Do successful persons remain in the sample? A combination of sequence analysis and log-linear path models to analyze the influence of occupational success on panel attrition

All researchers conducting life course panel studies face the problem of panel attrition. Panel attrition may lead to a biased sample if it is correlated with the outcome of the life courses. Studies of occupational trajectories of graduates and other labor market studies in particular may be affected by a "success bias." This means that unsuccessful persons have a lower response rate than successful persons. In this article the authors suggest a combination of two methods that can be used to evaluate the success bias in panel studies. In the first step an application of the optimal matching method is employed as a dynamic measure of occupational success. In the second step the causal relationship of the vocation or profession the person has been trained for, the cluster of unsuccessful "losers", and the response rare is investigated by means of a log-linear path model. The analysis demonstrates that,the data set examined is not affected by a success bias.