Team:Paris Bettencourt/Human Practice/Gender Study

From 2013.igem.org

Background

Science suffers from gender bias

Results

  • Revealed gender bias in synthetic biology by studying sex ratios in SB conferences and labs
  • Built a database of all iGEM teams reporting all available online information and sex ratios of teams and advisors
  • Conducted a statistical analysis of this data-set and showed among other results that success in iGEM is correlated to gender mix
  • Made recommendations to implement an active gender policy in iGEM

Aims

To investigate gender dynamics in iGEM and in synthetic biology research community at large in a quantitative manner

Introduction

   For every woman killed by TB, there are two men. Our review of the literature on gender bias and tuberculosis can be found here. If a disease can be biased, what about ourselves? iGEM? Synthetic biology? Gender bias in science may appear in different forms. Gender balance varies by discipline, by job title, by age or by region. Only 30% of researchers in Europe are women, while 92% of French university deans are men.

Hisorically, gender bias has affected the lives of scientists and the practice of science. However, assessing gender bias today in a living community is very difficult. History, stereotypes, limits of the disciplines, and the simple lack of data can prevent us, the synthetic biologists, from thinking about our own relationship to gender.


Most of those issues should not apply in synthetic biology. Synthetic biology is a new field. The argument of the heritage of some habits cannot be made. It is a mix of previously existing disciplines and therefore very open and should not reflect preexisting stereotypes. To study gender bias in iGEM and in synthetic biology we decided to follow a data driven approach. Studying in a quantitative manner this subjects had two main benefits. First it prevented us to apply our own biases and stereotypes on this subject. Secondly, it lead us to construct data base that we make freely available and let anyone test his own hypothesis on this controversial subject and form his own conclusions.

Synthetic biology field : general overview of gender equality in synthetic biology

   Gender repartition in synthetic biology can be looked at from different perspectives. For this study, two main ways were chosen: composition of labs and conferences. The main reasons for those choices were the accessibility of online data

as well as the necessity to get information not only about the general gender balance but also the sex ratio inside a defined category: PhD students, post docs, head of labs...

Synthetic biology labs, a good representation of gender (in)equality in science

   Teams of 50 synthetic biology labs were studied. The labs were chosen by their presence on the webpage http://syntheticbiology.org/Labs.html . For each lab, several numbers were reported in a table : total number of people in the team, number of women in the team, number of PhD students, post docs, head of labs, number of women PhD students, post docs, head of labs. From this, the sex ratios (number of women / total number of people) were then calculated for each of those categories.



The first conclusion that can be made is that women are generally under-represented in synthetic biology labs. 33% correspond to the average presence of women in research in Europe. Indeed according to the European Commission, 32% of researchers in Europe are women (She Figures, 2012).

The second finding also reflects well an already known reality in science : the glass ceiling. In 1995, the glass ceiling was defined by the U.S. Department of Labor, as a "political term used to describe "the unseen, yet unbreakable barrier that keeps minorities and women from rising to the upper rungs of the corporate ladder, regardless of their qualifications or achievements" . With only 17,85\% of heads of labs being women, synthetic biology is still doing slightly better than the average. According to a European study done in 2008 called Mapping the maze,getting women to the top in research., only 15% of women occupy top research position in Europe. However, the number of SB P.I. should be analyzed through the filter of history. In a new field, it would be expected in a world where bias would not be present anymore to have way more women at those positions.





Figure 1:Sex ratio in synthetic biology labs. The percentage of women by role in 50 synthetic biology labs. Error bars represent SD. The sex ratio of each lab is determined independently and then the mean of the labs was determined.



Labs Phd Students Post Docs Head of Labs
33,10 % 35,39 % 31,31 % 17,85 %

Speakers at SB Conferences : effects of an active gender policy

   SBX.0 conferences accompanied the development of synthetic biology. They provide a great way to investigate the evolution of gender ratio since the birth of synthetic biology. Moreover, the presence/absence of women as speakers is a known indicator of gender bias and specially of active gender policy. Indeed, several social mechanisms are in place lead to fewer female speakers that could be expected: self censorship, unconscious stereotypes, unconscious choice of only male speakers... However, having female speakers at conference is a key point. It allows women, to gain confidence but also to act as role model for women attending the conference.

To study SB conferences, available programs online were downloaded. Data referring to the number of speakers but also to posters were recorded. The data-set could not be completed for certain years due to the impossibility of finding the data online.



The sex ratio of the speakers have followed a very interesting evolution. It has been multiplied by 3 from SB1 to SB5. This could indicate a change of policy considering speakers. Most likely, the first conferences invited speakers without taking into consideration the gender dimension. Might it be due to some complaints or the raise in awareness of the conferences organizers, the numbers went up. This example is interesting because it clearly show an interest in the subject by the involved community.

Two main conclusions can be drawn on posters. First, the sex ratio of authors in posters has changed throughout the years. Secondly, this number is not as high as the sex ratio in labs. The question is why? The points described above could be underlying reasons, however it is very difficult to truly go beyond this with only those numbers.





Figure 2:Sex ratio in SB conferences. The proportion of speakers and poster presenters at SBX.0 conferences who are women. Data was gathered on-line from available programs.

Under represented and badly represented

   In order to try to better understand the dynamics of gender behind the posters numbers, the rank of authors were reported for each poster. Sex ratio were calculated for each rank, keeping in mind that in biology, the first author is often a Phd student or a post doc and the last author, the P.I.

As explained above, women are generally under-represented in synthetic biology labs, even less represented at conferences. When looking at the rank of author in posters, another bias appears. Indeed, women are more likely to be present as middle authors than first or last. This bias can be found in papers of different disciplines as shown on the graph realized on the eigenfactor.

The main finding considering gender in synthetic biology is that even though synthetic biology is new and interdisciplinary, it remains quite representative of existing gender bias in science. Therefore it can be concluded, that the issues that have kept women out of science and especially out of top research position are still present and will not be resolved with time. A strong and active policy appears necessary to bring more mixity and therefore diversity in this field.


Figure: Sex ratio according to rank of authors in SB posters. Authorship on Posters in SB conferences was collected and Women and Male authors are organized by their rank of authorship. Women tend to be middle authors more often then first or last authors.

Figure 4: Analysis of rank of authors according to gender in scientific publications.Women's rank of authorship in various journals and databases. There is over-all far more men cited as authors than women, and this is consistent across publications and scientific fields. Additionally, Authorship rank is shown on the bottom with the % of women over all shown as the solid line with the dot indicating the % of women within that authorship rank. Dots above the line indicate that women are over represented compared to the mean and dots below the line indicate that women are underrepresented, with the further they are from the line the greater the imbalance.

iGEM as a model : a fantastic database

Online Data

   All the data concerning iGEM were retrieved from the website : https://igem.org List of teams were retrieved from the webpages https://igem.org/Team_List.cgi?year=2012. List of project themes were retrieved from https://igem.org/Team_Tracks?year=2012. List of prices were retrieved https://igem.org/Results. List of judges were retrieved from: https://igem.org/Judge_List

Sex ratio determination :

   For each team, the official team profile was checked to count the number of student members, advisors and instructors. Then to determine the sex of particpants, wiki were used when names were not obvious, using pictures when they existed. When no pictures were available and names were not obviously referring to one sex, a google image search was done on the name (first and last name) and the sex was chosen as the most represented sex in the pictures (if 10 images of men come up and 30 of women, the participant was considered as a woman).

Database :

   Information for the first year of iGEM were difficult to find because of the non existence of available wiki pages and it was therefore decided not to take into account this year. Teams who withdrew during the competition were not taken into account since it was most of the time impossible to know the number of participants due to the absence of wiki. In the end our data set is composed of 662 teams over 5 years. For each team were reported : Year ; region ; name of the team ; number of student members ; number of women student members ; number of advisors ; number of women advisors ; number of instructors ; number of women instructors ; participation to MIT championship ; medal ; regional prices ; championship prices ;tracks.



Attrition by Career Stage

   With the introduction of High School iGEM competition, We have quantitative data about gender balance through career progression. By observing trends between the High School Division, Undergraduate and Overgraduate Divisions, Advisors and finally Judges; we can identify potential glass ceilings and find out why women are being lost through various career stages.

iGEM : a mirror of main gender problems

Teams sex ratio, a very robust value

  

The first thing that was examined was the evolution of sex ratio of teams in iGEM across continents and throughout the years.


The striking conclusion of this comparison is that the sex ratio is iGEM teams remains constant through the years and across continents (ANOVA's p-value for the different conditions > 0,5). This shows that women are underrepresented in iGEM teams.

Figure 5: Sex ratio of iGEM teams through the years and across continents. The proportion of team members of each gender over time and between regions in the collegiate iGEM competition. Bars represent the 95% Confidence interval.

Women do not supervise as much as men

   The second question investigated was the sex ratios for the different categories of people participating in iGEM. Indeed, iGEM is not only undergrad students. Advisors, instructors, judges also participate representing the complete professional ladder of synthetic biology. A category called Supervisors was created corresponding to instructors and advisers. Indeed, those terms are not understood and used in the same way in different continents. In some countries "advisers" means people who directly teach the teams (mostly grad students and post docs) whereas it means general mentors for others and vice versa.


Figure 6: Sex ratios in iGEM according to categories of people participating. The gender balance of students, Supervisors and Judges in iGEM collegiate competitions. Supervisors is taken as the combination of advisors and instructors due to variations on how individual teams differentiate between them. Bars are 95% confidence intervals.

   When executing comparisons tests , team members' sex ratio is found to be different from judges' and instructors' ones (p value < 0,01). However judges and advisers are not significantly different ( p value > 0,5). This result reveals a tendency of women to supervise less than men. Indeed, from team members to instructors, the sex ratio is divided by two. What is even more interesting is to compare those numbers to sex ratios of PhD students and post docs in labs. The sex ratio of instructors is 10 points lower.

Women constitute a pool of talent that is not mobilized. They participate but do not supervise teams. They are "lost" along the way. Indeed, in a study published last year in PNAS, researchers showed that P.I. were less prone to have a woman mentoring students than man. This unconscious bias can be translated by a lack of encouragement from P.I.s but also by a self censorship which is not taken into account by other supervisors as explained in an recently published article by Eileen Pollack (E. Pollack Why Are There Still So Few Women in Science? NY TImes October 2013).

Tracks and sex ratio in iGEM

   The third finding goes against an often-heard stereotype "women are more interested by applied research". In order to investigate this subject, tracks were reported for each project. In iGEM tracks correspond to general theme of the project : medicine, fundationnal research…Tracks were then looked at in terms of sex ratios. There is no significant difference between tracks. (ANOVA > 0,1).

Figure 7: Sex ratios and tracks in iGEM. The proportion of gender in teams grouped by the track that the team is entered in during collegiate iGEM competitions. There is no statistical significant difference between the gender balance between tracks.

High School Division is more balanced

When we looked at the High School Divison, we found that it had a higher percentage of women than the university iGEM teams. Additionally, The number of female advisors and instructors in the High School division is much higher than that of the Collegiate division and is approximately the same as the proportion of students. This indicates that there is a problem at the Collegiate level and that iGEM can be an important bridge for women to access new opportunities to lead in higher education



Figure 8: Sex ratios in High School iGEM by year and role. On the left, the proportion of students in the High School Competition that are women for each year. On the right, the proportion of women in each role for all years. Bars represent 95% Confidence Interval.

To conclude, studying the iGEM competition gives a unique quantitative insight on existing questions in the field of gender studies. It also constitutes an amazing argument to convince scientists of the existence of a gender issue in science. As explained by Rascun et al in a recent paper published in PNAS, scientists believe that those type of bias only exist in some labs, not their own, therefore very objective studies need to be conducted to clearly show the reality of the numbers. More over , Jo Handelsman a microbiologist involved in that paper underlined in a recent interview, that people often think that there is still an issue in physics or maths but that there are no more women issues in biology, which is not true. This study supports strongly the view that this general thinking is untrue.

In iGEM, is diversity a factor of success ?

   Several studies led by consulting groups (McKinsey and Company Women Matter, 2007) have shown that mixity in a team increases performance. The big question of what leads to success in iGEM was therefore investigated using the database with a special focus on gender. In order to be able to get a general idea about iGEM team success, a point system was put in place.

Points were attributed the following way.
For the medal: 1 point for bronze medal, 2 points for silver medal, 3 points for gold medal. For the world jamboree qualified teams: 2 points for every team taking part in 2010 and before (before regional jamborees existed) , 6 points for team qualified for world final (after 2010). For special prices (Best ...): 6 points were attributed for each regional price earned (only after 2010), 13 points for each price earned in the world final (all price worth 13 points before regional jamborees existed). For the final place in world final: 15 points for the sixth team, 20 points for the fifth team, 25 points for the fourth team, 30 points for the third team, 35 points for the second team, 40 points for the firth team.


Figure 9: Gender balance and succes in iGEM. The proportion of women in teams that have won prizes in iGEM compared to the proportion in teams over all. There is a significantly higher proportion of women in teams that win prizes (p=0.034).

The aim was to give each team a score that is proportional to the rewards it earned, taking in account that all teams were in world jamboree prior to 2011, without having to be qualified in regional jamborees.

Best score is for the Imperial College London team in 2011 (81 points). All teams (all years) average is 7.41 points, considering teams with no points (due to withdrew).

Correlations studies between this number of points and other variables show that that for all teams, the main variables explaining success in iGEM is the number of years of existence and the size of the team. It would therefore seem that mixity would not be a factor. However, when looking at correlations between variables of teams who truly succeeded (points > 20) , the variables that have a significant correlation with the number of points become the sex ratio and the number of supervisors. Therefore it could be hypothesized that beginning iGEM teams have to face major challenges but when the team existed for a few years and general organization or funding problems have been dealt with , diversity could be a factor for success.

In order to check if this could be seen in the best iGEM teams that existed, the sex ratio of of prize winner teams was compared to the one of participating teams with boostrap resampling giving a p-value of 0.035 This means that the sex ratio of winning teams (45%) is significantly different from the one of participating teams (37%)

Clues to improve mixity

   Women are not as represented as men in iGEM. Why should this be a problem ? Indeed, even if it might lead to success as explained above, the need to have gender equality could be questioned. However iGEM is an international competition. One of its main goals is to attract and educate young people as well as trying to have them solve real issues. Synthetic biology might be a key technology to solve the main challenges of the 21st century.

The world will need science and if iGEM only succeeds in motivating half of the population that could be interested, this would be a major failure to achieve its mission. Therefore, the last part of the study was aimed at understanding how could iGEM improve mixity within its own ranks.

From the data

   By looking at correlation between sex ratios and other variables, the most striking result is the link between team size and sex ratio. Teams of 2 or 3 people are almost only male teams. Even when taking out those very small teams, out of the data set the correlation holds up. This is a first lead.
The second analysis that was made regarding the data was to compare the detailed statistics of the 100 most female teams and 100 male teams. Again, it is found that the total team member is lower for male team (9,7 vs 7,8 (p-value 0,0019) we can hypothesize that having women instructors does matter to attract girls in teams. They serve as role models. Having a woman capable of studying and realizing a synthetic biology project is a direct signal to female students that it is also possible for them to do it. Having a woman adviser might also help girls better adapt in a group and reduce their fears about having to endure constant teasing or "male " ambiance.

From a survey

   Finally, a survey was conducted among iGEMers and former iGEMers to understand their motivations and activities in iGEM. The study was designed to be unbiased and to avoid stereotype threat (for example by putting the question about gender in the end among many other pieces of information). It is still available here. Participants in the survey had to rank from 1 to 5 (1 being not important, and 5 very important) answers to questions regarding personnal and professional motivations for participating in iGEM as well values and on what did they spend their time. 63 people answered among whom 32% were women.

It is interesting to notice that men and women answered almost exactly the same way regarding most of the questions. Women gave a little more importance for the value of fundamental research in iGEM while men graded a bit better "Changing the world". Motivations were approximately the same as well as time spent on each activity. Just a little fact was that men considered human practices a bit more important than women did but spent a little less time on it. There is only one main difference (more than one point out of five which is represented below) : the will to lead a project and lead a team. It is striking to see how much men are more motivated to lead teams than women. This is definitely to put in relation with the number of women advisers found and the impact it can then have on teams mixity. This could reflect women lack of self esteem in some parts of their work.

Figure 10: Results of survey : what did you hope to learn in iGEM? A survey was distributed to iGEM teams asking participants to rank their motivation to participate in iGEM in various subjects between 1-5. The only significant difference between the motivation of men and women in iGEM were in the subjects leading a team and leading a project.


Recommendations

   Considering all the results that were presented above, here is a list of recommendations for the iGEM foundation to pursue an active policy to improve mixity in iGEM.

  • Raise the number of women judges
  • Promote large teams
  • Write up a small paragraph to team heads to insist on the importance of motivating young women to be advisers.
  • Giving Bonus point when the team have women advisers

And finally, add in iGEM requirements a Gender reflection. By having teams filling out the database that was built and answering the survey and write a small paragraph about how they see mixity in their team and what it could bring, it would drastically raise the awareness of young men and women about the gender problem in science. Having an up-to-date database is also a great way to see improvements in a quantitative manner. It would allow a direct assessment of the effects of an active gender policy which would be a unique example in science. iGEM could become a leader in that fight and prepare the new generation of scientists to finally get rid of the gender inequality in science

Litterature

  • P. Allotey, M. Gyapong Gender in tuberculosis research INT J TUBERC LUNG DIS 2008
  • M. Calid, S. Rasul, S Ullah Khan, M; Saeed Gender differences in delay to s to tuberculosis diagnosis and treatment outcome
    European Commission She figures 2012
    European Commission, Mapping the gaze : getting more women to the top in Research 2008.
  • C.B. Holmes, H. Hausler, P. Hunn : A review of sex differences in the epidemiology of tuberculosis
  • A. N. Martinez J. T. Rhee, P. M. Small,‡M. A. Behr Sex differences in the epidemiology of tuberculosis in San Francisco INT J TUBERC LUNG DIS 4(1):26–31 2000
  • Moss-Racusin et al, (2012) Science faculty’s subtle gender biases favor male students PNAS
  • McKinsey and Company Women Matter, 2007
  • Olivier Neyrolles, Lluis Quintana-Murci Sexual Inequality in Tuberculosis, Plos Medicine 2009
  • Nosek et al. (2009) National differences in gender–science stereotypes predict national sex differences in science and math achievement. PNAS June 30, 2009 vol. 106 no. 26 10593–10597
  • E. Pollack Why Are There Still So Few Women in Science? NY TImes OCtober 2013
  • Al S. Rhines The role of sex differences in the prevalence and transmission of tuberculosis : Tuberculosis 2013
  • M. W. Uplekar, S. Rangan, M. G. Weiss, J. Ogden, M. W. Borgdorff, P. Hudelson Attention to gender issues in tuberculosis control INT J TUBERC LUNG DIS 4(1):26–31 2001

Attributions

We would like to thank Flora Vincent, President of WAX Science association for her precious help in analyzing the results, and Kim de Mora and Kitwa from the iGEM foundation, for helping spreading the survey.

This project was designed and accomplished by Aude Bernheim, Clovis Basier, Matt Deyell, Marguerite Benony and Sebastian Jaramillo in consultation with Edwin Wintermute and Ariel Lindner.

Centre for Research and Interdisciplinarity (CRI)
Faculty of Medicine Cochin Port-Royal, South wing, 2nd floor
Paris Descartes University
24, rue du Faubourg Saint Jacques
75014 Paris, France
+33 1 44 41 25 22/25
team2013@igem-paris.org
Hit Counter by Digits
Copyright (c) 2013 igem.org. All rights reserved.