Better or Longer? The evolution of weekly number of completed interviews over the fieldwork period in the European Social Survey

Developing adaptive or responsive fieldwork design relies on the availability of auxiliary information. An example of such data are paradata that have obtained growing attention in recent years to detect, prevent or correct for survey errors. In this context, fieldwork length and timing of contact attempts are typical time-related paradata and are easy to derive from contact forms or call records. In general more effort should lead to better results. Therefore longer fieldwork and carefully timed contact attempts are expected to be positively correlated with fieldwork performance indicators such as response rate, contact rate or refusal rate. We analyse the six first rounds of the European Social survey for which contact forms are available for most of the participating countries containing information about each contact item (date, time of day, outcome, etc.). At first glance, a negative correlation is observed between fieldwork performance indicators and fieldwork length and timing effort. Timing effort means compliance to the ESS guidelines to increase the chance of contact -at least four contact attempts, at least one in the evening and one on the weekend, best spread over at least two weeks. Concentrating on the contact rate, controlling for the country’s survey climate measured by the percentage of contact at the first contact attempt, we found positive relations with fieldwork length and timing effort; non-significant for the length but significant for the care taken in the timing of the contact attempts. This result shows the importance of carefully planning and implementing the contact strategy before the start of the fieldwork process. Such an adaptive design, based on timing of contact attempts seem to have a greater effect than extending the fieldwork effort to increase response rates.status: publishe