Why does a new consent procedure during an ongoing longitudinal study influence data quality

Christoph Homuth, Monja Schmitt, Christian Lorenz, Daniel Mann

Abstract


The quality of social studies largely depends on the representative nature of the underlying sample. Panel mortality, for this reason, is a particularly problematic phenomenon in longitudinal studies. It cannot be tackled through a larger initial sample like in simple cross-sectional studies, but mainly by maintaining a steady willingness to participate. Since ministerial requirements at the state level in Germany more and more frequently require repeated renewal of the consent forms signed by participants of longitudinal studies in the field of institutional education, questions arise concerning the effect of this procedure on the willingness to participate. Likewise, it was necessary to renew the consent of parts of the sample within the context of two longitudinal studies by the Bamberg research group Bildungsprozesse, Kompetenzentwicklung und Selektionsentscheidungen im Vorschul- und Schulalter (BiKS). By analyzing the panel participation and drop-out rates before and after the additional consent we show the size of the negative effect on sample size. By using a counterfactual analysis of the participation probabilities of several subgroups we can estimate the negative effect on sample composition.

Keywords


Sample attrition; Selection bias; Active refusal to participate; Passive refusal to participate; Longitudinal study

Full Text:

PDF (Deutsch)


Copyright Waxmann 2009-2018 - Imprint
Journal for Educational Research Online/Journal für Bildungsforschung Online (ISSN 1866-6671)