Occupation coding during the interview

Currently, most surveys ask for occupation with open-ended questions. The verbatim responses are coded afterwards, which is error-prone and expensive. We describe an alternative approach that allows occupation coding during the interview. Our new technique utilizes a supervised learning algorithm to predict candidate job categories. These suggestions are presented to the respondent, who can in turn choose the most adequate occupation. 72.4% of the respondents selected an occupation when the new instrument was tested in a telephone survey, implicating potential cost savings. To aid further improvements, we identify a number of factors how to increase quality and reduce interview duration.