The paper is split into three main sections: a review of existing data sources; an analysis of characteristics of the main reference datasets used for the production of research and official statistics; and an investigation into the imputation code used by the LPC to produce official statistics.
The finding of the review of ASHE, LFS and BHPS is that the notion of rounding in wage setting and wage reporting is confirmed, and the sources of that rounding detailed. For employers, one suggestion is that rounding wages is a response to administrative burdens. For employees, rounding in survey reponses is largely unrelated to any characteristics of the respondent; this means that, although the LFS is clearly subject to measurement error, it can be assumed to be random for many statistical purposes. The report recommends that LPC should focus its attention on occupational differences in wages rather than industrial sector.
On imputation, the study reviewed the current LPC code for addressing gaps in the data available for describing subsections of the community. The study found that while the code has a number of problems, the alternatives (no imputation, or using an alternative wage measure) seemed to less justified statistically. The report recommends that the LPC review and rewrite the code.