Wage Data Collected by Telephone Interviews : An Analysis of the Item Nonresponse Problem and its Implications for the Estimation of Wage Functions

In this paper, the item nonresponse problem associated with wage data collected by telephone surveys is analysed. Using data from the 1998 Swiss Labour Force Survey (SLFS) and data on interviewers, it is shown that the response decision can only partially be explained by the chracteristics of the interview situation, the respondent, and the interviewer. This suggests that the response inclination is, to a large extent, randomly distributed amongst the underlying population. It is therefore argued that wage functions estimated using only the observed wage data are not biased by the large (wage) item nonresponse encountered in telephone interviews.

[http://econpapers.repec.org/article/sesarsjes/2000-i-4.htm]