Antje Kirchner

Messung und Erklärung von Schwarzarbeit in Deutschland - Eine empirische Befragungsstudie unter besonderer Berück- sichtigung des Problems der sozialen Erwünschtheit* Measuring and Explaining Undeclared Work in Germany - An Empirical Survey with a Special Focus on Social Desirability Bias

Zusammenfassung Der vorliegende Beitrag geht der Frage nach, wie das Ausmaß von Schwarzarbeit in Deutschland im Rahmen von Befragungen der allgemeinen Bevölkerung mçglichst valide geschätzt werden kann. In einem experimentellen Design wird die konventionelle direkte Befragungstechnik mit zwei Spezialtechniken, der Randomized-Response- Technik (RRT) und der Item-Count-Technik (ICT), verglichen. Die RRTund die ICT wurden für die Messung besonders heikler Verhaltensweisen entwickelt und sollen durch eine Erhçhung der Anonymität in der Interviewsituation sozial erwünschtes Antwortverhalten reduzieren. Unsere Befunde zeigen, dass die häufig angenommene Wirkung der beiden Spezialtechniken auf die Bereitschaft der Befragten, sozial unerwünschtes Verhalten zu berichten, nicht eindeutig ausfällt. Zudem werden theoretisch bedeutsame Einflussfaktoren von Schwarzarbeit diskutiert und deren Wirkung im Rahmen von multiplen Regressionsanalysen empirisch überprüft. Neben Gelegenheitsstrukturen sind vor allem soziale Normen gute Prädiktoren für die individuelle Entscheidung schwarzzuarbeiten. Summary This article explores methods used to obtain a higher validity in estimates of the prevalence of undeclared work in Germany in surveys within the general population. Using an experimental design two “dejeopardizing” techniques are compared as alternatives to direct questioning when asking sensitive questions: the randomized response technique (RRT) and the item count technique (ICT). These techniques were specifically developed to reduce misreporting on sensitive topics: The goal is to elicit a higher proportion of honest answers from respondents by increasing the anonymity of the question-and-answer process. Our results suggest that neither RRT nor ICT provide unambiguous results with respect to more successful elicitation of reports of socially undesirable behavior. In addition, the theoretically significant influence of background variables is investigated empirically by means of multiple regression. Factors which foster illicit work are, aside from opportunity structures, social norms, which contribute significantly to the explanation of individual decisions to engage in undeclared work.

An Introduction to Machine Learning Methods for Survey Researchers

Machine learning techniques comprise an array of computer-intensive methods that aim at discovering patterns in data using flexible, often nonparametric, methods for modeling and variable selection. These methods offer an expansion to the more traditional methods, such as OLS or logistic regression, which have been used by survey researchers and social scientists. Many of the machine learning methods do not require the distributional assumptions of the more traditional methods, and many do not require explicit model specification prior to estimation. Machine learning methods are beginning to be used for various aspects of survey research including responsive/adaptive designs, data processing and nonresponse adjustments and weighting. This special issue aims to familiarize survey researchers and social scientists with the basic concepts in machine learning and highlights five common methods. Specifically, articles in this issue will offer an accessible introduction to: LASSO models, support vector machines, neural networks, and classification and regression trees and random forests. In addition to a detailed description, each article will highlight how the respective method is being used in survey research along with an application of the method to a common example. The introductory article will provide an accessible introduction to some commonly used concepts and terms associated with machine learning modeling and evaluation. The introduction also provides a description of the data set that was used as the common application example for each of the five machine learning methods.