Coding with ISCO, problems and solutions
|Convenor||Professor Dominique Joye (University of Lausanne )|
|Coordinator 1||Dr Evi Scholz (GESIS)|
|Coordinator 2||Mrs Cornelia Zuell (GESIS)|
The International Standard Classification of Occupations (ISCO) aspires to provide a systematic classification of all occupations exercised by the whole civilian working population.
Only 45 % of the occupational data collected could be automatically coded into ISCO. We started a project to code occupational data directly to ISCO in the field during the computer assisted data collection phase. Tests enabled that in three or four steps via a tree structure more than 80 % of all occupational data could be coded into ISCO in the field. An error analysis was possible.
Coding of occupational data is a crucial task in surveys. Usually, information on occupations is recorded in open-ended questions on job title, main duties and activities. Coding into ISCO-08 requires the assignment to one of the more than 500 individual 4-digit codes of the scheme. The quality and validity of the resulting variables are often low. In this paper, I will present the results of an experiment where different coding services have coded the same occupational survey data. The aim is to identify systematic differences and to develop coding rules to guarantee a higher coding reliability and
Currently, most surveys ask for occupation with open-ended questions. The verbatim responses are coded afterwards, which is error-prone and expensive. When textual answers have a low level of detail, exact coding may be impossible. We improve this process by asking response-dependent questions during the interview. Candidate job categories are predicted with a machine learning algorithm. When chances are high that the forecasted job is correct, the interviewer presents a short list of related jobs to the respondent. We compare the new technique to traditional coding practice and describe and quantify different error sources for both methods.
The goal of developing a new German Classification of Occupations (KldB 2010) was to create an up-to-date classification of occupations for Germany which on the one hand enables an almost one-to-one linkage to the ISCO 2008 and satisfies the occupation-specific structuring of the German labor market on the other hand. In the presentation we first give a short overview of developing KldB 2010, especially how we ensured the high compatibility to ISCO 2008. Secondly, we suggest how to code information on occupations into KldB 2010 and then convert into ISCO 2008.