unclear, but they may reflect, to some extent at least, the specialized publishing niche that BioScience occupies. Black scientists were 3.5% of the PhD recipients in the biological sciences, despite representing 13% of the US population, according to 2018 National Science Foundation (NSF) data (NSF 2019). Although our finding that 4% of BioScience authors identify as African American or Black seems in line with the proportion of Black biologists in the United States, this figure is woefully lower than the general population. A similar case exists for Latinx scientists, who make up 6% of biology PhD recipients and 10% of our sample but roughly 19% of the US population. More work is needed throughout the pipeline, and all scientific entities have a part to play. Moreover, the same NSF (2019) study indicated that 53% of biology PhDs were granted to female scientists, and women make up 50.1% of the general population. Although 46% of the surveyed BioScience authors identified as female, this percentage is slightly lower than both of these other statistics, suggesting more work still needs to be done to achieve proportionality. It would be useful to compare our data with those of other scientific journals, but there are surprisingly few publicly available data sets, and the results are not standardized (Else and Perkel 2022) and have often been focused on relatively few demographic characteristics. For AAAS Science journals’ authors and reviewers, those identifying as Black represented less than 1% of the total, although this number is hard to assess because of missing data (AAAS 2020). For cardiology journals, a recent tally suggested that only 2% of their authors are Black, although these results were not self-identified (Else and Perkel 2022). For American Chemical Society journals, Black corresponding authors represented 1% of the total (ACS 2021). In our results, Black authors represent 4% of the journal's total. Similarly, the reported levels of Latinx authors in these journals appear significantly lower than the 10% that are represented in our sample. In terms of gender, the male authors and reviewers for AAAS Science journals were found to outnumber their female authors 2.3:1, whereas in American Chemical Society journals, the corresponding published female authors represented 23% of the total. In our sample, the male authors of BioScience barely outnumbered the female authors (1.2:1). It will be important for journals to continually track these statistics in the future, using more standardized methods so we can develop a sense of how author demography evolves over time and across scientific fields. Also, importantly, few journals seem to track sexuality or age, and few look at the intersectionality of these variables, but these are important factors to consider; they add richness to these types of data and should be more universally embraced. Indeed, one of our respondents commented that representation from people with disabilities is an important area of diversity that was missed in our survey. We will consider this for inclusion in future data collection, and we will continue our commitment to accessibility (OUP 2021). Our ultimate aim is representation in science that reflects the broader society it serves, but achieving this goal will require the efforts of the whole of the scientific enterprise, from educational systems to research funding agencies and publication platforms. At BioScience, we will work in concert with the AIBS Diversity Committee to chart out next steps, which may include improved networking and outreach, as well as a focus on increased diversity on our editorial board and among our commissioned authors. Similarly, because our younger demographic is relatively more diverse, focused outreach to early-career researchers will be key in ensuring that BioScience is home to the diverse perspectives that will shape science's future. Overall, the results from this survey are a baseline—a first step but an important one. Supplemental data are available at BIOSCI online. Stephen Gallo is the chief scientist, and James M. Verdier is the senior editor of BioScience, at the American Institute of Biological Sciences, in Herndon, Virginia, in the United States. Scott L. Collins is affiliated with the Department of Biology at the University of New Mexico, in Albuquerque, New Mexico, in the United States, and is editor in chief of BioScience.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.