BACKGROUND AND OBJECTIVES

Gender-based communication differences are described in educational online communities, but have not been rigorously evaluated in medical online communities. Understanding gender differences in communication may provide insight into gender disparities in the medical profession. Our objective was to describe gender differences in post frequency, content, and language styles on the American Academy of Pediatrics Section on Hospital Medicine (SOHM) listserv.

METHODS

Posts were obtained from publicly available SOHM listserv archives. The first month of every quarter of 2019 and 2020 were reviewed. Two reviewers assigned a post topic (clinical, research, etc) and format (question vs statement) to all deidentified original posts (K = 1.0 topic, 0.89 format). Six trained reviewers assigned language styles (intraclass coefficient = 0.73, indicating good agreement).

RESULTS

We analyzed 1592 posts: 287 original posts and 1305 responses. Frequency: Women authored 50% of posts. The 9 most frequent posters (7 men, 2 women) accounted for 19.5% of posts. Content: Men’s posts had more words than women’s (132.51 vs 112.3, P ≤ .01). Men were more likely to post about health policy and research (P < .001). Men were more likely to post statements compared with women (39% vs 21%, P < .001). Style: Men’s posts were more likely to be coded adversarial (12.3% vs 5.5%, P < .001) authoritative (12.2% vs 6.5%, P < .001) or self-amplifying (6.5% vs 3.6%, P < .001).

CONCLUSIONS

Women contribute disproportionately fewer posts to the American Academy of Pediatrics SOHM listserv compared with their percentage in the subspecialty. We noted significant gender differences in language style and content, which may impact career development and online community inclusion.

Online communication has rapidly become an integral component of practice for health care providers, and the importance for networking and connecting is only anticipated to grow in the future.1  Online platforms, such as listservs, are important components of community-building,2  allowing providers to quickly share resources and offer mutual support without geographical constraints.3  Despite an initial hope that online communication would minimize biases by removing external cues, gender differences have emerged. In the 1990s, Herring described that men were more likely to post messages to online educational communities4  and use language styles described as adversarial, authoritative, and self-promoting, whereas women were more likely to use language styles described as supportive and attenuative.5  Gender differences in language styles have been correlated with career advancement, with traditionally masculine styles linked to higher rates of promotion in the business profession.6 

Despite a continual increase in the number of physicians connecting online, including physicians of all career stages, little is known about gender differences in posting and language styles on online medical professional communities. One study exploring professional benefits of social media found that men physicians perceive more advancement from social media, including more invited speaking and scholarship opportunities, compared with women physicians.7  This perceived difference may be partially explained by the increased amplification (eg, likes, retweets) men receive on social media compared with women.8  Increased amplification on social media is correlated with future citations,9  an important component of advancement in academic medicine.

As physicians increasingly connect via online platforms, it is important to understand gender differences in online communication that may impact career advancement, perceived leadership, and online community inclusion. One online physician community is the American Academy of Pediatrics (AAP), Section on Hospital Medicine (SOHM) listserv, which has 4133 subscribers (Niccole Alexander, e-mail conversation, December 2022). Because of its active online discussions and large audience, the SOHM listserv is an important platform for provider name recognition and amplification of scholarly work. Despite pediatric hospital medicine (PHM) being composed of 70% women, gender disparities have been described in this pediatric subspecialty.10,11  In this cross-sectional analysis, we aimed to describe any gender differences in post frequency, post content, and language styles on the AAP SOHM listserv.

Posts were extracted from the publicly available AAP SOHM listserv archives. All posts from the first month of every quarter (January, April, July, October) were included from 2019 and 2020. Because active posting on the listserv may have seasonality, sampling a full month from each quarter allowed the entire year to be sampled while also preserving threads of discussion. Job postings and AAP posts (eg, AAP newsletters) were not included. For the remaining included posts, all identifying information was removed (names, e-mails, hospitals, institutions, etc) and replaced with a generic term (ie, [name], [hospital]). A priori exclusion criteria were: Anonymous posts, moderator posts, administrative manager posts, posts “on behalf of a colleague,” and posts >1000 words. The latter were excluded because of difficulty assigning clear topics and language styles to longer posts. Posts were divided into 2 categories, original posts and responses, on the basis of whether a post was created as a new thread or a response to a previous post.

All original posts were assigned 1 post topic by 1 of 2 coauthors (J.A., H.F.) with the following categories: Clinical, systems, research, medical education, diversity equity inclusion, career development, health policy, board certification, other. Original posts were also categorized as either a question or a statement based on the perceived intent of the post. Before data collection using a random sample of 100 posts, the 2 reviewers had 100% agreement (K = 1.0) on post topic and 95% agreement (K = 0.89) on question versus statement. The 2 reviewers were unaware of post date and any identifying characteristics of the author during these assignments. Number of responses and post word count were manually generated by Excel. Acronyms (eg, AAP) were counted as 1 word.

Categories for language themes were created on the basis of previous studies.4,5  Definitions of the language themes were refined over 4 training sessions (Table 1). After 3 training sessions, 6 reviewers (3 women and 3 men) each received the same 100 random posts from the sample frame. The overall intraclass coefficient was 0.73 (95% confidence interval 0.66–0.79), indicating “good agreement”12  across reviewers. Individual language styles intraclass coefficient were: Adversarial 0.75, attenuation 0.8, authoritative 0.9, supportive 0.93, and self-amplification 0.8. After training, each reviewer received roughly 260 random posts to assign language themes. Original posts and responses were kept together to allow for context. Language styles were assigned as either present or absent, and each post could have >1 language style. The 6 reviewers were unaware of the author name, gender, and date of post. These 6 reviewers did not participate in coding gender.

TABLE 1

Language Style Definitions and Examples

StyleDefinitionExamples
Adversarial Language characterized by contentious assertions, sarcasm, put-downs, and/or antagonism “To think like that is absolute garbage…”
“Ok seriously? What a joke.” 
Attenuative Language characterized by hedging, expressing doubt, apologizing, and/or proposing ideas that are uncertain “I’m really sorry to bother everyone…”
“Every time I ask the listserv something, I feel dumb for asking.” 
Authoritative Language characterized by self-confidence, expert stance, and/or strong assertions “Everyone should be doing it this way.”
“I’ve been reading about this for years; there has never been a study showing…” 
Supportive Language characterized by appreciation, agreement, gratitude, and/or community-building “I want to say how much I appreciate this community.”
“Kudos to the planning committee for putting together the best conference.” 
Self-amplifying The act of showcasing one’s capabilities, skills, or achievements to others “Putting a plug for everyone to attend my [conference] session with [name]…”
“My coauthors and I recently published this paper looking at…” 
StyleDefinitionExamples
Adversarial Language characterized by contentious assertions, sarcasm, put-downs, and/or antagonism “To think like that is absolute garbage…”
“Ok seriously? What a joke.” 
Attenuative Language characterized by hedging, expressing doubt, apologizing, and/or proposing ideas that are uncertain “I’m really sorry to bother everyone…”
“Every time I ask the listserv something, I feel dumb for asking.” 
Authoritative Language characterized by self-confidence, expert stance, and/or strong assertions “Everyone should be doing it this way.”
“I’ve been reading about this for years; there has never been a study showing…” 
Supportive Language characterized by appreciation, agreement, gratitude, and/or community-building “I want to say how much I appreciate this community.”
“Kudos to the planning committee for putting together the best conference.” 
Self-amplifying The act of showcasing one’s capabilities, skills, or achievements to others “Putting a plug for everyone to attend my [conference] session with [name]…”
“My coauthors and I recently published this paper looking at…” 

Perceived gender was classified as 3 categories woman (she/her), man (he/his), or nonbinary (they/them) via review of name, online pronouns, or publicly searchable profiles based on previous methodology.11  No individuals were perceived to identify as they/them, so we report gender as binary (ie, woman/man). Gender was coded independently by 3 authors who did not review language styles (100% concordance).

Parametric and nonparametric statistics including t tests, and χ2 tests, as well as logistic regression, were used to analyze the data. In instances of binomial comparison where assumptions of independence were violated, McNemar’s χ2 was employed. R and RStudio, as well as Exploratory Desktop v.6.11 (Exploratory, Inc, United States, https://exploratory.io) were used for analysis. Statistical significance was defined as P < .05.

This study was deemed exempt by the University of Chicago board review.

A total of 1620 posts were retrieved from the AAP SOHM listserv archives. Twenty-eight posts were excluded, including posts written anonymously (7), from the listserv moderator (2), from the administrative manager (9), sent “on behalf of a colleague” (2), or >1000 words (8). The remaining 1592 posts were divided into 2 categories: original posts (n = 287) and response posts (n = 1305). A total of 578 unique posters were identified.

There was no statistically significant difference in total posts, original posts, or response posts based on gender (Table 2). Fifty percent of the total posts were authored by women. On average, each unique poster contributed 2.8 (SD 5.0, median 1, mode 1, range 1–56) posts. The 9 most frequent posters (7 men, 2 women) accounted for 19.5% of posts. Within this group of top posters, there were no statistically significant gender differences in post frequency.

TABLE 2

Post Frequency Based on Poster Gender

Total (n = 1592)aOriginal (n = 287)aResponse (n = 1305)aMean Words/Postb
Men 796 (50%) 140 (48.7%) 656 (50.3%) 132.5 
Women 796 (50%) 147 (51.3%) 649 (49.7%) 112.3 
Total (n = 1592)aOriginal (n = 287)aResponse (n = 1305)aMean Words/Postb
Men 796 (50%) 140 (48.7%) 656 (50.3%) 132.5 
Women 796 (50%) 147 (51.3%) 649 (49.7%) 112.3 

a Differences not statistically significant.

bP = .01.

On average, each original post was associated with 4.2 responses (SD 6.0, median 2, range 0–46). There was no difference in the frequency of responses authored by men versus women. Original posts from men tended to result in a higher, though nonsignificant (P = .14), number of responses (4.7 [SD 7.2] and 3.6 [SD 4.6], respectively). There was no difference in overall posting between 2019 and 2020 (P = .2), including no difference in all posts authored by women in 2019 (390) and 2020 (406), and no difference in all posts authored by men in 2019 (426) and 2020 (370).

Men’s posts had more words, on average, than women’s posts (132.5 vs 112.3, P ≤ .01). Original posts from men were more likely characterized as statements as compared with original posts from women (39% vs 21%, P < .001). There were differences in the original post topics (P ≤ .001), with men more likely to post about research, health policy, or other topics (Fig 1). There were no instances of women posting on health policy. More women posted on clinical and systems topics compared with men; however, these differences were not statistically significant.

FIGURE 1

Original post topics by post author gender.

FIGURE 1

Original post topics by post author gender.

Close modal

Of the posts reviewed, 45.5% (n = 731) of posts received at least 1 language style. Supportive was the most common language style assigned to 419 posts (26.3%). Attenuative, authoritative, and adversarial posts were coded at similar frequencies (171 [10.7%], 149 [9.4%], 142 [8.9%], respectively). Self-amplifying posts were the least frequently noted, 81 (5.1%). There were significant differences in language themes based on gender (P ≤ .001) (Table 3). Posts authored by men were more often coded as adversarial (98 [12.3%] vs 44 [5.5%] P < .001), authoritative (97 [12.2%] vs 52 [6.5%], P < .001] or self-amplifying (52 [6.5%] vs 29 [3.6%], P < .001) compared with posts authored by women. There were no gender differences in rates of attenuative or supportive posts.

TABLE 3

Post Language Style by Poster Gender

Total (n = 1592)Men (n = 796)Women (n = 796)P
Adversarial, n (%) 142 (8.9%) 98 (12.3%) 44 (5.5%) <.001 
Attenuative 171 (10.7%) 96 (12.1%) 75 (9.4%) .089 
Authoritative 149 (9.4%) 97 (12.2%) 52 (6.5%) <.001 
Supportive 419 (26.3%) 208 (26.1%) 211 (26.5%) .864 
Self-amplifying 81 (5.1%) 52 (6.5%) 29 (3.6%) <.001 
Total (n = 1592)Men (n = 796)Women (n = 796)P
Adversarial, n (%) 142 (8.9%) 98 (12.3%) 44 (5.5%) <.001 
Attenuative 171 (10.7%) 96 (12.1%) 75 (9.4%) .089 
Authoritative 149 (9.4%) 97 (12.2%) 52 (6.5%) <.001 
Supportive 419 (26.3%) 208 (26.1%) 211 (26.5%) .864 
Self-amplifying 81 (5.1%) 52 (6.5%) 29 (3.6%) <.001 

In this cross-sectional analysis of the AAP SOHM listserv, we found notable gender differences in online posting, post content, and language styles, which have important implications for gender disparities in our profession. In our sample, men and women post at similar frequencies (50%); however, the field of PHM is 70% women, indicating that women may be underrepresented on this platform. Previous studies have linked gender differences in online amplification9  and language styles to career advancement,6  thus our results may provide insight into gender disparities previously described in PHM.

The field of PHM is estimated at 70% women on the basis of multiple proxies, including the largest associated medical society (AAP SOHM), recent PHM fellowship graduates, and a random sample of university-based PHM faculty.10  Despite their majority, women are underrepresented as divisional/program leaders,10  senior authors, and invited speakers11  in PHM. In our sample of the AAP SOHM listserv, women authored 50% of the posts. The exact gender representation of AAP SOHM listserv participants is unknown because of the presence of inactive and duplicate accounts, but if the listserv mirrors PHM as a field, it can be inferred to be 70% women, thus women’s voices may be underrepresented on this platform. Previous studies in social media have described links between physician online presence and professional benefits, including speaking invitations, collaboration opportunities, and future citations.7,9  Although it is unclear how posting on the AAP SOHM listserv influences career advancement in PHM, the potential underrepresentation of women posters on the listserv may impact important components to career advancement, such as women’s name recognition, speaking invitations, and invited research collaborations.

Understanding any barriers to posting, for women and all members of a community, is critical to having an inclusive online communication board. The presence of adversarial posts might be 1 important barrier. Previous studies on online platforms have found that women are less likely to post or engage in conversations or platforms deemed to be adversarial.5,13  Although all genders disliked adversarial conversations, women had a stronger aversion and were less likely to engage in these conversations. Although adversarial language only occurred for 8.9% of the posts in our sample, these posts are often memorable and shape the culture of the online community. Minimizing adversarial language through the promotion of thoughtful posting and individual self-reflection may be 1 important action to make online communities more inclusive, encourage more diverse voices, and maintain open discussion.

Adversarial language may also have an important role in conversations about health policy. Previous studies have described that political and policy-based conversations are more likely to result in adversarial discussion.5,13  Although the small number of original posts about health policy in our sample precluded further analysis, it is notable that no women posted about health policy. Herring described that men were more likely to post online messages compared with women, and this difference was more prominent during active discussions around controversial topics of high interest to the online educational community.5  At a time when physicians are increasingly participating in national health policy discussion, including gun violence and reproductive rights, it is critical to have diverse voices and perspectives engaging in these conversations. Decreasing adversarial language may encourage more members of a community, including women, to engage in discussion around these important health policy topics.

In our sample, men’s posts were longer, more likely to be statements, and more likely to use language described as adversarial, authoritative, or self-amplifying. These findings are consistent with previous studies4,5  and have been associated with higher likelihood of promotion for men in business.6  The authors postulate that making statements and using authoritative or self-amplifying language increases perceived expertise, and this may have important implications for how members of a community define leaders. Despite this, encouraging women to do more self-amplification may not be an effective strategy because of gender stereotypes associated with self-promotion.14  Amplification of others’ expertise and scholarship on online platforms (eg, sharing someone else’s achievements) is 1 important strategy to mitigate the gender differences around language styles, specifically around self-promotion. In addition to benefiting women, amplification of others’ publications and scholarship is an effective action to support members underrepresented in medicine.15 

We did not find any gender differences in supportive or attenuative posts which have previously been described as more common language styles for women.4,5  In our sample of the AAP SOHM listserv, supportive was the most common language style and was used at similar frequencies by men and women. Supportive posts are an important component of community-building for online platforms.2  Online platforms, such as listservs, serve as communities of practice16  for health care providers which are protective against burnout.17  As health care providers face increasing amount of burnout18  and novel medical crises, these online communities of practice offer support and an ability to rapidly share knowledge and resources.3 

In our sample, we found that a small number of posters accounted for almost one-fifth of the posts in our sample, suggesting that the voices frequently heard on this platform are not necessarily representative of the >4000 subscribers in the community. As online communities grow in importance for physicians, further research is needed to explore any potential online structures or practices serving as barriers to posting to ensure professional online platforms remain open and accessible to all members. Additional studies should explore why providers choose to join and participate in online communities, and why some providers choose not to join and/or leave these platforms. It is also important to understand both the reach of posts and how different language styles may affect members of the online community.

Limitations of this study include the unknown gender breakdown of the AAP SOHM listserv limiting the ability to make direct statistical comparison between active posters and the listserv at large. We also cannot determine which posts were read versus deleted, therefore reach cannot be assessed. Although we had diverse reviewers regarding gender, race/ethnicity, practice setting, and experience level with good interreviewer reliability, interpretation of language styles is ultimately subjective. These 6 reviewers who assigned language styles did not participate in classifying gender; however, it is possible that conscious or unconscious identification of gender occurred, which could influence language style assignment. In addition, a small number of posters (mostly men) accounted for a large percentage of posts. We did not perform a cluster analysis of frequent posters because of a higher risk of unconscious identification; therefore, it is unclear whether certain language styles are more common in this group. Lastly, we did not assess other aspects of poster identity important to diversity, including race/ethnicity, sexual orientation, age, and disability status.

In conclusion, given the gender breakdown of our field and the AAP SOHM, our findings suggest that women contribute disproportionately fewer posts to the AAP SOHM listserv. We also found notable gender differences in post content and language styles, with men contributing more adversarial, authoritative, and self-amplifying posts. Further research is needed to explore any online structures or practices serving as barriers to participating in online posting to ensure these growing online platforms remain accessible to members and do not further exacerbate inequities.

Drs Allan and Fromme conceptualized and designed the study, collected data, participated in the analysis and interpretation of the data, and drafted the initial manuscript; Dr Black participated in the study concept and design, designed the data collection instruments, coordinated and supervised data collection, and conducted the initial analyses; Drs Lemiuex, Gilliam, Yardley, Kim, Ganem, and Black participated in the study concept and design, collected data, and participated in the analysis and interpretation of the data; Drs Schroeder and Ralston participated in the study concept and design, designed the data collection instruments, collected data, and participated in the analysis and interpretation of the data; and all authors critically reviewed and revised the manuscript, approved the final manuscript as submitted, and agree to be accountable for all aspects of the work.

FUNDING: No external funding.

CONFLICT OF INTEREST DISCLOSURES: The authors have indicated they have no conflicts of interest relevant to this article to disclose.

1
Karim
S
,
Uddin
S
,
Imam
T
,
Moni
MA
.
A systematic review of network studies based on administrative health data
.
Int J Environ Res Public Health
.
2020
;
17
(
7
):
2568
2
Herrmann
F
. Listserver communication: the discourse of community-building.
Schnase
JL
,
Cunnius
EL
, Eds.
CSLC ’95: The First International Conference on Computer Support for Collaborative Learning
.
L. Erlbaum Associates, Inc
:
1995
;
165
168
3
Smith
GD
,
Thomas
SL
,
Orscheln
RC
, et al
.
Listserv to support community pediatric clinicians during the COVID-19 pandemic
.
J Pediatric Infect Dis Soc
.
2023
;
12
(
4
):
254
255
4
Herring
SC
.
Gender differences in computer-mediated communication: bringing familiar baggage to the new frontier
. Available at: http://cpsr.org/issues/womenintech/herring2/. Accessed November 15, 2021
5
Herring
SC
. Gender and democracy in computer-mediated communication.
Kling
R
, Ed.
Computerization and Controversy: Value Conflicts and Social Choices
. 2nd ed.
Academic Press
:
1995
;
2
:
476
489
6
Weinberg
FJ
,
Trevino
LJ
,
Cleveland
AO
.
Gendered communication and career outcomes: a construct validation and prediction of hierarchical advancement and nonhierarchical rewards
.
Communic Res
.
2015
;
46
(
4
):
456
502
7
Woitowich
NC
,
Arora
VM
,
Pendergrast
T
,
Gottlieb
M
,
Trueger
NS
,
Jain
S
.
Gender differences in physician use of social media for professional advancement
.
JAMA Netw Open
.
2021
;
4
(
5
):
e219834
8
Zhu
JM
,
Pelullo
AP
,
Hassan
S
,
Siderowf
L
,
Merchant
RM
,
Werner
RM
.
Gender differences in twitter use and influence among health policy and health services researchers
.
JAMA Intern Med
.
2019
;
179
(
12
):
1726
1729
9
Klar
S
,
Krupnikov
Y
,
Ryan
JB
,
Searles
K
,
Shmargad
Y
.
Using social media to promote academic research: identifying the benefits of Twitter for sharing academic work
.
PLoS One
.
2020
;
15
(
4
):
e0229446
10
Allan
JM
,
Kim
JL
,
Ralston
SL
, et al
.
Gender distribution in pediatric hospital medicine leadership
.
J Hosp Med
.
2021
;
16
(
1
):
31
33
11
Kim
JL
,
Allan
JM
,
Fromme
HB
,
Forster
CS
,
Shaughnessy
E
,
Ralston
S
.
Gender distribution of scholarship and measures of national recognition in hospital medicine
.
Hosp Pediatr
.
2022
;
12
(
2
):
117
124
12
Cicchetti
DV
.
Guidelines, criteria, and rules of thumb for evaluating normed and standardized assessment instruments in psychology
.
Psychol Assess
.
1994
;
6
(
4
):
284
290
13
Maximova
OB
,
Lukyanova
GO
.
Gender differences online: self-representation and involvement in political communication on Facebook
.
Heliyon
.
2020
;
6
(
12
):
e05613
14
Exley
CL
,
Kessler
JB
.
The gender gap in self-promotion
.
Q J Econ
.
2022
;
137
(
3
):
1345
1381
15
Amplifying diverse voices in microbiology
.
Nat Microbiol
.
2022
;
7
(
10
):
1501
1502
16
Murty
SA
,
Gilmore
K
,
Richards
KA
,
Altilio
T
.
Using a listserv to develop a community of practice in end-of-life, hospice, and palliative care social work
.
J Soc Work End Life Palliat Care
.
2012
;
8
(
1
):
77
101
17
Delgado
J
,
de Groot
J
,
McCaffrey
G
,
Dimitropoulos
G
,
Sitter
KC
,
Austin
W
.
Communities of practice: acknowledging vulnerability to improve resilience in health care teams
.
J Med Ethics
.
2020
;
47
(
7
):
488
493
18
Shanafelt
TD
,
West
CP
,
Dyrbye
LN
, et al
.
Changes in burnout and satisfaction with work-life integration in physicians during the first 2 years of the COVID-19 pandemic
.
Mayo Clin Proc
.
2022
;
97
(
12
):
2248
2258