Relevance & Research Question: The aim of this research is to evaluate whether measures taken at the data collection stage could improve the estimates in on-line panels. Attention is also paid to survey costs reduction. This is a relevant question since response rates are declining and increasing effort to achieve preset response rates is required during the survey process. Optimizing the data collection effort could help in reducing the effects of the abovementioned situation.
Methods & Data: To achieve our task we propose to apply a form of responsive design in the framework of panel data collection. The main idea underlying this method is to intervene in the data collection process, in order to achieve an ultimate set of responding units “better balanced” or “more representative” than if no special effort is made. The intervention points in the data inflow are chosen by monitoring the data collection by indicators of balance and representativity. These indicators are computable from selected auxiliary variables, known for the responding units as well as for the non-responding ones. During the on-line data collection many variables on the participation process become available and let this approach feasible.
To evaluate bias changes we use data from the PAADEL panel. This is an on-going probability-based Italian household panel for the agro-food sector managed by the University of Bergamo.
Results: A tracking protocol of the recruitment process has been written up. During the recruitment process of the panel a database has been created on the basis of the abovementioned protocol. Therefore information for studying the impact of responsive design on the estimates is available. Starting from the collected data, we artificially have reproduced a set of experimental responsive designs based on alternative interventions in the data collection. We evaluate whether this method improves the estimates in terms of bias reduction; some thoughts on the consequences on the variability of the estimates are also proposed.
Added Value: A step toward the application of the abovementioned innovative approach to data collection.