Scottish Household Survey: response rates, reissuing and survey quality

This paper assesses the impact of reissuing on survey estimates using data from the Scottish Household Survey, 2014 and 2016.


Footnotes

1. This could be because the respondent was unable/would prefer not to complete the interview at that time, or because the interviewer failed to contact the respondent. The initial interviewer is required to make 6+ calls on different days and at different times to try and establish contact.

2. Such as Freeth & Sparks, 2004.

3. Details of the approach to weighting can be found in the 2016 technical report. https://www2.gov.scot/Resource/0054/00546545.pdf

4. These are base sizes before any adjustment for the design effect.

5. Sample size divided by the square of the design factor.

6. Sample size before adjustment for the design effect.

7. Sample size before adjustment for the design effect.

8. Estimates and confidence intervals for the Issue 1 sample assume that the overall achieved sample size and design effects would be the same - in other words, what we would expect the final data to look like had no reissuing been carried out but the response rate targets had been adjusted down by around 10 percentage points.

9. Sample sizes for Figures 4.1 to 4.4 are as provided in Table 4.1

10. Full details of each of these are provided in Tables A1.1 to A1.10 in Appendix 1.

11. Central Region includes Stirling, Falkirk, & Clackmannanshire council areas.

12. 22% and 25% for single adults in 2014 and 2016.

Contact

Email: shs@gov.scot

Back to top