The existing studies that aim to analyse the symbolic impact of women's representation on citizens' political attitudes face a number of methodological challenges. Among those there are: (i) the risk of spurious correlations – it is hard to prove whether the change that is observed is really due to the sex of the politician(s) and not to something else; (ii) the possibility of a reciprocal relationship – it can very well be that those countries where the gender gap in political engagement is already lower are also those more likely to have a higher presence of women in political power; and (iii) endogeneity because the people who are object of analysis are the same people who voted for the female politicians whose impact is being analysed. We propose a reflection on the advantages and disadvantages of some alternative methodological avenues, namely the use of experimental surveys, panel surveys and quasi-experimental designs.
