Collecting effectively information from people in web: a marketing challenge for social sciences

The aim of this research is to discuss how to effectively increase survey response and retention rates in online panels, a burning issue which can be viewed as a marketing problem in the social sciences. The author firstly considers the theory behind the methods for motivating people to participate in a survey, to respond and to remain for the full duration of a study. Then three hypotheses are formulated and tested in order to investigate the effects of combined strategies on participation and attrition. By means of a web-based panel experiment the research draws the attention to empirical evidence of phenomena that could be exploited in order to increase survey response and panel retention rates, namely the respondent tendency to lie and the trend of his trust in the inquirers. In order to enhance the construct validity a quantitative experiment and a qualitative research are integrated in a mixed approach design. The findings suggest that the trend of dropout is weaker when the relationship with the inquirers is framed as based on something like reciprocity than when it isn’t. Furthermore increasing tendency to lie in completing the questionnaire and diminishing declared trust in survey authors and sponsors seem to work as effective predictors of dropouts and non-responses in following experiment waves. And so two indices are proposed that can help panel managers to monitor the willingness to respond: the Pinocchio Ratio and the Candide Ratio. In short the research implies that stressing the reciprocity features of relationship can’t increase survey response but can reduce panel attrition in web-based studies. Moreover it points out that the more respondents tend to lie, the less they are willing to participate in following panel waves. Furthermore the less inquirers enjoy the trust of people under investigations, the more panelists are likely to drop out and to non-respond. Nevertheless this study with the reported online experiment is not without limitations: on the one hand having recourse to all the members of a sub-population and not to a probability sample of the overall population may weaken the chance of generalize the findings. On the other hand using different amount of questionnaires and diverse intervals between waves and reminders could affect the results. The main practical implication for marketing research and social sciences investigations at-large lies in the remark that reciprocity framework used as incentive doesn’t seem to be very effective in increasing response rate of one-time survey: therefore spending on implementation could be deemed unnecessary to do. On the contrary, in panels where the relationship with the inquirers is framed as based on reciprocity something happens. So it’s worthwhile considering the possibility of inserting in online questionnaires some items that can provide answers which work as predictors of future dropouts or non-responses. These items enable inquirers to monitor willingness to respond through the proposed indices and allow panel managers to intervene promptly in order to limit attrition. The originality and the value of the study come from the choice made by the author of collecting and processing a large e-mails data bank that represents an overall sub-population with some descriptive characteristics well known. In fact the previous knowledge of some respondent’s socio-demographic aspects enabled an easier analysis of lies which was added to the estimate of declared trust: that allowed the construction of the two indices working as predictors, namely the Pinocchio Ratio and the Candide Ratio.