| dc.description.abstract | Contributing to a growing debate about `symbolic' or non-policy effects of gender-balanced legislatures, my thesis sets out to tackle issues of (a) measurement, combining experimental and observational evidence of the effect of female politicians as role models on women's political engagement; and (b) mediation, considering the underlying mechanisms convincing on the individual-level of voter psychology, explaining why role models are powerful in engaging fellow women in the electorate.
Firstly, I triangulate results from an eye-tracking experiment investigating attentional bias to gender balance in manipulated picture stimuli of political groups; an online experiment investigating measures of psychological engagement with politics as a function of gender balance in the same picture stimuli; and British Election Study panel data investigating campaign effects on psychological engagement with politics as a function of the gender balance among candidates running in the 2010 and 2015 UK parliamentary constituencies.
My results suggest two general types of role model effects: one of `tokenism' where women's striking minority presence impacts political attention and the probability of learning about politics, and one of `linear' effects where a gradual increase in women's presence in political groups towards parity translates into a gradual increase in political self-efficacy and confidence about political knowledge.
Secondly, I develop and test hypotheses about mediation in terms of implicit mechanisms not requiring that citizens consider the policy output of their representatives, drawing heavily on the stereotype threat literature especially on the role of affect. Using a more classical, regression-based approach to mediation analysis, along with a novel crossover experiment or `design-based' mediation analysis, I present preliminary evidence that, following exposure to role models, women experienced fewer self-evaluative threats as evidenced by anxiety, explaining effects on self-efficacy in politics. I present an additional study scrutinising affect, and show that the action-oriented anger may result in approach of the source of threat, reversing stereotype threat effects under `men-only' politics.
Thirdly, I develop and test hypotheses about mediation in terms of instrumental mechanisms that do require expectations or associations about policy output. Through similar approaches to mediation analysis, I show that though women expect better policy across two domains with more female politicians on board, greater competency attributed to elites is, if anything, negatively related to self-efficacy in politics. In a full-experimental study, I find no evidence that women's greater self-efficacy is due to expectations about women-friendly policy pursued by role models. | en_GB |
