Establishing the empirical significance of female underrepresentation in political and civic life is a stubbornly difficult problem. The main challenge is that the selection of individuals for most positions – elected officials or political appointments – is far from random. This makes it difficult to separate the causal effect of female officials from the circumstances that lead to their appointments.
For example, the majority of female members of the US Congress represent progressive/liberal districts. A large body of correlational evidence suggests that male and female politicians differ systematically in basic policy positions, active policy focus, parliamentary approach and legislative effectiveness, implying that gender composition affects a range of substantive outcomes (e.g. Pearson and Dancey 2011, Anzia and Berry 2011, Volden et al. 2013, Wittmer and Bouche 2013). However, it is difficult to disentangle the influence of a politician’s gender on a vote from that of the political ideology of the constituency that they were chosen to represent.
Perhaps more subtly, this same selection problem also hampers the identification of gender differences using variation within jurisdictions. In this case, the positions of those appointed (regardless of gender) must be broadly compatible with the person or constituency doing the selecting. This tends to diminish gender differences among those selected relative to the larger population of qualified candidates. Strikingly, even the apparent gold standard of the random assignment of court cases to judges does not avoid this fundamental selection problem, as both female and male judges must have views and judicial approaches that permitted them to be appointed in the same district in the first place. This is perhaps reflected in George’s (2001) conclusion that the empirical literature has failed to find any broad and consistent gender-based differences in judging, and that many researchers have concluded that judge characteristics, of which gender is just one example, simply do not explain much of the variation in judicial decisions (e.g. Ashenfelter et al. 1995).
The ‘great experiment’: Women join English juries for first time
In a recent paper, we turned to the history books to help shed some light on this difficult question (Anwar et al. 2016). Specifically, we studied the impact of adding females into the English jury pool – the group of people from whom the final jurors are selected – on criminal court outcomes. Termed the ‘great experiment’ by newspapers of the time, women became eligible to serve on English juries with the passing of the Sex Disqualification (Removal) Act of 1919.
Our focus on this historical period is motivated by several considerations. First, even though eligible citizens may be randomly called for jury duty in modern settings, the pre-trial jury selection process provides attorneys and judges considerable discretion over which members of the jury pool are actually seated for the trial, re-introducing a version of the fundamental selection issue described above. Second, variation in the composition of modern juries is typically centred at a point of approximate gender balance; results are generally only informative about variation near this point of gender balance – for example, identifying the difference in outcomes for 12-person juries with six versus seven female members.
In contrast, our analysis estimates the impact of including female jurors on what had been all-male juries, providing a sharper and more meaningful comparison for assessing the broad impact of female representation.
We use an original data set of more than 3,000 criminal cases from archival hand-written court records of the First and Second Courts of the Old Bailey – the London criminal court – from 1918 to 1926, including, most importantly, the names of all individuals seated on the jury. The ability to directly observe the gender of seated jurors distinguishes this archival data from that in other jurisdictions with similar reforms (the female jury movement was also occurring in the US at the time), and allows us to directly assess the extent to which the reform was implemented. Though the reform was passed in 1919, it was not implemented until 1921 due to the logistical issues of updating the jury books from which the pools were drawn. More than half of seated juries in 1921 had at least one female juror and more than 80% of cases in the following years had female jurors. There is also evidence of some ‘selection’ of females onto juries according to case characteristics, especially for sex offenses – females are less likely to be represented on such cases.
This kind of differential pre-trial selection motivates our primary empirical strategy of comparing verdicts from the full set of trials before and after the introduction of the reform. This reduced form approach focuses on the margin of the basic inclusion (or not) of female jurors and avoids the potential problem of pre-trial jury selection by not explicitly conditioning on the number of seated female jurors in the post-reform period. The estimated effect encompasses many potential mechanisms, the most obvious of which is that eligibility leads directly to female representation on juries, thereby affecting deliberations and decisions. But the reform may also indirectly affect verdicts even in cases in which females are not seated on the jury. This could occur, for instance, if the judge makes comments to the entire jury pool about whether the nature of the case is ‘sensitive’ for female ears or if pre-trial selection is focused on striking potential female jurors, which could affect the resulting distribution of seated male jurors.
Our analysis yields a number of key findings.
- Bundling all cases together, the reform did not have a significant impact on the likelihood of conviction.
- But, female representation on juries significantly increased conviction rates for sex offense cases by 16 percentage points, and decreased those for property and violent crimes by 10 and 13 percentage points, respectively.
- The decrease in conviction rates for violent crimes as a whole conceals opposing results based on the gender of the victim.
Prior to the reform, the conviction rate differential between male and female victim violent crime cases was essentially zero; post-reform, this conviction differential between female and male victim cases increased to 20 percentage points.
In a secondary analysis, we study the impact of seated female jurors for the subset of cases in which the jury was carried over from a previous trial. We expect issues involving the non-random seating of female jurors to be muted in these cases and, in fact, the gender composition of ‘carried-over’ juries is not correlated with the gender of the victim in violent crime cases. This analysis reveals a pattern of results that is remarkably consistent with the initial pre-post reform analysis, indicating that adding females to the seated jury sharply increased conviction rates for violent crime cases against women.
Why did adding female jurors lead to more convictions when the victim was a woman? While our data does not allow us to explicitly determine this, there are many potential explanations. One set of explanations involves biases on the part of female and/or male jurors towards female victims. For instance, female jurors may have over-sympathised with female victims, or alternatively all-male juries might have perceived violence against women as less harmful. Of course, these findings can also be generated by mechanisms that do not involve biases. For instance, the quality of evidence may not be constant across cases if juries composed of both males and females interpret evidence differently than all male juries. Without knowing more about the exact nature of the evidence, it is impossible to distinguish among these explanations. An important consideration in weighing these possibilities, however, is that juries were still overwhelmingly male after the reform, requiring female jurors to affect the votes of a large number of male colleagues in order to impact trial outcomes.
Concluding remarks: A lesson from history?
Taken as a whole, the results of our analysis imply that female representation on juries substantially affects the likelihood of conviction for a subset of cases – sexual and violent crimes – in which female jurors might have viewed the alleged behaviour or its impact on the victim from a different perspective than their male counterparts. While the present study is based on data from nearly a century ago, the magnitude of the results raises concerns about the basic fairness and effectiveness of modern criminal justice systems that exclude – or severely limit the role of – women. Such concerns are not limited to societies that completely exclude women from serving as judges and jurors but extend to courts or institutions, such as the military, where the historical overrepresentation of men in positions of authority continues to limit the role of women in judicial decision-making.
Anwar, S, P Bayer and R Hjalmarsson (2016) “A jury of her peers: The impact of the first female jurors on criminal verdicts”, CEPR Discussion Paper No. 11116.
Ashenfelter, O et al (1995) “Politics and the judiciary: The influence of judicial background on case outcomes”, Journal of Legal Studies, 24: 257.
Anzia, S F and C R Berry (2011) “The Jackie (and Jill) Robinson effect: Why do congresswomen outperform congressmen?”, American Journal of Political Science.
George, T (2001) “Court fixing”, Arizona Law Review, 43.
Pearson, K and L Dancey (2011) “Elevating women's voices in congress: Speech participation in the House of Representatives”, Political Research Quarterly, 64: 910-923.
Volden, C , A E Wiseman and D E Wittmer (2013) “When are women more effective lawmakers than men?”, American Journal of Political Science, 57(2): 326–341.
Wittmer, D E and V Bouche (2013) “The limits of gendered leadership: Policy implications of female leadership on “women's issues”, Politics & Gender, 9(3): 245-275.