2. Literature and Research Questions
Explanatory approaches for the gender gap in academic research are diverse and range from individual characteristics (such as self-efficacy), to structural factors (such as family, networks, discrimination etc.). A highly cited reason for female researchers’ lower career success is their averagely lower research productivity (
Fox 2005;
Jagsi et al. 2006;
Sidhu et al. 2009;
Sugimoto et al. 2013;
Symonds et al. 2006). Indeed, publishing is one of the most important factors for employment in academic research, acquiring prestigious positions and attaining full professorship (
Lutter and Schröder 2014;
Plümper and Schimmelfennig 2007). Building a publication record in early career stages is, hence, an important step for the preparation of a successful academic research career (
Feldon et al. 2017). However, it is questionable whether fewer publications merely are a cause, or also an effect of lower status and/or disadvantageous job characteristics of female academics. As
Fox (
2005) puts it: “Publication productivity reflects women’s depressed rank and status, and partially accounts for it” (p. 31). Female postdocs in Germany are, for instance, four times more likely to work on a temporary contract and their contracts are shorter in comparison to those of their male counterparts (
Konsortium Bundesbericht Wissenschaftlicher Nachwuchs 2013). Under such circumstances, publishing is probably more difficult. In addition, empirical results hint that women engage more in teaching than men do (
Link et al. 2008;
Winslow 2010). More time invested in teaching equals less time that can be invested in research, resulting in a penalty in publication productivity (
Fox 1992).
Thus, rather than explaining women’s underrepresentation with lower productivity, it should be a prior goal to explain publication differences by gender. While some studies do not support lower publication productivity by females (
Schubert and Engelage 2011), this can be due to the fields of study analyzed, but also to local/geographical differences (
Sugimoto et al. 2013). A recent study by
Sugimoto et al. (
2013), analyzing publication patterns worldwide and in various disciplines, supports that “[m]en dominate scientific production in nearly every country” (
Sugimoto et al. 2013, p. 212). This applied to co- and first-authorships.
In addition to geographical differences and the differences among disciplines, career stages also need to be considered: While Ph.D. students, as surveyed by
Schubert and Engelage (
2011), may not vary so greatly with respect to their work and family circumstances, these might, however, become visible in the postdoc phase. It has been shown, for instance, that females’ publication productivity is diminished one year after childbirth more so than for men (
Hunter and Leahey 2010). Since academic researchers, fearing a negative impact on their career progression, start their family in later career stages (
Metz-Göckel et al. 2014), Ph.D. students are less likely to have their first child in comparison to postdocs. Hence, childcare responsibilities will probably contribute to gender differences in a more advanced career stage. Different career stages are also likely associated to the variety of pursued activities. Ph.D. students may vary less with respect to the time invested in different academic activities, as their employment circumstances are more standardized. In the US, they usually pursue their Ph.D. within a graduate school and are not part of the academic staff. In Germany, the Ph.D. is most commonly achieved while having the position of a research assistant, on a scholarship, or within a graduate school (
Kreckel 2011). While most Ph.D. students are conducting their thesis work as academic staff members, graduate schools with scholarship funding are becoming more popular (
Konsortium Bundesbericht Wissenschaftlicher Nachwuchs 2017). In the life sciences, the Ph.D. is also recommended for students who want to pursue any career outside of academic research and is, therefore, almost considered the standard degree (
Brockmann and Kühl 2015)–unlike in other disciplines, e.g., the social sciences (
Destatis 2016). Since Ph.D. students were, until recently, not assessed by their universities, there is a lack of knowledge across all disciplines with respect to variables such as proportion of cumulative versus monographic theses, international collaboration, etc. (
Konsortium Bundesbericht Wissenschaftlicher Nachwuchs 2017). With the recent introduction of a new law, however, universities are now obliged to gather information about their Ph.D. students and this knowledge gap will be closed in the future (
Konsortium Bundesbericht Wissenschaftlicher Nachwuchs 2017). During the Ph.D. period it is plausible to assume that most Ph.D. students invests most of their time in conducting the research related to their Ph.D. project.
With respect to the life sciences, a recent study by Feldon and colleagues (2017) analyzed first year Ph.D. students. The study finds that female researchers, in contrast to their male counterparts, spent more time on supervised research in the laboratory, but were “rewarded” with fewer publications (mostly as a co-author). Research self-efficacy beliefs did not explain these differences. Indeed, a recent German study did not find gender differences in research self-efficacy beliefs in life sciences Ph.D. graduates (
Epstein and Fischer 2017). Since in the study of Feldon and colleagues first year Ph.D. students mainly published as co-authors, it is possible that supervisors rewarded male and female Ph.D. students differently, and thus objective and transparent standards may be missing. This finding could hint at the so-called Matilda Effect (
Rossiter 1993), describing a “systematic underrecognition of female scientists” (
Knobloch-Westerwick et al. 2013). The Matilda Effect was recently tested and supported within an experimental research setting: study participants rated contributions of researches less favorably when they were labelled as coming from a female rather than a male researcher (
Knobloch-Westerwick et al. 2013). This is in line with research showing gender stereotypes being linked to the evaluation of male and female behavior at the workplace. Expectations based on gender stereotypes, e.g., men are competent and women are caring, “can compromise a woman’s career progress” (
Heilman 2012, p. 114), since expectations also influence the perception of performance (
summarized in Heilman 2012). Moreover, it has been found that successful females in male-dominated areas (such as academia) are liked less than equally successful males (
Heilman et al. 2004). Being disliked can have detrimental effects on females’ career progression and is related to salary and job opportunities (
Heilman et al. 2004).
Considering the importance of networks for successful collaboration, the integration into the scientific community might be an important factor contributing to productivity differences. There is a “broad consensus that embeddedness in academic social networks—notably informal networks is both crucial for doing research, and for achieving a career” (
Kegen 2013, p. 80). Collaborating with other researchers has also been related to publication outcome (
Landry et al. 1996;
Lee and Bozeman 2005). Moreover, collaborations are necessary for co-authorships, which can help to build up a publication record within a shorter time, than if only focusing on lead authorship (
Feldon et al. 2017). Empirical evidence suggests that females are less integrated into their scientific communities (
Kyvik and Teigen 1996;
Schubert and Engelage 2011), or that their network ties are less beneficial (
Feeney and Bernal 2010;
Fuchs et al. 2001). A study of
Fuchs et al. (
2001) did not support gender differences with respect to scientific community integration; however, the authors found that contact to the scientific community only increased males’ but not females’ likelihood to stay employed in academic research (
Fuchs et al. 2001). Accordingly,
Schubert and Engelage (
2011) found that male Ph.D. graduates acquired their first job more frequently thanks to social ties than women. These gender dependent outcomes of social ties are probably a matter of quality: quality of the tie (e.g., weak or strong) and/or characteristics of the contact/the social network and the social capital it entails. Furthermore, other research studies suggest that females tend to have a more local/domestic, and less international network in comparison to men (
Abramo et al. 2013;
Sugimoto et al. 2013). Homophily could also play a crucial role in recreating the disadvantageous position of women in the academic setting (
Bozeman and Corley 2004;
Kegen 2013). Bozeman and Corley find that 84 percent of collaborators of females in non-tenure track positions are also females. Since female researchers, on average, have lower academic ranks, it is likely that homophily is one reason female researchers seem to benefit less from their networks.
Concerning the occasionally conflicting results about the collaboration tendencies of each gender, the field of study is also an important factor. For instance, empirical studies suggest smaller gender differences in the social sciences (
Hunter and Leahey 2008;
Sugimoto et al. 2013). However, while
Hunter and Leahey (
2008) did not find gender differences in the collaboration frequency in a sample of researchers in sociology, they did not include the quality and benefit from collaborations of each gender in their study. As stated above, the career status may be of further importance when analyzing gender differences in integration/collaboration and interpreting empirical results. Investigating researchers who are well established in their field will probably reveal different results, since the population is already highly selected. Moreover, the problem of a selective sample at least partly applies to the mere analysis of publications, since it excludes researchers who miss publishing opportunities, e.g., due to a lack of collaborations.
Furthermore, integration into the scientific community could affect female and male researchers differently in their career decisions. Research has suggested that males are more driven by their achievements, whereas females seem to be more influenced by social cues (
Hoffman 1972;
Widom and Burke 1978;
Mottaz 1986;
Kim 2005). While in the German context, female life sciences Ph.D. graduates seem to aspire for an academic research career as much as their male counterparts, at least at that early career stage (
Epstein and Fischer 2017), it is possible that different variables influence their intentions. Male Ph.D. graduates may be more driven by their objective achievements (publications) and females may bestow a greater importance to cues of social embeddedness. In this context, the article explores three research questions:
Research Question 1: Do male and female Ph.D. students differ with respect to their pursued tasks during the Ph.D. and their publication outcomes?
Research Question 2: Are Ph.D. characteristics (pursued tasks and integration into the scientific community) associated differently with publication outcomes, depending on gender?
Research Question 3: Do Ph.D. characteristics and outcomes affect the intention to pursue an academic research career differently, depending on gender?
5. Discussion
With respect to Research Question 1 (cf. Chapter 2) the results of our study showed no significant difference in the relative time spent on the tasks during the Ph.D. (research, teaching, and administrative tasks) between female and male Ph.D. students in the life sciences. However, in line with previous research results (
Fox 2005;
Sidhu et al. 2009;
Sugimoto et al. 2013;
Symonds et al. 2006), women published significantly less. This applied to first author articles, but not the number of articles as a co-author. Surprisingly, a cooperative atmosphere in the scientific community was linked to a lower number of first author publication for females. This result may indicate that females are more motivated to publish in competitive environments. However, since the effect size was rather small, this potential link is rather a topic for future research.
Since research articles in the life sciences usually have a large number of co-authors (
Tscharntke et al. 2007), the position and number of co-authors is an important indicator of author’s contribution (
Abramo et al. 2013). These aspects were not assessed and limit our results as we only captured the number of co-authored articles but neither the relative position in the authors’ list nor the number of co-authors. This may be one reason, why gender differences could only be found for first authorship but not for co-authors. The perceived integration into the scientific community and working group did not differ by gender. Since we did not measure social ties directly, respondents’ perception could have deviated from their actual embeddedness.
Research Question 2 asked, whether Ph.D. characteristics are differently associated to Ph.D. outcomes by gender. In agreement with previous research, indicating that female Ph.D. students benefit less from supervised research by publication output (
Feldon et al. 2017), the results of our analyses show that only for males integration into the scientific community is related to significantly more publications as a first author. This could mean various things: it may be possible that males are more successful in using their social capital i.e., asking contacts for advice, finding established researchers as co-authors, etc. As already mentioned in the introduction of this paper, males and females’ networks differ with respect to the gender of their contacts and (inter)nationality. Their professional contacts could, moreover, differ substantially in rank and experience. As stated earlier, the indirect measure of social ties, i.e., missing information on the quantity and quality of males’ and females’ scientific contacts, could have biased the results of males and females’ embeddedness. Females may just feel as embedded as males, but their contacts could be of lower rank, or less prestigious, and their ties may be weaker. Our results are limited with respect to these details since these measures were not included. With respect to success at using one’s social capital it is important to note that in Germany, Ph.D. students are highly dependent on their supervisors (
Berning and Falk 2005). Attempts involving the introduction of more structure and an increase in the number of supervisors to the Ph.D. have been made to resolve this situation. Hence, with respect to attaining first authorship on the project they work on during their Ph.D., students may yet be at the mercy of their supervisors. In addition, the common practice of “gift authorship”
4 in the life sciences (cf.
Tscharntke et al. 2007) is an indicator that fair practices in attributing authorship may not always be present. Supervisors’ behavior towards male and female Ph.D. students and practices of attributing authorship during and beyond the Ph.D. should be analyzed by future studies.
In consideration of evidence supporting the Matilda Effect (cf. Introduction), it is also possible that male Ph.D. students are more often offered lead authorship in cooperative projects. As described in the Introduction, a person’s gender is linked to performance evaluations (
summarized in Heilman 2012). Also a recent study (
Sarsons 2017) provides evidence “that a person’s gender influences the way others interpret information about his or her ability” (
Sarsons 2017, p. 1). In her study, female surgeons were not only punished (drop in referrals) more after the death of a patient; they also benefited less from good patient outcomes compared to their male colleagues (lower increase in referrals). Further, results pointing to the different liking expressed towards successful females and males (
Heilman et al. 2004) could be meaningful: If talented female researchers are disliked more than their male counterparts, superiors and colleagues may be less inclined to support their career progression. The reason(s) why males seem to profit more from their social embeddedness is an important topic to be investigated in the future.
Pertaining to the kind of compiled thesis, working on a monograph thesis in comparison to a cumulative thesis, was associated with fewer articles as a first author. The result is not surprising, since a cumulative thesis is linked to a certain number of mandatory publications. The result clearly suggests that a certain publication obligation to attain a Ph.D., is beneficial in the life sciences. Since paper publications in peer reviewed articles are most important for a future successful academic research career (
Lutter and Schröder 2014;
Plümper and Schimmelfennig 2007), this aspect should be considered for regulations in Ph.D. programs.
The third research question asked, whether female and male Ph.D. graduates are influenced differently in their career decisions through Ph.D. characteristics and outcomes. While intrinsic research motivation highly correlated with the intention to pursue an academic research career for both genders, first author publications were only significant within the male group. For females, co-authored articles and feeling integrated into the scientific community were significantly related to their academic career intentions. Since upper ranks in academic research are still male-dominated and working together with other scientists is especially crucial for being successful in the life sciences, it may be possible that females ascribe more importance to their network for career success than men.
The importance males and females ascribe to different aspects for their career progress should be analyzed in the future. Moreover, the kinds of collaborations male and female Ph.D. students have within and outside their working group should be analyzed, to better understand why the same level of self-rated integration and cooperation is related to a greater number of first author publications for male but not female Ph.D. students. Since first author publications are among the most important factors for furthering one’s career in academic research and eventually being appointed as a full professor (
Lutter and Schröder 2014), it is important to understand the pitfalls which lead to a lower first author publication outcome for females. Understanding the mechanism behind the gender difference of our study (effect of scientific community integration only positively associated to first author publications for male Ph.D. students), is crucial to derive practical implications. Supporting female Ph.D. students with their attempts to build a network in the scientific community and/or sensitizing supervisors with respect to that topic could be fruitful. Further, Ph.D. students can benefit from their scientific contacts on many other levels. A sense of integration itself may be related to a higher motivation to pursue an academic research career, elicit creative new research ideas and projects, etc.
Our results are not readily transferable to other fields with different gender ratios and disciplinary cultures. More research is needed to analyze whether these patterns can be found in other domains, such as, e.g., the social sciences or math-intensive fields. Furthermore, our results are limited with respect to missing variables in the dataset: childcare responsibilities could have had a different impact on males’ and females’ publication output. As noted, however, academic researchers often postpone their family planning to later career stages out of concerns regarding their career progression (
Metz-Göckel et al. 2014). For Ph.D. students, this aspect may not yet have been relevant. In alignment with previous findings, the majority of our participants was still childless (cf. chp. 3.1.). As mentioned before, data regarding the number of co-authors and position as co-author were missing. Since this is a crucial aspect in the assessment of research productivity in the life sciences, (
Abramo et al. 2013) our results on co-authorship may be biased. The number of co-authors and their relative position in authors’ lists should be analyzed by future research.