Scottish Employer Perspectives Survey 2019: technical report

Technical report accompanying the Scottish EPS 2019 research report providing background information on the methodology used.


1. The last UK-wide EPS was carried out in 2016, principally commissioned by the Department for Education. Earlier editions of the UK-wide EPS were principally commissioned by the UK Commission for Employment and Skills (UKCES).

2. Note that the number of interviews sums to greater than 2,652 as ROAs are defined by Local Authorities (LA), and some LAs are in more than one ROA. Also note that the population data are rounded to the nearest 100.

3. The initial counts provided were incorrectly based on employees not employment. Although establishments with one owner or working proprietor but no employees were correctly excluded, those with one employee but no owners or working proprietors were incorrectly included if they were part of a chain, while those with two (or more) owners or working proprietors but no employees were incorrectly excluded. As a result the population counts from IDBR had to be re-run. The net effect was minimal with around eight thousand incorrect exclusions but almost the same number of incorrect inclusions, and the total number of eligible establishments fell slightly from 151,400 to 151,168. However, there were quite large differences within size band, and the number of establishments with employment of 2-4 in the correct, re-run data (75,977) was lower than in the first run of the data (80,403). This was accompanied by an increase in those within the 5-9 category.

A useful check to reduce the risk of this mistake happening is to request unit and employment counts for small employers by the individual employment size number (i.e. those employing two people, those employing three, those employing four), rather than at the grouped 2-4 size band category. The employment figure for those establishments with employment of exactly two should be twice the number of establishments, and so on, and if not it is likely that the figures show employee figure not employment.



Back to top