Mechanisms of Panel Conditioning in Longitudinal Studies_Silber (PaCo_Si)
Longitudinal surveys are the major instrument of empirical social sciences through which both stable and changing societal patterns are observed, theories are tested, and evidence-based policy recommendations are derived. Panel conditioning effects (PCE), that is changes in actual behavior, attitudes, or knowledge, or changes in response behavior as a result of previous survey participation, endanger these most important aims of longitudinal research: The valid measurement of stability and change. The research gap in the area of PCE exists because of a lack of studies that allow testing for causal inferences. Specifically, in order to understand the mechanisms of PCE as well as to propose means to reduce measurement error due to PCE it is essential to use an experimental study design. Thus, we propose an innovative study design, which encompasses a 2 X 6-factorial experiment and a 2-factorcal experiment. The first factor manipulates the frequency of answering the same questions (from one to six times) and the second factor varies the measurement interval (2-months and 4-months between waves). This study design will be implemented in two existing panel studies: first in the GESIS Panel, an offline recruited probability-based mixed-mode (online and mail) panel and second an online access panel. We selected the online mode, as the main survey mode because this mode becomes increasingly important for panel study. The overall goal of the proposed project, therefore, is to better understand the magnitude (i.e., effect size) of PCE and the mechanisms that cause PCE in order to advance the theoretical and empirical knowledge, as a basis for providing recommendations and best practice advice for researchers conducting longitudinal studies. The project is organized along three objectives. The first objective is to differentiate between different mechanisms leading to PCE: changes in reporting attitudes, behaviors, and knowledge due to reflection, survey satisficing and social desirability bias. The second objective is to identify the effect size and the significance of PCE in relation to question types represented in social surveys (attitudes, behaviors, and knowledge), different number of times the questions are asked (conditioning frequency), and different time intervals between the questions (conditioning interval). The third objective is to give recommendations for correction methods that will function as best practice advice. These recommendations will help to decrease measurement inaccuracy for the different question types. The project will also have a decisive impact on evidence-based practice because it will help large-scale longitudinal studies with substantial public policy implications to account for potential measurement errors.