Data editing and quality of daily diaries in the Italian Time use survey

In the Italian Time use survey, carried out by Istat (Italian National Statistical Institute) in 2002–2003, the coding process translates the sentences reported by the respondents into codes, but association text-code is not a one to one easy linkage, because the respondents describe the activities performed using the common language. So, a process of data editing is requested in order to improve data quality. The possibility to use sentences (recorded in the data entry) has definitely improved the data editing process. The main questions to analyse concern:

(a)	particular ancillary codes used by coders to point out daily diary’s problems;

(b)	incompatibility between codes and codes and between codes and text;

(c)	errors in the intervals’ sequences.

This paper shows methodological and contextual choices in order to build the process of data editing. In particular, it is focused on the analysis of techniques and deterministic rules adopted and on the analysis of imputation system to correct adequately the wrong codes, with reference to the entire diary. This paper was presented at the XXVI IATUR (International Association for Time Use Research) Conference held in Rome, 27–29 October 2004. The authors share contents and views expressed in this paper. However, the paragraphs are so drafted: Salvatore F. Allegra drew up the paragraphs 1, 2, 3, 7; Barbara Baldazzi drew up the paragraphs 4, 5, 6, 8. Salvatore F. Allegra and Barbara Baldazzi drew up the paragraph 9.