Gender differences in how authors “hype” the importance
of their research

Ira Mills,a Suze Kundu,b Bob JA Schijvenaarsb

aParexel International, Hackensack, NJ, USA, bDigital Science, London, UK

KEY FINDINGS

  • The percentage of publications with hyped words in abstracts has increased over time from 2002 to 2021
  • Use of "hype" words in abstracts was observed more frequently among male authors than their female counterparts

Most frequently used hyped words

Why does this matter?

  • In our research, we observed a gender imbalance in the use of "hyped words" more commonly found in publications authored by males
  • This may unfairly impact how we measure excellence in publications
  • More conservative data interpretation by authors may reduce this gendered imbalance
  • By using technology to quickly analyze the landscape of publications through a demographic lens, we can monitor progress toward the goal of more balanced activities

Author video

CONCLUSIONS

  • This study confirms increased hyping over time in publications across a broad range of science and medical journals with high impact factors (IFs), mostly by male authors
  • More conservative data interpretation by authors may reduce this gendered imbalance

INTRODUCTION

  • Recent studies suggest that authors “hype” their findings in publications and this was observed more frequently with male authors in journals with high IFs (>10)1-5
  • We examined these gendered differences using a scholarly information platform

Click the numbered list icon above to view the references

METHODS

  • General science and medical journal original research articles (IF>10) in 2002-2021 were assessed for use of hyped words by gender
  • Using the Dimensions platform (https://www.dimensions.ai), we identified author gender using the genderize.io web service, which requires a first name and an optional country code, and applied a 90% probability threshold

Click the paperclip icon above to view additional Methods

Methods (Cont’d)
  • Using the Dimensions author disambiguation we could derive author first names in many publications with first initials only
  • The hyped words were taken from Lerchenmueller et al,5 and supplemented with handpicked words close to any of these based on their GloVe encodings
  • Regular expressions were built for each hype word that took some spelling variations into account (eg, favorable/ favourable) and applied to all publications with an abstract in Dimensions

How did we do this?

  • Thanks to the capabilities of modern data science and the comprehensive coverage of the Dimensions database combined with the powerful guessing program (genderize.io)

RESULTS

  • A total of 443,104 publications from 159 journals (IF >10) were surveyed from 2002 (18K) to 2021 (32K) for use of hyped words (Table)
  • Across all publications, male/female/unknown gender for first and last author were 220K/100K/100K (52%/24%/24%) and 310K/50K/75K (71%/12%/17%), respectively
  • The number of hyped words in abstracts increased over time, with greater use by male first authors (Figure 1) consistently found across the most commonly used hyped words (Figure 2)
  • Similar trends in gender were observed by last author or combined first/last author and in titles

Table

High-tiered journals (IF >10) with "hyped" words in the abstract from 2002–2021a

Figure 1

Percentage of publications with "hype" words in the abstract over time (2002-2021)

Figure 2

Number of publications with "hyped" words in the abstract by first author gender

Study limitations

Polling Questions

Q1: Which of the observations of “hyped” word usage by authors in reporting scientific findings do you find surprising?
Increased over time period studied (2002-2021)
More common with male vs female authors
Gender-based differences in high‑tiered (>10 IF) journals
None of the above
All of the above

Please select one of the above options to vote

Q2: Do you feel that authors should be more conservative in describing their research findings in publications?
Yes
No

Please select one of the above options to vote

Click to view our Call for Action

Call for Action
  • Many organizations are aware of their gender imbalance and are putting in place initiatives to help mitigate. However, if this unconscious bias is not recognized, uninclusive cultural behaviors will continue to persist, further driving imbalance
  • We all need to take responsibility for equality, diversity, and inclusion and create frameworks to support these elements

Acknowledgments

This study was supported by Parexel International and Digital Science. The poster was built using Scign ePoster technology.
©2022 Parexel International Corporation. All rights reserved. Logos are the property of their respective owners. Poster presented virtually at the 18th Annual Meeting of ISMPP, May 9–11, 2022, Washington, DC.

Key sections
Thank you for your response

Your input is valued

Questions and Comments

* required field

Submit
REFERENCES