Survey nonresponse research: appendices

Footnotes

1 Press and Journal (13 December 2021) ‘New blood needed: Plea as Scottish donor numbers at their lowest this century’

2 Meeting of the Parliament: 12/01/2022 | Scottish Parliament Website

3 See Scottish Government (2024) Households in Scotland by housing tenure: SHS 2023, additional tables.

4 See Scottish Household Survey: response rates, reissuing and survey quality 2018 and SJCS methods papers on response rate and survey bias

5 Issues of interviewer capacity on SHeS also impacted on the proportion of addresses reissued.

6 The key inconsistency in the response rate calculations was how addresses with outcomes where eligibility was uncertain were treated.

7 The small user PAF was used for the matching in Scottish Index of Multiple Deprivation (SIMD) 2020 and the 6-fold urban/rural indicator using datazones.

8 Scottish Index of Multiple Deprivation 2020 - gov.scot

9 Scottish Government Urban Rural Classification 2020 - gov.scot

10 This achieved a response rate of 77%.

11 For 1999 to 2011, the SHS sample was drawn for two-years sweeps, and then allocated to individual fieldwork years.

12 The first SHCS ran in 1991. We have not yet managed to track down the technical report

13 For SHS 1999 to 2011, response rate are shown across biannual sampling waves. For the SCJS, fieldwork is conducted April-to-March rather than Jan-to-Dec. Estimates have been charted against the year of the first three quarters of fieldwork.

14 This is due to a number of reasons: Ethical considerations around reissuing, especially the proportion of addresses were some had refused to take part; Evidence that reissuing had a marginal impact on results and on nonresponse bias; Consequently, because reissue interviews are more expensive than first issue interviews, it meant that the level of use did not provide value for money.

15 We do not have data on the extent of reissuing in SHeS post-Covid.

16 NA indicates not available. “X“ indicates survey did not run in this year.

17 96% of non-contacts at first issue were reissued in SHS 2015. See Supplementary Table 1 at the end of this Appendix.

18 Non-contact as proportion of all eligible addresses.

19 ‘Ill at home’ increased from 0.7% to 1.0%, ‘in hospital’ from 0.8% to 1.4%, physically or mentally unable to take part from 1.8% to 2.7%, and ‘other reasons’ from 1.9% to 2.6%

20 The data available on first issue response rates varies by survey. For the SHS, 2012-2019 and 2022,2023. For SHeS, 2012, and 2015 to 2019. For SCJS, 2018 and 2022 only.

21 Data for the SCJS is only available for one wave, 2018, but shows a similar pattern.

22 Because data on first issue response is available for different waves of the different surveys, we analysis of change over time separately for the SHS and SHeS.

23 A similar pattern of a relatively uniform drop was seen for the SCJS between 2018 and 2022.

24 As reissuing strategies are more likely to be focused on achieving interview targets in Scotland or at sub-national level, than on minimising variations in nonresponse, they are less likely to impact patterns of nonresponse for individual/household-level factors than area-level factors.

25 A notable exception is the 45-54 age group where the ratio was generally stable between 2012 and 2019 (1.01 and 0.99 respectively) but saw a drop between 2019 in 2022 (from 0.99 to 0.91).

26 See Office for Statistics Regulation website for an explanation of these designations.

27 An indicator of level of exposure to nicotine, from tobacco.