PROBING "DONT KNOW" ANSWERS: EFFECTS ON SURVEY ESTIMATES AND VARIABLE RELATIONSHIPS

Don't Know responses to survey questions are ambiguous becausethe same words are used by respondents to mean different things—ignorance,indecision, or uncertainty about the meaning of the questionasked. In order to clarify the meaning of such answers, surveyinterviewers are frequently trained to probe Don't Know answersat least once before the answer is considered final. We arguethat unconditional probing of Don't Know answers may not bea desirable practice, particularly as regards knowledge items.Large unintended effects on responses to four knowledge itemsresulted when two groups of interviewers, who administered thesame survey questions, probed Don't Know responses at differentrates. Strong evidence is provided in the article that the probingof Don't Know answers encouraged guesswork on the part of uninformedrespondents, giving rise to significant distributional differencesand differences in means across half samples for the affectedvariables. However, relationships between variables appear tobe largely unaffected by probing effects. In order to formulatevalid probing guidelines for interviewer training, further researchis needed to establish whether the findings for knowledge questionsgeneralize to factual questions.     

EMPLOYMENT AND DEADWEIGHT LOSS EFFECTS OF OBSERVED NONWAGE LABOR COSTS

To assess the employment effects of labor costs, it is crucial to have reliable estimates of the labor cost elasticity of labor demand. Using a matched firm‐worker data set, we estimate a long‐run unconditional labor demand function, exploiting information on workers to correct for endogeneity in the determination of wages. We evaluate the employment and deadweight loss effects of observed employers' contributions imposed by labor laws (health insurance, training, and taxes) as well as of observed workers' deductions (social security and income tax). We find that nonwage labor costs reduce employment by 17% for white collars and by 53% for blue collars, with associated deadweight losses of 10% and 35% of total contributions, respectively. Since most firms undercomply with mandated employers' and workers' contributions, we find that full compliance would imply employment losses of 4% for white collars and 12% for blue collars, with respective associated deadweight losses of 2% and 6%. (JEL J23, J32)