Online Program

More Information, Less Item Nonresponse? Implementation of a Question Experiment in the IAB Job Vacancy Survey (309626)

The IAB Job Vacancy Survey of the German Institute for Employment Research collects, for a randomly chosen last hiring process, detailed information on search and vacancy durations. The linkage of these specific date variables to a comprehensive set of variables about the recruitment process and to the employment biographies of the hired applicant enables researchers to conduct unique analyses of the structure of labor demand, the efficiency of search and recruitment methods, or the quality of the resulting job match. However, the duration questions themselves are rather burdensome as they ask about such things as the earliest hiring date, the date of the start of the personnel search or the date of a decision for an applicant. Consequently, the item nonresponse rate for these questions has been relatively higher over the years (up to 15 percent). As a result, important analyses of vacancy duration and labor market tightness are threatened by a loss of power and potential item nonresponse bias. To address this risk and evaluate possible solutions to reduce the item nonresponse rate, we conducted a question experiment by providing additional information (e.g. examples) to respondents for each of these questions. The additional information was intended to assist respondents in answering the questions and provide concrete examples to minimize confusion. The experiment was implemented as an online survey which ran parallel to the main (mail/web) survey, which did not include the additional question information. The experiment revealed surprising results, indicating that more information is not necessarily better for improving item nonresponse rates.