Sage on the Stage

Abstract

Who presents at conferences matters. Presenting research benefits speakers, and presenters shape the conclusions audiences draw about who can succeed in a field. This is particularly important for members of historically underrepresented or disadvantaged groups, such as women. We investigated gender representation over a 13-year period among speakers at the largest social and personality psychology conference. On average, women were underrepresented as speakers, though this effect diminished over time. Chairs appeared to serve as gatekeepers: In symposia chaired by women, almost half of the invited speakers were women, whereas in symposia chaired by men, it was a third. The representation of women as speakers varied significantly by academic rank, with women underrepresented at lower ranks but not as full professors, and by topic. Women also tended to present with a smaller, less varied array of individuals than men, though this could be explained by women’s lower average academic rank.

Keywords

gender higher education diversity gender representation social network development

Presenting research at a conference has obvious positive consequences for speakers. Talks increase visibility for the speakers and their research, particularly among colleagues and high-status others who may cite their work, write their evaluation letters for tenure, recommend them for grant review boards, and invite them to future conferences, panels, and editorial boards (Kite et al., 2001). Less obvious are the consequences for audience members. Beyond the specific knowledge conveyed by the presentations, audience members also learn about what kind of person is likely to be successful in their field. For example, when graduate students and junior faculty who are women or members of other underrepresented or historically disadvantaged groups see speakers who share their identity, this signals that they too can succeed (Asgari, Dasgupta, & Cote, 2010; Stout, Dasgupta, Hunsinger, & McManus, 2011). Conversely, a lack of similar others implies that success is unlikely for them (Ceci, Williams, & Barnett, 2009). Given these important potential consequences of who is “on stage” at conferences, we analyze women’s representation in peer-reviewed symposia at the annual conference of the Society for Personality and Social Psychology (SPSP), the largest organization for social and personality psychologists, from 2003 to 2015.

The Issue of Representation

A variety of scientific fields have examined female representation on speaker panels and enacted policies to address inequities (e.g., Casadevall, 2015). Their reports document underrepresentation, having proportionally fewer women speakers on panels than women members of a society, because it may suggest bias against women. They also document absolute representation because the salience of women speakers may increase perceptions that a field is women friendly and improve women audience members’ beliefs about the ease with which they can succeed (Sanders, Willemsen, & Millar, 2009). Therefore, fields in which women predominate (e.g., Isbell, Young, & Harcourt, 2012; Simon, Morris, & Smith, 2007) and those in which women are a numerical minority (e.g., Casadevall & Handelsman, 2014; Roberts & Verhoef, 2016) have concerns about representation.

Research on underrepresentation is motivated by broader concerns that members of historically disadvantaged and/or underrepresented groups leave many academic fields at disproportionate rates at more senior ranks, a pattern called the “leaky pipeline” (Pell, 1996). This pattern has been identified in science, technology, engineering, and mathematics (STEM) fields (Hill, Corbett, & St. Rose, 2010) and within American psychology (e.g., American Psychological Association [APA] Task Force on Women in Academe, 2000). Recent data on psychology as a field suggest that gender differences in promotion from assistant onward may no longer exist (Box-Steffensmeier et al., 2015; Ginther & Kahn, 2014), but it is unclear whether such patterns hold for individual specialties, such as social psychology. Women enter social psychology at a higher rate than men but are underrepresented at the professorial levels. For example, from 1994 to 2014, women earned on average 63% of social psychology PhDs, and in all but three of those years (1995, 2001, and 2002), women earned more than 60% of social psychology PhDs (National Science Foundation, 1994-2014). Given this data, assistant and associate professor percentages for 2014 should be more than 60% if men’s and women’s careers progressed at equal rates (e.g., Ginther & Kahn, 2014). However, in 2014 only 54.1% of assistant professors, 45.8% of associate professors, and 28.8% of full professors in social/personality psychology in U.S. graduate departments of psychology were women (Wicherski, Hamp, Christidis, & Stamm, 2014).

Access to Resources

Analysis of conference speakers illustrates how women are represented at the conference and in the field. Analysis of the connections between speakers illustrates the access that women versus men have to different resources (Burt, Kilduff, & Tasselli, 2013). Using co-occurrences on symposia as a measure of presenters’ connections to each other (i.e., Professor A and B are considered connected if they presented in the same symposium), social network analysis (SNA) can be used to examine how men’s and women’s professional networks provide them with access to resources. One relevant network characteristic, structural holes, is the degree to which people are members of overlapping networks and provides a means of measuring access to unique resources. For example, imagine Professor A was in one symposium with Professor B, a neuroscientist, and in another symposium with Professor C, a developmental psychologist, and Professors B and C never appeared together. Professor A’s network would be characterized as having a structural hole, because Professor B and C are not connected. A network with more structural holes provides greater access to different and unique kinds of information and resources (Burt et al., 2013). Moreover, as the bridge between two nonoverlapping networks (i.e., Professor B’s and Professor C’s networks), Professor A can serve as a gateway for information and is central to the research relationship. Such networks are associated with greater creativity and innovation (Perry-Smith, 2006), such as receiving more patents (Wang, Rodan, Fruin, & Xu, 2014).

In addition to analysis of overlap between the members of an individual’s network, analysis of the number of connections a speaker has with other speakers, known as degree centrality, indicates the extent of their network and their position within the field (Borgatti & Everett, 2006). A large network can be beneficial to career success. For example, if excluded from men’s networks, women faculty may face challenges in their professional socialization and advancement, especially when men hold most high-status positions (e.g., Brass, 1985). Limited access to professional networks can also decrease women’s access to information and support (Ibarra, 1993; Rose, 1985; Xu & Martin, 2011).

Current Research

The present work examines gender representation among speakers and chairs in peer-reviewed symposia at the annual SPSP conference between 2003 and 2015. We drew on a variety of theories and empirical findings to generate a series of testable hypotheses. First, we considered gender representation generally. Then, drilling down, we examined how different symposia characteristics were associated with differences in women’s representation (i.e., chair gender, symposia topic). Next, stepping back, we examined women’s representation compared with various base-rate measures, and whether the pattern of representation changed over time. Finally, taking a broader look at the connections between speakers across all years, we examined the characteristics of women’s research networks, as a proxy for their access to resources and position within the broader fields of social and personality psychology. In the sections that follow, we provide the hypotheses generated at each stage of analysis and the rationale for each hypothesis.

Representation in General and in Context

Hypothesis 1a (H1a): There are more men speakers than women speakers in symposia.

Hypothesis 1b (H1b): There are more men chairs than women chairs of symposia.

Hypothesis 2 (H2): Women invited speakers are of higher status than men invited speakers.

Based on patterns of women’s participation in meetings of other academic societies (e.g., Casadevall & Handelsman, 2014; Isbell et al., 2012; Simon et al., 2007), as well as of women’s authorship in psychology (Brown & Goh, 2016; Ceci, Ginther, Kahn, & Williams, 2014; Cikara, Rudman, & Fiske, 2012), we predicted that there would be more men in speaking and chair roles in symposia than women (H1a and H1b). In addition, as research suggests that members of minority or stigmatized groups must be exceptional performers to succeed (Biernat, Manis, & Nelson, 1991), we used academic rank as a proxy for experience and status and predicted that women would need to be of higher rank to be invited to present (H2).

Hypothesis 3 (H3): There are more women invited speakers in symposia with women chairs.

Hypothesis 4 (H4): Gender representation differs by topic.

We also predicted that symposia would have a greater percentage of women speakers if at least one chair was a woman (H3). Previous studies of academic conferences have found that symposia with only men chairs have the smallest percentage of women speakers, relative to only women chairs and a mix of men and women chairs (Casadevall & Handelsman, 2014; Isbell et al., 2012). Our prediction was also based on the extensive evidence for gender-based homophily (McPherson, Smith-Lovin, & Cook, 2001). In organizational settings, the proportion of men’s networks that is other men is greater than the proportion of women’s networks that is other women, even when the organization is relatively gender-balanced (Brass, 1985; Ibarra, 1992, 1997; Xu & Martin, 2011).

Within personality and social psychology, there are many topic areas, although the boundaries between them may be permeable. Given that women, relative to men, are more likely to endorse communal goals and prefer domains that feature a focus on others, tenderness, and warmth (Diekman, Clark, Johnston, Brown, & Steinberg, 2011), we predicted that topics more associated with communal values that are stereotypically female (e.g., collaborative, prosocial) would have a greater representation of women speakers than topics more associated with stereotypically male and agentic qualities (e.g., competitive, quantitative).

Underrepresentation and Bias

Hypothesis 5a (H5a): Women are underrepresented as speakers in accepted symposia.

Hypothesis 5b (H5b): Women are underrepresented as speakers in submitted symposia.

Hypothesis 5c (H5c): Women are underrepresented as chairs in submitted symposia.

Hypothesis 5d (H5d): Gender representation in symposia influences acceptance.

It is important to determine whether women’s level of representation constitutes underrepresentation by comparing it with the percentage of women in relevant comparison populations. Studies of other academic conferences have found that women are often underrepresented as speakers relative to their presence as attendees (Casadevall, 2015; Casadevall & Handelsman, 2014) and to their membership in relevant organizations (Isbell et al., 2012; Simon et al., 2007). As a result, we predicted that women would be underrepresented as speakers in accepted symposia relative to their membership in SPSP (H5a).

In addition, we examined women’s representation in accepted and submitted symposia (H5a-H5c) separately for different academic ranks. Given the decreasing percentages of women at increasing academic ranks in social and personality psychology (e.g., Wicherski et al., 2014), if most presenters are at higher academic ranks, women could be underrepresented in the aggregate but well-represented at some academic ranks (i.e., Simpson’s paradox; see Bickel, Hammel, & O’Connell, 1975, for an example involving gender representation).

Finally, to explore potential explanations for gender representation differences, we also examined women’s representation as speakers and chairs across all submitted symposia (H5b and H5c) and compared representation in accepted and rejected symposia (H5d). If women are underrepresented across submitted symposia but have similar representation in accepted and rejected symposia, this would suggest that gender representation differences are caused by differences in submission rates. If women are better represented in rejected than accepted symposia, this would be consistent with a bias against symposia with more women speakers (e.g., Knobloch-Westerwick, Glynn, & Huge, 2013; Roberts & Verhoef, 2016) and/or decreases in symposium quality with more women speakers.

Patterns Over Time

Hypothesis 6 (H6): Women’s representation has changed over time.

One explanation for women’s underrepresentation in various fields and positions has been historical limitation of women’s access to education and opportunities. According to this argument, gender equity will be achieved over time as increasing numbers of women attend college, pursue advanced degrees, and enter academia. This hypothesis is in line with recent data documenting reduced or eliminated gender differences in promotion to different academic ranks (e.g., Box-Steffensmeier et al., 2015; Ceci et al., 2014; Ginther & Kahn, 2014), and with improvements over time in women’s representation among presenters at conferences in other fields (Simon et al., 2007). To test whether women’s representation increased over time, we looked for temporal effects for Hypotheses H1a, H1b, H2, and H3.²

Access to Resources

Hypothesis 7a (H7a): Women presenters have less access to nonoverlapping resources (i.e., their networks have fewer structural holes) than men presenters.

Hypothesis 7b (H7b): Women presenters are less connected to social resources (i.e., have lower degree centrality) than men presenters.

Previous research has found that men in organizations are connected to more people than are women (e.g., Ibarra, 1992; Mehra, Kilduff, & Brass, 1998). In addition, some research suggests that women’s social networks are denser, consisting of more interconnected people and therefore having fewer structural holes, than men’s networks (e.g., Mehra et al., 1998; Moore, 1990; but see Ibarra, 1993, 1997, for an opposing viewpoint). Therefore, we predicted that women’s research networks, as defined by co-occurrence on symposia, would have fewer structural holes than men’s. Researchers’ networks were considered to have more structural holes to the extent that their copresenters did not present with each other and did not share other connections. We also predicted that, using a measure of degree centrality, women’s networks would reveal them to be less central to the field than men. Researchers were considered more central as they copresented with a greater number of different people.

Method

Data Collection

All conference programs from 2003 to 2013 were downloaded from the SPSP website. Spreadsheets corresponding to the 2014 and 2015 conferences, which included both accepted and rejected symposia, were obtained from the program chairs. Symposia that were submitted but rejected were unavailable for other years. SPSP had its first stand-alone conference in 2000, and the reviewing procedures and submission pool for the early years of the conference may not be representative. For the downloaded programs, research assistants manually copied symposia titles, names of chairs, discussants, and first authors of presentations, along with their affiliations, into a spreadsheet. Presidential symposia, special sessions, and data blitzes were excluded from analyses to maintain our focus on peer-reviewed symposia organized by chairs rather than invited by a committee.

The resulting database included all accepted symposia from 2003 through 2013, and all accepted and rejected symposia from 2014 to 2015. Each entry (N = 6,415) corresponded to a role: symposium chair (n = 1,813, 28.3%), speaker (n = 4,476, 69.8%), or discussant (n = 126, 2.0%).³ During this time period, SPSP restricted individuals from having two or more speaking roles in the same year but not from holding other combinations of roles (e.g., chair and speaker), so the same person could be represented multiple times within 1 year. These data were cleaned to standardize participant names (so the same individual was always listed under the same name) and institutional affiliations.

Data Preparation

People

Research assistants and the first two authors coded each individual in the database for gender (0 = male) and academic rank at the time of the presentation (see online supplemental materials for coding instructions). They drew upon personal knowledge, online information, the Social Psychology Network, and personal networks, with decisions based on names, photographs, and pronouns used in biographical sketches. Gender identification at the time of the conference was used. Of the 6,415 entries, 3,452 (53.8%) were filled by individuals identified as men, 2,962 (46.2%) by individuals identified as women, and one by someone whose gender could not be identified. These represented 2,612 unique individuals: 1,329 men (50.9%), 1,282 women (49.1%), and one individual whose gender could not be identified. Using the U.S. academic ranking system as a framework, rank was coded as undergraduate (n = 4, <0.1%), graduate student (n = 1,574, 24.5%), postdoc/lecturer (n = 631, 9.8%), assistant professor (n = 1,729, 27.0%), associate professor (n = 962, 15.0%), full professor (n = 1,420, 22.1%), or nonacademic (n = 54, 0.8%). Twenty-six positions (0.4%) were academic but did not fit these classifications (e.g., lab coordinator), and the position for 15 entries (0.2%) could not be identified. Undergraduate research assistants completed a first round of coding, flagging ambiguous cases. The first two authors created equivalents for different European systems and evaluated titles such as “research scientist” or “research fellow” on a case-by-case basis. In some cases, academic rank was extrapolated from other data available on CVs (e.g., years since earning PhD, time in an academic position). Individuals could appear multiple times in the database (M = 2.46 times, SD = 2.32, range = 1-29), with their rank changing from year to year.

Symposia

The average accepted symposium had 3.98 speakers (SD = 0.36, range = 2-5) and 1.58 chairs (SD = 0.51, range = 1-4). We used keywords to classify each symposium by topic area. Keywords were first used by SPSP in 2009 and were created by one of the program chairs, (W. Gardner, personal communication, August 27, 2016). Keywords were initially used to categorize poster presentations, so they could be grouped by topic area. Keywords were also used to ensure that symposia on similar topics would not be scheduled at the same time. Program chairs coded the symposia idiosyncratically. The keywords were passed from one program chair to the next without instructions for coding. Chairs added and eliminated keywords as they saw fit (S. Srivastava, personal communication, August 24, 2016).

For the purpose of this analysis, the list of keywords used in 2014 was used to classify all symposia. Each symposium was assigned at least one keyword. Keywords were assigned based on a symposium’s title and abstract as well as the abstracts of its individual presentations. In the first round of coding, undergraduate psychology student research assistants coded the symposia, and all codings were reviewed by the first or second author. Any symposia that were ambiguous were coded by the first two authors, based on their experience in the field and the rules detailed in the online supplemental materials, as well as symposia titles and abstracts. Of the accepted symposia, 15 (1.8%) were only assigned one keyword. Rules for classification and new keywords were created when several similar symposia did not align with existing keywords. For example, the original keyword list did not include “social cognition” but did include “person perception,” the origin of the field of social cognition. Therefore, this keyword was expanded to include social cognition. If a single symposium did not clearly align with any keywords or with other symposia, it was classified as “other.” We tried to avoid creating overly narrow categories while also keeping the categories meaningful.

Results

Analysis Strategy

Except where indicated, analyses were conducted at the level of symposium because during the studied time period, symposia, not individual talks, were submitted to the SPSP conference and acceptance decisions were made only for entire symposia. Except for testing of H5b to H5d, only accepted symposia were examined (n = 840). Throughout, the term speaker refers only to first authors of talks presented in a symposium and the term presenters refers to speakers and chairs of symposia. When used in analyses, academic rank was coded as 0 = undergraduate student, 1 = graduate student, 2 = postdoctoral fellow/lecturer, 3 = assistant professor, 4 = associate professor, and 5 = full professor. Nonacademic positions were coded as 6 and other was coded as 7. Table 1 presents descriptive information by year of conference.

Table 1.

Descriptive Statistics of Sample by Year of Conference.

Year of conference	Number of symposia	Average % of women speakers in symposia	Average % of women invited speakers in symposia	Average % of women chairs in symposia	Median rank of women invited speakers	Median rank of men invited speakers
2003	36	37.4	36.4	39.8	3	4
2004	39	40.0	40.4	39.7	3	5
2005	40	38.0	37.0	35.0	3	4
2006	51	38.4	38.2	36.3	3	4
2007	61	40.9	37.1	43.4	3	4
2008	66	49.8	49.0	50.5	3	4
2009	73	37.4	33.9	47.7	3	4
2010	75	42.0	41.7	40.9	3	4
2011	74	42.7	39.0	52.0	3	4
2012	75	44.7	39.7	50.0	3	3
2013	90	44.1	43.4	45.6	3	3
2014	79	53.2	51.5	57.6	3	3
2015	81	50.2	43.9	56.2	3	3

Note. Only accepted symposia are included. Only invited speakers with an academic position are included in the rank statistics. Rank was coded as 0 = undergraduate student, 1 = graduate student, 2 = postdoctoral fellow/lecturer, 3 = assistant professor, 4 = associate professor, and 5 = full professor.

Tests of Hypotheses

H1a: There are more men speakers than women speakers in symposia.

Of the 3,344 speakers in accepted symposia, 1,879 (56.2%) were men and 1,465 (43.8%) were women. Figure 1 shows the frequency with which different percentages of women within a symposium occurred. As can be seen, a larger percentage of symposia had majority men speakers (45.1%) than had majority women speakers (28.3%). A one-sample t test revealed that the average percentage of women speakers in a symposium (M = 43.8%, SD = 28.7%) differed significantly from 50%, t(839) = −6. 26, p < .001, 95% CI_difference [4.26%, 8.15%], d = 0.22. Thus, there were more men than women in speaking roles in the average symposium.

Figure 1.

Representation of women as speakers in accepted symposia, 2003-2015.

One could reason that with increasing opportunities to speak (e.g., having five speakers rather than three), there might be greater diversity of speakers. To test this, we ran separate chi-square analyses for each symposium size comparing the distribution of women’s representation in symposia against an equal distribution. There were fewer women than men speakers in symposia, regardless of size (see Table 2). Among symposia with three, χ²(3) = 9.10, p = .028, Cramer’s V = .22; four, χ²(4) = 130.13, p < .001, Cramer’s V = .21; and five, χ²(5) = 15.68, p = .008, Cramer’s V = .26, speakers, the distribution significantly differed from an equal distribution (only one symposium had two speakers, so no analysis was run for this amount). The correlation between the number of speakers on a symposium and the percentage of women speakers was nonsignificant, r(838) = .006, p = .86, Fishers’s Z_r = 0.0006, also indicating that increasing the number of speaker slots on a symposium did not increase women’s participation.

H1b: There are more men chairs than women chairs in symposia.

Table 2.

Representation of Women as Speakers Across Accepted Symposia by Size of Symposium, 2003-2015.

Size of symposium	Number of women speakers	Number of symposia	% of symposia
Three speakers (n = 61)	0.00	15	24.6
	1.00	18	29.5
	2.00	22	36.1
	3.00	6	9.8
Four speakers (n = 731)	0.00	110	15.0
	1.00	206	28.2
	2.00	223	30.5
	3.00	136	18.6
	4.00	56	7.7
Five speakers (n = 47)	0.00	4	8.5
	1.00	11	23.4
	2.00	14	29.8
	3.00	11	23.4
	4.00	6	12.8
	5.00	1	2.1

A one-sample t test revealed a marginally significant tendency for symposia (M = 47.1%, SD = 43.1%) to have fewer than 50% women chairs, t(839) = −1.945, p = .052, 95% CI_difference [−0.03%, 5.8%], d = 0.07. Most symposia had one (n = 360) or two (n = 473) chairs, with the remaining few having three (n = 6) or four (n = 1). Because having multiple chairs provides more opportunities to have a woman chair, we separately examined symposia with one versus multiple chairs. For symposia with one chair, a man (n = 191, 53.1%) was as likely as a woman (n = 169, 46.9%) to be chair, χ²(1) = 1.34, p = .246, Cramer’s V = .004. Symposia with multiple chairs were more likely to have a mix of men and women chairs⁴ (n = 212, 44.2%) than all men chairs (n = 147, 30.6%) or all women chairs (n = 121, 25.2%), χ²(2) = 27.46, p < .001, Cramer’s V = .03. There was no relationship between the number of chairs in a symposium and the percentage of women chairs, r(838) = −.003, p = .934, Fishers’s Z_r = 0.0003.

H2: Women invited speakers are of higher status than men invited speakers.

We examined the academic rank of invited speakers to determine whether women had to be more experienced in the field or have higher status than men to be invited to present. Nonchair speakers, by definition, were invited to present. Speakers whose academic rank was not identified, whose rank was classified as “other,” or who held nonacademic positions were not included in this analysis. This left 2,381 invited speakers, 1,389 men (58.3%) and 992 women (41.7%).

Contrary to the hypothesis, a Mann–Whitney test revealed that women invited speakers (Mdn = 3.00) were of lower rank than men invited speakers (Mdn = 4.00), U = 527,335.00, Z = −10.065, p < .001, r = .21. The distributions of men and women invited speakers were not equal at different career stages, χ²(4) = 104.72, p < .001, Cramer’s V = .10. As shown in Figure 2, congruent with the fact that a majority of individuals receiving social psychology PhDs during this time period were women, invited speakers who were graduate students were more likely to be women than men, t(418) = 3.41, p = .001, d = 0.17, and invited speakers who were postdocs were just as likely to be women as to be men, t(184) = 0.37, p = .71, d = 0.03. However, in all faculty categories (i.e., assistant, associate, and full professor), invited speakers were more likely to be men, with the gender gap growing from assistant professors, t(661) = −3.53, p < .001, d = 0.14; to associate professors, t(426) = −4.19, p < .001, d = 0.20; to full professors, t(687) = −12.49, p < .001, d = 0.48.

Figure 2.

Distribution of invited speakers in accepted symposia by gender and academic rank, 2003-2015.

These findings highlight the importance of academic rank in understanding women’s representation at the SPSP conference. They are best interpreted in conjunction with information about the percentage of women at different academic ranks in the broader population, and we continue explorations of academic rank in testing Hypothesis 5.

H3: There are more women invited speakers on symposia with women chairs.

Of the 839⁵ symposia that had nonchair speakers, 337 (40.2%) had all men chairs, 212 (25.3%) had a mix of men and women chairs, and 290 (36.4%) had all women chairs. A Kruskal–Wallis test revealed that the gender of the chairs was related to the gender of the invited speakers, p < .001. When all chairs were men, 33.8% (SD = 29.4%) of the invited speakers in a symposium were women. This percentage increased as the proportion of women chairs increased, to 42.5% (SD = 32.8%) when there was a mix of men and women chairs and to 49.6% (SD = 33.0%) when all chairs were women.⁶

As shown in Figure 3, a chi-square test examining the hypothesis that the distributions of the proportion of women invited speakers in symposia were the same across the different proportions of women chairs revealed that women were more likely to be invited to speak when there was at least one woman chair, χ²(18) = 84.06, p < .001, Cramer’s V = .07.⁷

H4: Gender representation differs by topic.

Figure 3.

Representation of women as invited speakers in accepted symposia by gender of chairs, 2003-2015.

We predicted that women’s representation would differ by topic area. We used one-sample t tests to compare the average percentage of women speakers in symposia associated with each keyword to 50%. (As the keyword “other” does not represent a substantive topic, we will not discuss results for it, though it is included in Table 3 for reference.) As shown in Table 3, the percentage of women deviated significantly from 50% for 15 of the 30 substantive keywords. Only three topics had on average more women than men speakers in their associated symposia. For the 15 most frequently occurring substantive keywords, statistically significant differences were found 8 times, with six topics (social cognition, motivation, methods, self-regulation, norms and social influence, and judgment and decision making) having majority men speakers, and two topics (close relationships and stereotyping/prejudice) having majority women speakers.

Table 3.

Average Percentage of Women Speakers in Accepted Symposia by Keyword, 2003-2015.

Keyword	Number of symposia listing keyword	Average % of women speakers	One sample t statistic	p value	Cohen’s d	Rated masculinity	Rated femininity
Person perception/social cognition	141	39.6	−4.86	<.001	0.41	3.84	2.93
Self/identity	121	45.2	−1.81	.07	0.16	3.23	3.15
Other	114	34.3	−6.09	<.001	0.57
Close relationships/interpersonal	100	60.0	3.76	<.001	0.38	3.25	3.64
Emotion	98	46.9	−1.05	.30	0.11	3.49	3.09
Motivation/goals	82	36.7	−4.40	<.001	0.49	3.37	3.04
Stereotyping/prejudice	80	57.3	2.28	.03	0.26	3.48	3.57
Methods/statistics/research integrity	70	29.8	−6.49	<.001	0.78	4.15	2.79
Culture	68	53.0	0.81	.42	0.10	2.75	3.72
Self-regulation	67	35.5	−4.52	<.001	0.55	3.56	2.96
Intergroup relations	59	51.4	0.41	.68	0.05	3.35	3.62
Mental health/well-being	58	54.8	1.29	.20	0.17	3.07	3.73
Norms and social influence	58	40.2	−2.89	.01	0.38	3.14	3.30
Social neuroscience	55	43.5	−1.63	.11	0.22	4.09	2.78
Groups/intragroup processes/power	53	52.6	0.67	.50	0.09	3.49	3.34
Judgment/decision-making	51	33.0	−4.77	<.001	0.67	3.94	2.89
Attitudes/persuasion	47	34.9	−3.99	<.001	0.58
Personality processes/traits	45	32.7	−4.76	<.001	0.71
Morality	41	29.9	−5.26	<.001	0.82
Physical health	32	57.0	1.55	.13	0.27
Prosocial behavior	32	35.2	−2.55	.02	0.45
Individual differences	30	46.9	−0.74	.47	0.13
Politics	26	31.2	−4.61	<.001	0.90
Evolution	23	46.7	−0.53	.60	0.11
Belonging/rejection	19	61.8	2.03	.06	0.46
Gender	18	66.7	2.29	.04	0.54
Field research/interventions	17	49.7	−0.04	.97	0.01
Applied social psychology	17	44.8	−0.76	.46	0.19
Diversity	15	62.2	1.63	.13	0.42
Religion/spirituality	15	30.0	−2.70	.02	0.70
Aggression/antisocial	13	36.5	−2.11	.06	0.59

Note. Keywords are listed in order of descending number of symposia listing that keyword. In all, 825 symposia were associated with two different keywords and are represented twice; 15 were associated with only one keyword and are represented once. The one-sample t tests test the average percentage of women speakers against 50%. Rated masculinity is the average of ratings of competitive and quantitative, and rated femininity is the average of ratings of collaborative and prosocial, all on 5-point scales. These ratings were only collected for the 15 most frequently occurring substantive keywords (i.e., excluding “other”; see main text for details).

We predicted that perceptions of a topic area as stereotypically masculine versus feminine would be associated with the percentage of women speakers (Diekman et al., 2011). To test this, we conducted a brief study examining social and personality psychologists’ perceptions of the 15 most frequently occurring substantive keywords (which were each associated with at least 50 symposia). Social and personality psychologists were recruited via Facebook and the SPSP Connect Forum for a study of “Perceptions of Social and Personality Psychology.” Participants (N = 101) completed a Qualtrics survey online in which they were presented with a list of these 15 keywords (presented in random order) and asked to rate how well each of six adjectives (competitive, collaborative, quantitative, creative, high status, prosocial) described each of the areas. Participants responded using a 5-point scale (1 = not at all descriptive, 2 = a little descriptive, 3 = somewhat descriptive, 4 = mostly descriptive, 5 = very descriptive). Ratings of competitive and quantitative, r(13) = .55, p = .03, were averaged for a measure of how stereotypically masculine a topic was, and ratings of prosocial and collaborative, r(13) = .60, p = .02, were averaged for a measure of how stereotypically feminine a topic was (see Table 3, for means by keyword). These two measures were negatively correlated, r(13) = −.82, p < .001. We then tested whether gender representation was associated with perceptions of topic areas. Topics rated as more stereotypically feminine had a higher percentage of women speakers, r(13) = .86, p < .001, and topics rated as more stereotypically masculine had a lower percentage of women speakers, r(13) = −.60, p = .02. There were no significant correlations between the percentage of women speakers and perceptions of how much a topic was described as high status, r(13) = −.34, p = .21, or creative, r(13) = −.27, p = .33. Although these results cannot speak to causality, they do provide one framework for understanding differences in gender representation.

Underrepresentation

Women’s underrepresentation relative to the population in the field was examined among three groups: speakers in accepted symposia, speakers in submitted symposia, and chairs in submitted symposia. We used demographic information about 2015 SPSP membership as a comparison baseline. Reliable demographic data on SPSP membership for other years were not available, as response rates to voluntary surveys were low (less than 60%) or data were not maintained (C. Rummel, personal communication, August 24, 2016). In 2015, SPSP reported that women constituted 56.25% of the SPSP membership.

For each group, we examined underrepresentation at each academic rank. To establish base rates, we used several sources. For graduate students, we used data from years 2003 to 2014 (2015 data were not yet available) of the annual census of U.S. social psychology PhD recipients by the National Science Foundation (the Survey of Earned Doctorates; National Science Foundation 1994-2014), as we could not obtain data on the gender distribution among social and personality graduate students. For assistant, associate, and full professors, we used data from years 2004 to 2014 (2003 and 2015 data were unavailable) of the annual survey of U.S. graduate departments⁸ by the APA Center for Workforce Studies (the Faculty Salary Survey; American Psychological Association Center for Workforce Studies 2004-2014). By averaging across the available years of data for a given academic rank, we arrived at base rates of 64.7% women for graduate students, 50.8% women for assistant professors, 45.8% women for associate professors, and 28.1% women for full professors. Postdoctoral researchers and lecturers were excluded from these comparisons because data on base rates by gender were not available.

Women’s representation in accepted versus rejected symposia was also compared, to determine whether acceptance decisions were related to the percentage of women speakers in a symposium.

H5a: Women are underrepresented as speakers in accepted symposia.

A one-sample t test comparing the average percentage of women speakers in 2015 accepted symposia (n = 81, M = 50.2%, SD = 28.8%) with the percentage of women SPSP members in 2015 revealed a marginally significant difference such that women were underrepresented relative to the SPSP membership, t(80) = −1.89, p = .063, 95% CI_difference [0.03%, 12.4%], d = 0.21.

Next, underrepresentation of women as speakers in accepted symposia by rank, relative to the base rates listed above, was examined. Women were underrepresented among graduate student speakers, M = 58.5%, SD = 49.3%, t(708) = −3.33, p = .001, d = 0.13; assistant professors speakers, M = 44.9%, SD = 49.7%, t(961) = −3.67, p < .001, d = 0.12; and associate professors speakers, M = 39.3%, SD = 48.9%, t(518) = −3.03, p = .003, d = 0.13. However, women were well-represented among full professor speakers, M = 29.0%, SD = 45.4%, t(815) = 0.59, p = .55, d = 0.02. In short, women were underrepresented as speakers in accepted symposia at every academic rank, relative to the population, except at the highest rank of full professor.

H5b: Women are underrepresented as speakers in submitted symposia.

A one-sample t test revealed that the average percentage of women speakers (M = 51.8%, SD = 27.1%) in all submitted symposia in 2015 (n = 244) differed significantly from the percentage of women SPSP members, t(243) = −2.56, p = .01, 95% CI_difference [1.0%, 7.8%], d = 0.16.

Next, we examined underrepresentation of women as speakers in all submitted symposia in 2014 and 2015 by rank. Women were well-represented among graduate student speakers, M = 60.8%, SD = 48.9%, t(548) = −1.85, p = .06, d = 0.08; assistant professor speakers, M = 51.8%, SD = 50.0%, t(443) = 0.42, p = .67, d = 0.02; and associate professor speakers, M = 46.4%, SD = 50.0%, t(251) = 0.20, p = .84, d = 0.01. However, women were overrepresented among full professor speakers, M = 39.5%, SD = 49.0%, t(285) = 3.94, p < .001, d = 0.23.

H5c: Women are underrepresented as chairs in submitted symposia.

To test whether women were underrepresented as chairs, we compared, for 2015, the average percentage of women chairs⁹ (M = 52.9%, SD = 41.2%) in submitted symposia (n = 244) with women’s membership in SPSP that year. A one-sample t test revealed that these two percentages did not differ significantly, t(243) = −1.28, p = .20, 95% CI_difference [−1.8%, 8.6%], d = 0.08. That is, in 2015, women were chairing and submitting symposia at a rate proportional to their percentage of the SPSP membership and were not underrepresented as applicants.

Next, we examined underrepresentation of women as chairs in all submitted symposia in 2014 and 2015 by rank. Women were underrepresented among assistant professor chairs, M = 42.2%, SD = 49.5%, t(191) = −2.41, p = .02, d = 0.17, but well-represented among graduate student chairs, M = 63.1%, SD = 48.4%, t(232) = −0.51, p = .61, d = 0.03, and associate professor chairs, M = 46.1%, SD = 50.1%, t(114) = 0.06, p = .95, d = 0.01. Women were overrepresented among full professor chairs, M = 50.9%, SD = 50.2%, t(109) = 4.76, p < .001, d = 0.45. As in the previous analysis of speakers in submitted symposia, women were overrepresented at the highest rank as chairs in submitted symposia. In addition, they were underrepresented at the rank of assistant professor.

H5d: Gender representation in symposia influences acceptance.

We compared gender representation in accepted (n = 160) and rejected (n = 291) symposia using data from 2014 and 2015. The average percentage of women speakers in accepted symposia (51.7%, SD = 28.4%) and rejected symposia (53.1%, SD = 27.5%) did not differ significantly, t(449) = 0.53, p = .60, 95% CI_difference [−3.9%, 6.9%], d = 0.003. Thus, for 2014 and 2015, there was no evidence consistent with bias against women speakers at the level of accepting versus rejecting symposia.

To examine whether the gender of a symposium’s chairs influenced whether it was accepted, a chi-square test compared the gender mix of chairs in accepted and rejected symposia (see Table 4). The distribution of women chairs was the same across accepted and rejected symposia, χ²(2) = 2.95, p = .23, Cramer’s V = .06. If anything, a marginally significant correlation suggests that the likelihood of a symposia being accepted increased with the percentage of women chairs, r(449) = .08, p = .09.

H6: Women’s representation has changed over time.

Table 4.

Distribution of the Presence of Women Chairs in Accepted and Rejected Symposia, 2014-2015.

	Gender of chairs			Total
	All men	One man and one women	All women	Total
Accepted symposia	44	50	66	160
Rejected symposia	102	87	102	290
Total	146	137	168	451

To understand how patterns of representation changed over time, we examined some of the previous hypotheses using longitudinal analyses.

We previously found that there were more men than women speakers (H1a) and a tendency toward more men than women chairs (H1b). To test whether these patterns changed over time, we conducted one-way ANOVAs on the average percentage of women speakers and chairs in symposia with year as a between-subjects factor. As can be seen in Table 1, over time, the increase in the percentage of women speakers in symposia was significant, F(12, 827) = 2.19, p = .01, η² = .03; for linear contrast, p < .001, η² = .01, and the increase in the percentage of women chairs was marginally significant, F(12, 827) = 1.73, p = .06, η² = .02; for linear contrast, p < .001, η² = .01.

We previously found that among invited speakers, women were of lower academic rank than men (H2). To examine this pattern over time, we conducted a 2 (gender) × 13 (year) between-subjects ANOVA on the average rank of invited speakers. In addition to confirming the previous finding that men were of higher rank than women, F(1, 2355) = 97.14, p < .001, η² = .04, a significant main effect of year emerged, F(12, 2355) = 3.31, p < .001, η² = .02. A linear contrast revealed that, over time, the average rank of invited speakers decreased, p < .001, η² = .02, as can be seen in Table 1. There was no evidence that rank differences between men and women invited speakers changed over time, as the two-way interaction was nonsignificant, F(12, 2355) = 0.44, p = .95, η² = .002.

In a more liberal test of this question, we ran a series of one-way ANOVAs looking at the effect of year on the percentage of women among invited speakers at each of the five academic ranks. There was no significant effect of year on the percentage of women among invited speakers who were graduate students, F(12, 406) = 0.71, p = .75, η² = .02; postdocs, F(12, 172) = 0.38, p = .97, η² = .03; assistant professors, F(12, 649) = 0.51, p = .91, η² = .01; associate professors, F(12, 414) = 0.85, p = .60, η² = .02; or full professors, F(12, 675) = 1.47, p = .13, η² = .03. Again, we found no evidence that rank differences between men and women invited speakers changed over time.

We previously found that symposia with more men chairs had fewer women invited speakers (H3). To test whether this association changed over time, we conducted a 3 (chairs’ gender: all men, mix of men and women, and all women) × 13 (year) between-subjects ANOVA on the average percentage of women invited speakers in symposia. We replicated the earlier effect of chairs’ gender, F(2, 800) = 16.17, p < .001, η² = .02. This pattern was not moderated by year, F(24, 800) = 0.49, p = .98, η² = .005, and there was no significant main effect of year, F(12, 800) = 1.37, p = .17, η² = .007. The pattern in which symposia that had more men chairs had fewer women invited speakers did not appear to change over time.

To summarize, women’s representation as speakers and chairs in accepted symposia has increased over the 13-year period examined. However, gender differences in the academic rank of invited speakers in a symposium remained steady, as did the association between chair gender and invited speaker gender.

H7a: Women presenters have less access to nonoverlapping resources (i.e., their networks have fewer structural holes) than men presenters.

H7b: Women presenters are less connected to social resources (i.e., have lower degree centrality) than men presenters.

To test these hypotheses, we used SNA (Wasserman & Faust, 1994). The analysis examined all co-occurrences on symposia between all presenters on accepted symposia from 2003 to 2015. First, we examined the extent to which presenters had structural holes in their copresenting network, indicating their access to nonoverlapping resources. The extent of structural holes was calculated using the constraint model (see Wang et al., 2014, for relevant formulas), taking into account the size of a person’s copresenting network and the degree to which that person’s copresenters were directly or indirectly connected to one another. Women (n = 886) had fewer structural holes in their copresenting networks (M = 1.53, SD = 0.23) than men (n = 1,014; M = 1.57, SD = 0.24), t(1898) = −3.31, p < .001, 95% CI_difference [−0.056, −0.014], d = 0.17. Women’s copresenting networks offered access to fewer unique resources than did men’s.

Next, we examined the number of different resources and connections presenters had using degree centrality, the number of co-occurrences on symposia with different people. Women had lower degree centrality (M = 6.23, SD = 5.57) than men (M = 7.15, SD = 6.01), t(1898) = −3.44, p < .001, 95% CI_difference [−1.44, −0.39], d = 0.16. Thus, women were less central in the field and had connections to fewer resources than did men.

As the women in our sample are on average of lower academic rank than the men, differences in social networks between genders might be explained by differences in rank. To test this possibility, we regressed a presenters’ highest academic rank¹⁰ while at the SPSP conference, the presenter’s gender, and the interaction of the two onto the measures of structural holes and degree centrality. Only presenters who had been in an academic position (e.g., graduate student) while presenting (n = 1,880) were included in these analyses; all nonacademic positions were coded as missing values. Academic rank was positively related to the number of structural holes, β = .03, SE = 0.01, t(1876) = 6.81, p < .001, but gender no longer had a significant effect, β = −.03, SE = 0.02, t(1876) = −1.24, p = .22, and their interaction was not significant, β = .01, SE = 0.02, t(1876) = 0.92, p = .36. Similarly, academic rank was positively related to degree centrality, β = .84, SE = 0.13, t(1876) = 6.69, p < .001, but gender no longer had a significant effect, β = −1.00, SE = 0.60, t(1876) = −1.68, p = .09, and their interaction was not significant, β = .25, SE = 0.19, t(1876) = 1.36, p = .17. This suggests that the previously found differences in the networks of men and women presenters could be accounted for by women’s lower average academic rank.

Discussion

Social psychology is the scientific study of people and how they are affected by their environments. As such, the science should be informed by a diversity of voices and perspectives. The present research considers one facet of diversity by examining gender representation in SPSP symposia from 2003 to 2015, as well as exploring the characteristics of symposia, such as topic area and chair gender.

The Big Picture: Women Are Underrepresented but Not Always

At the highest level of analysis, the overall picture was mixed. There were reasons for optimism and areas for improvement. On one hand, across all symposia, there were more men speakers than women speakers. When walking into a symposium, 28% of the time, audience members saw only one woman presenting her research, and 15.5% of the time, no women at all. In addition, although many junior women participated in SPSP symposia, women were underrepresented at every rank except the full professor level, relative to their percentage in the population. Finally, men chairs were more likely to invite other men to speak in their symposia than to invite other women. This last effect is notable, given that there were previously only two ways to speak at the SPSP conference: (a) organize a symposium or (b) be invited to speak in someone else’s symposium.

On the other hand, the disparities were not large, as can be seen from our effect sizes, and the representation of women has increased over time. In 2015, women were organizing and submitting symposia as chairs at a rate proportional to their membership in SPSP. In addition, in the years tested, having more women speakers or women chairs did not affect the likelihood that a symposium was accepted. The latter analysis is based on only 2 years of data, and the programs from those years are confounded with the motivations of the program chairs (in 2014, the program chairs actively attended to gender issues in creating the conference review panel and program [C. Kaiser, personal communication, April 27, 2016]), limiting the conclusions that can be drawn. However, this finding does suggest, as other fields have found (Casadevall, 2015), that conscious attention to the gender representation of panels can have meaningful effects.

Analysis of symposia by topic likewise provided a mixed picture of gender representation. Among the 15 most common topics, two (close relationships and stereotyping/prejudice) had on average significantly more women speakers than men speakers on their symposia, and seven did not significantly deviate from 50/50 representation (self/identity, emotion, culture, intergroup relations, mental health/well-being, social neuroscience, and groups/intragroup processes/power). However, six topics (social cognition, motivation, methods, self-regulation, norms and social influence, and judgment and decision making) had on average significantly more men speakers than women speakers on their symposia. We did not have base-rate information for the population of social psychologists studying each topic, so could not test underrepresentation relative to any population. However, as prolific researchers often presented on several topic areas, base rate information may not be useful. We do know that perceptions of topic areas as collaborative and prosocial were positively associated with women’s representation, and perceptions of topic areas as competitive and quantitative were negatively associated with women’s representation. Just as low participation of women and minorities in STEM disciplines is a concern, researchers in areas with low women’s participation may use these findings to consider issues of parity, as a lack of diversity in researchers may constrain the kinds of research questions that are pursued and the creativity of the field (e.g., Inbar & Lammers, 2012).

Looking at representation over time provided reasons for both optimism and concern. Representation of women is increasing over time. However, in the 13 years of conference data studied, there was no evidence that the pattern of women being underrepresented as speakers at all ranks except full professor was changing. Differences between men and women in productivity and impact (Brown & Goh, 2016; Cikara et al., 2012; Eagly & Miller, in press; Nosek et al., 2010) may explain overall gender differences in representation but cannot explain why these gender differences disappear among full professors, as men full professors still publish more and are better cited than women full professors (Eagly & Miller, in press). Another potential explanation for this pattern is that women must be more established and successful than men to be invited to join a symposium (Banaji & Greenwald, 1995; Biernat et al., 1991), although our test of Hypothesis 2 did not support this explanation. Or, less senior women may be more likely to turn down invitations to speak (Schroeder et al., 2013) because of childbearing and family obligations, more common at those ages, that may reduce their ability to attend conferences (Xu & Martin, 2011). Future research should continue to test these and other explanations.

To further understand women’s professional progress, we explored their access to professional resources through examination of their professional networks. In the present analysis, we found that men’s and women’s copresenting networks, and therefore resources available to them, differed. Women were on symposia with fewer different individuals and with individuals who likely had overlapping resources. Men, on the contrary, were on symposia with researchers who were more independent from one another, which likely increased the range of resources to which they had access. These gender differences disappeared when the lower average academic rank of women was taken into account, suggesting that these deficiencies in women’s networks may decrease and eventually disappear as more women achieve more senior academic positions. This analysis reflected how women and men interacted with their colleagues in the past; further understanding of network influence could help increase parity in the field.

Future Actions

In addition to suggesting future research directions, these results point to a number of actions that could be taken by those who convene and organize conferences. Encouragingly, the organizers of the SPSP conference have recently adopted some of these actions,¹¹ in addition to greater vigilance and attention on the part of symposia chairs and program committees. For example, in the years under review, the only way to speak at the SPSP conference was as part of a symposium, which relies heavily on having or creating a network. Such a policy may disadvantage women for several reasons. Women may be less likely to be invited as speakers because they are less central in social networks and have less diverse set of connections than men, as suggested by our copresenting network data. Furthermore, men chairs are less likely to invite women to speak than to invite men, but women chairs do not evidence a parallel preference in favor of women speakers. This means that even if women are well-represented as chairs (as was true in recent years; see Table 1), they will be underrepresented as speakers. In addition, women organizing symposia may have greater difficulty recruiting and attracting high-status speakers or speakers outside their network, than men. SPSP created opportunities for submitting single-paper oral presentations to the 2017 conference, in addition to symposia, which could address these issues by reducing reliance on social networks, reducing the role of symposia chairs as gatekeepers, and making it easier to submit a presentation.

In addition to changing the structure of the submission process, in 2015, SPSP adopted a double-blind review procedure. Prior to this, reviewers knew the identities of the submitting authors, limiting reviewers’ ability to assess symposia quality independent of biases associated with author characteristics.¹² For example, reviewers might have been biased toward individuals who were members of their social networks or were of high status. In addition, gender bias could have emerged, as research of similar quality may receive lower ratings when attributed to a woman scientist versus a man scientist (Knobloch-Westerwick et al., 2013; Roberts & Verhoef, 2016).

Finally, these analyses speak to the representation of women as speakers at the SPSP conference but could not speak to the overall representation of women at the conference (e.g., as attendees, as poster presenters) or to the representation of other minority group members. Not only was it beyond the scope of the present research to examine poster presentations, but demographic information about submissions or attendees had not been obtained or maintained by SPSP. It is heartening that in response to requests for greater transparency and revelations of unequal representation, the SPSP Executive Committee has begun to collect demographic information on membership, attendees, and submissions, which should aid future research.

Significance and Conclusion

Our data fit a larger pattern of recent work suggesting that women’s impact on psychology is less than that of men’s (Brown & Goh, 2016; Cikara et al., 2012; Eagly & Miller, in press; Nosek et al., 2010). Here, we focused only on women, but with accurate and transparent tracking of other demographic variables, a fuller picture of who is presenting at the conference and whether that representation is equitable can emerge. Representation matters. For individuals, presenting research at a conference increases the likelihood that research will be read (de Leon & McQuillin, 2015) and cited (Winnik et al., 2012). Given that citation counts are often considered when measuring the impact of a researcher’s work, and factored into hiring, retention, and promotion decisions, increasing citation counts increases the likelihood of professional success. Although we cannot directly attribute differences to fewer conference appearances, recent research has shown that women’s research is cited less than men’s, in general (Larivière, Ni, Gingras, Cronin, & Sugimoto, 2013), and specifically in social and personality psychology (Eagly & Miller, in press; Nosek et al., 2010).

For the audience, the identity of speakers matters (Murphy, Steele, & Gross, 2007; Sonnert, Fox, & Adkins, 2007). At early career stages, individuals are working to create a viable self-concept and construe role models as guides to the kinds of attributes they want to acquire and develop. In mid-career stages, role models help people refine their self-concepts. Successful similar others serve as role models for individuals developing their professional self-concepts (Gibson, 2003; Sealy & Singh, 2006). In the absence of such role models, individuals may not gain the socialization necessary for success. Relatedly, exposure to similar successful and counter-stereotypic women has been shown to change women’s self-concepts and reduce negative self-stereotypes (Asgari et al., 2012; Dasgupta & Asgari, 2004). Finally, role model gender has more impact on women than men (Lockwood, 2006).

For these and many other reasons, we should attend to the visibility of women and other minority groups in positions of status. Psychology is the study of humanity in different situations; because we know that who we are affects what we study (e.g., Inbar & Lammers, 2012), social and personality psychologists should represent all aspects of humanity and a wide range of lived experiences.

Footnotes

Acknowledgements

The authors thank Rebecca Sandoval, Thuha Cao, Pratyusha Meka, and the attendees at the Duck Conference on Social Cognition (2014, 2015) for their assistance and encouragement.

Authors’ Note

The Pamela K. Smith and Camille S. Johnson contributed equally to this article.

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, authorship, and/or publication of this article.

Supplemental Material

The supplemental material is available at online.

Notes

References

American Association of University Professors. (2015, March-April). The compensation survey (survey report, Table 12). Academe, 101(2). Retrieved from http://www.aaup.org/sites/default/files/files/2014%20salary%20report/table12.pdf

American Psychological Association Center for Workforce Studies. (2004-2014). Faculty Salary Survey. Washington, DC: Author.

American Psychological Association Task Force on Women in Academe. (2000). Women in academe: Two steps forward, one step back. Washington, DC: American Psychological Association.

Asgari

Dasgupta

Cote

N. G.

(2010). When does contact with successful ingroup members change self-stereotypes? A longitudinal study comparing the effect of quantity vs. quality of contact with successful individuals. Social Psychology, 41, 203-211. doi:10.1027/1864-9335/a000028

Asgari

Dasgupta

Stout

J. G.

(2012). When do counterstereotypic ingroup members inspire versus deflate? The effect of successful professional women on young women’s leadership self-concept. Personality and Social Psychology Bulletin, 38, 370-383. Retrieved from https://dx-doi-org.web.bisu.edu.cn/10.1177/0146167211431968.

Banaji

M. R.

Greenwald

A. G.

(1995). Implicit gender stereotyping in judgments of fame. Journal of Personality and Social Psychology, 68, 181-198.

Bickel

P. J.

Hammel

E. A.

O’Connell

J. W.

(1975). Sex bias is graduate admissions: Data from Berkeley. Science, 187, 398-404.

Biernat

Manis

Nelson

T. E.

(1991). Stereotypes and standards of judgment. Journal of Personality and Social Psychology, 60, 485-499.

Borgatti

S. P.

Everett

M. G.

(2006). A graph-theoretic perspective on centrality. Social Networks, 28, 466-484.

10.

Box-Steffensmeier

J. M.

Cunha

R. C.

Varbanov

R. A.

Hoh

Y. S.

Knisley

M. L.

Holmes

M. A.

(2015). Survival analysis of faculty retention and promotion in the social sciences by gender. PLoS ONE, 10, e0143093. doi:10.1371/journal.pone.0143093

11.

Brass

(1985). Men’s and women’s networks: A study of interaction patterns and influence in an organization. Academy of Management Journal, 28, 327-343.

12.

Brown

A. J.

Goh

J. X.

(2016). Some evidence for a gender gap in personality and social psychology. Social Psychological & Personality Science, 7, 437-443.

13.

Burt

R. S.

Kilduff

Tasselli

(2013). Social network analysis: Foundations and frontiers on advantage. Annual Review of Psychology, 64, 527-547. doi:10.1146/annurev-psych-113011-143828

14.

Casadevall

(2015). Achieving speaker gender equity at the American Society for Microbiology General Meeting. mBio, 6, e01146-15. doi:10.1128/mBio.01146-15

15.

Casadevall

Handelsman

(2014). The presence of female conveners correlates with a higher proportion of female speakers at scientific symposia. mBio, 5, e00846-13. doi:10.1128/mBio.00846-13

16.

Ceci

S. J.

Ginther

D. K.

Kahn

Williams

W. M.

(2014). Women in academic science: A changing landscape. Psychological Science in the Public Interest, 15, 75-141.

17.

Ceci

S. J.

Williams

W. M.

Barnett

S. M.

(2009). Women’s underrepresentation in science: Sociocultural and biological considerations. Psychological Bulletin, 135, 218-261. doi:10.1037/a0014412

18.

Cikara

Rudman

Fiske

(2012). Dearth by a thousand cuts?: Accounting for gender differences in top-ranked publication rates in social psychology. Journal of Social Issues, 68, 263-285.

19.

Dasgupta

Asgari

(2004). Seeing is believing: Exposure to counterstereotypic women leaders and its effect on the malleability of automatic gender stereotyping. Journal of Experimental Social Psychology, 40, 642-658.

20.

de Leon

F. L. L.

McQuillin

(2015). The role of conferences on the pathway to academic impact: Evidence from a natural experiment. Retrieved from http://ssrn.com/abstract=2507361

21.

Diekman

A. B.

Clark

E. K.

Johnston

A. M.

Brown

E. R.

Steinberg

(2011). Malleability in communal goals and beliefs influences attraction to STEM careers: Evidence for a goal congruity perspective. Journal of Personality and Social Psychology, 101, 902-918. doi:10.1037/a0025199

22.

Eagly

A. H.

Miller

(2016). Scientific eminence: Where are the women? Perspectives on Psychological Science, 11(6), 899-90. DOI: 10.1177/174569161666391.

23.

Gibson

D. E.

(2003). Developing the professional self-concept: Role model construals in early, middle, and late career stages. Organization Science, 14, 591-610. doi:10.1287/orsc.14.5.591.16767

24.

Ginther

Kahn

(2014). Academic women’s careers in the social sciences. In Lanteri

Vromen

(Eds.), The economics of economists (pp. 285-315). Cambridge, UK: Cambridge University Press. doi:10.1017/CBO9781139059145.015

25.

Hill

Corbett

St. Rose

(2010). Why so few? Women in science, technology, engineering, and mathematics. Washington, DC: American Association of University Women.

26.

Hoffer

T. B.

Hess

Welch

Jr. Williams

(2007). Doctorate recipients from United States Universities: Summary report 2006. Chicago, IL: National Opinion Research Center.

27.

Ibarra

(1992). Homophily and differential returns: Sex differences in network structure and access in an advertising firm. Administrative Science Quarterly, 37, 422-447.

28.

Ibarra

(1993). Personal networks of women and minorities in management: A conceptual framework. Academy of Management Review, 18, 56-87.

29.

Ibarra

(1997). Paving an alternative route: Gender differences in managerial networks. Social Psychology Quarterly, 60, 91-102.

30.

Inbar

Lammers

(2012). Political diversity in social and personality psychology. Perspectives on Psychological Science, 7, 496-503. doi:10.1177/1745691612448792

31.

Isbell

L. A.

Young

T. P.

Harcourt

A. H.

(2012). Stag parties linger: Continued gender bias in a female-rich scientific discipline. PLoS ONE, 7, e49682. doi:10.1371/journal.pone.0049682

32.

Kite

M. E.

Russo

N. F.

Brehm

S. S.

Fouad

N. A.

Hall

C. C. I.

Hyde

J. S.

Keita

G. P.

(2001). Women psychologists in academe: Mixed progress, unwarranted complacency. American Psychologist, 56, 1080-1098. doi:10.1037/0003-066X.56.12.1080

33.

Knobloch-Westerwick

Glynn

C. J.

Huge

(2013). The Matilda Effect in science communication: An experiment on gender bias in publication quality perceptions and collaboration interest. Science Communication, 35, 603-625. doi:10.1177/1075547012472684

34.

Larivière

C. Q.

Gingras

Cronin

Sugimoto

C. R.

(2013). Global gender disparities in science. Nature, 504, 211-213. doi:10.1038/504211a

35.

Lockwood

(2006). “Someone like me can be successful”: Do college students need same-gender role models? Psychology of Women Quarterly, 30, 36-46. doi:10.1111/j.1471-6402.2006.00260.x.

36.

McPherson

Smith-Lovin

Cook

J. M.

(2001). Birds of a feather: Homophily in social networks. Annual Review of Sociology, 27, 415-444.

37.

Mehra

Kilduff

Brass

D. J.

(1998). At the margins: A distinctiveness approach to the social identity and social networks of underrepresented groups. Academy of Management Journal, 41, 441-452.

38.

Moore

(1990). Structural determinants of men’s and women’s personal networks. American Sociological Review, 55, 726-735. Retrieved from http://www.jstor.org/stable/2095868.

39.

Murphy

M. C.

Steele

C. M.

Gross

J. J.

(2007). Signaling threat: How situational cues affect women in math, science, and engineering settings. Psychological Science, 18, 879-885.

40.

National Science Foundation. (1994-2014). Survey of earned doctorates. Retrieved from http://www.nsf.gov/statistics/doctorates

41.

National Science Foundation, National Center for Science and Engineering Statistics. (2015). Doctorate recipients from U.S. universities: 2014 (Special Report NSF 16-300). Arlington, VA: Author. Retrieved from http://www.nsf.gov/statistics/2016/nsf16300/

42.

Nosek

B. A.

Graham

Lindner

N. M.

Kesebir

Hawkins

C. B.

Hahn

Tenney

E. R.

(2010). Cumulative and career-stage citation impact of social-personality psychology programs and their members. Personality and Social Psychology Bulletin, 36, 1283-1300. doi:10.1177/0146167210378111

43.

Pell

A. N.

(1996). Fixing the leaky pipeline: Women scientists in academia. Journal of Animal Science, 74, 2843-2848.

44.

Perry-Smith

J. E.

(2006). Social yet creative: The role of social relationships in facilitating individual creativity. Academy of Management Journal, 49, 85-101.

45.

Roberts

S. G.

Verhoef

(2016). Double-blind reviewing at EvoLang 11 reveals gender bias. Journal of Language Evolution, 1, 163-167.

46.

Rose

S. M.

(1985). Professional networks of junior faculty in psychology. Psychology of Women Quarterly, 9, 533-547.

47.

Sanders

Willemsen

T. M.

Millar

C. C. J. M.

(2009). Views from above the glass ceiling: Does the academic environment influence women professors’ careers and experiences? Sex Roles, 60, 301-312. doi:10.1007/s11199-008-9547-7

48.

Schroeder

Dugdale

H. L.

Radersma

Hinsch

Buehler

D. M.

Saul

. . . Horrocks

N. P. C.

(2013). Fewer invited talks by women in evolutionary biology symposia. Journal of Evolutionary Biology, 26, 2063-2069.

49.

Sealy

Singh

(2006, August). Role models, work identity, and senior women’s career progression: Why are role models important? Academy of Management Annual Meeting Proceedings, 1, E1-E6. doi:10.5465/AMBPP.2006.22898277

50.

Simon

J. L.

Morris

E. K.

Smith

N. G.

(2007). Trends in women’s participation at the meetings of the Association for Behavior Analysis: 1975-2005. The Behavior Analyst, 30, 181-196.

51.

Sonnert

Fox

M. F.

Adkins

(2007). Undergraduate women in science and engineering: Effects over faculty, fields, and institutions over time. Social Science Quarterly, 88, 1333-1356.

52.

Stout

J. G.

Dasgupta

Hunsinger

McManus

M. A.

(2011). STEMing the tide: Using ingroup experts to inoculate women’s self-concept in science, technology, engineering, and mathematics (STEM). Journal of Personality and Social Psychology, 100, 255-270. doi:10.1037/a0021385

53.

Wang

Rodan

Fruin

(2014). Knowledge networks, collaboration networks, and exploratory innovation. Academy of Management Journal, 57, 484-514.

54.

Wasserman

Faust

(1994). Social network analysis. Cambridge, UK: Cambridge University Press.

55.

Wicherski

Hamp

Christidis

Stamm

(2014). 2013-2014: Faculty salaries in graduate departments of psychology. Washington, DC: American Psychological Association.

56.

Winnik

Raptis

D. A.

Walker

J. H.

Hasun

Speer

Clavien

P. A.

. . . Matter

C. M.

(2012). From abstract to impact in cardiovascular research: Factors predicting publication and citation. European Heart Journal, 33, 3034-3045. doi:10.1093/eurheartj/ehs113

57.

Martin

(2011). Gender differences in STEM disciplines: From the aspects of informal professional networking and faculty career development. Gender Issues, 28, 134-154. doi:10.1007/s12147-011-9104-5

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.17 MB

0.37 MB