This conference paper shall discuss preliminary findings from a survey experiment carried out in slums in the city of Bhubanswar in India, during January - March 2020. The survey experiment focuses on analysing the effect that different types of clientelist promises have on the vote choice of slum voters. The experiment does this by presenting slum voters with pairs of hypothetical candidates, who vary along four dimensions - party affiliation, (social) caste identity, type of clientelist offer, and degree of coerciveness in the clientelist offer made. Recent elections in India have seen intense criticism of party candidates who have openly discussed elections as a transaction in their campaign speeches. The effect of such 'quid pro quo' statements on the choice of poor voters is less explored than the effect of positive clientelist inducements. In addition to this experimental component, the survey also probes the different types of political and civic participation in which the respondents engage - from party rallies, to community groups - in order to draw further conclusions about the relationship between political participation, clientelism, and vote choice. This study builds upon extensive qualitative work done in the slums of Bhubaneswar (Orissa) that examines how those living in urban slums simultaneously participate in protest politics, mobilise votes, and engage in clientelist negotiations with politicians. The study is part of a broader dissertation project that explores the relationship between political mobilisation by citizens and the institutions of electoral politics.