Variation in levels of survey participation and in the quality of individual responses, as well as the problem of attrition in panel studies are ongoing core concerns for all survey researchers. Various reward and incentive structures are employed in order to maximise response rates and response quality. One component of this structure is the content and tone of covering letters or emails that ask respondents to participate (or, in some cases to continue to participate) in surveys. However, there is little evidence of a systematic approach to testing the effects of formulation, tone, and content of the messages that accompany survey requests on response propensity and response quality in the existing literature. We intend to fill this gap using a quasi-experimental research design in which 14,000 respondents to an online survey undertaken by Kieskompas.nl (a Dutch electoral advice website) are re-contacted and asked to participate in an additional survey. We formulate 8 different email messages (as well as using a 'baseline’ boilerplate letter) that tap into three dimensions of variation: in terms of content we contrast altruistic and egoistic appeals, and in terms of tone we examine two dimensions: formal versus informal writing styles, and linguistically simple versus linguistically complex formulations. We analyze the differential effects stemming from various formulations on response propensity and several data quality measures. Our analysis also seeks to uncover how demographic and personality-related characteristics condition such effects.