Socio-demographic, smoking and drinking characteristics in GB: A comparison of independent telephone and face-to-face Smoking and Alcohol Toolkit surveys conducted in March 2022

,


Introduction
In March 2020, to deal with the rapidly growing number of seriously ill COVID-19 patients, the UK government introduced a set of measures to limit mixing between households (UK Government, 2020).Along with other household surveys, the methodology of the monthly Smoking and Alcohol Toolkit Study (STS/ATS -surveys conducted by Ipsos Mori) adapted quickly and switched from face-to-face to telephone interviewing in April 2020.This ensured that data collection continued during the COVID-19 pandemic.Due to successive waves of COVID-19 infections, telephone data collection for the STS/ATS is still being done by telephone in 2022.
It is possible that the change in data collection to telephone affected key indicators of smoking and alcohol use for at least two reasons.First, a participant might give different answers to the same question when asked via telephone rather than face-to-face.For instance, participants may be less willing to disclose sensitive information over the phone, or more likely respond they "don't know".Secondly, the characteristics of participants recruited for telephone interviews might differ systematically from those recruited face-to-face (Ellis and Krosnick, 1999).In March 2022, Ipsos Mori conducted an independent face-to-face survey in parallel to the ongoing telephone survey.
In this report we describe key socio-demographic, smoking and alcohol use indicators obtained via the telephone and face-to-face surveys to explore the possible impact that the change in survey modality may have had on key estimates.
Data on core socio-demographic (age, sex, social grade and region), smoking (including nicotine product use) and alcohol measures are presented first for the overall sample in Great Britain and then within each nation (England, Scotland and Wales).Further details on the design of the STS, including face-to-face and telephone sampling techniques, can be found elsewhere (Kock et al., 2021).Sample weighting uses the rim (marginal) weighting technique to match the Great Britain sociodemographic population profile relevant to the time each monthly survey was collected.Thus, respondents with characteristics that are under-represented receive a larger weight, while those who are over-represented receive a smaller weight.Comparisons with sales data and other national surveys show that the STS recruits a representative sample of the population with regard to key demographic variables and smoking indicators (Fidler et al., 2011;Jackson et al., 2019).
The sample dataset consists of STS/ATS respondents recruited during parallel telephone and face-to-face survey waves in March 2022.

Measures
The following measures were assessed among all respondents: age, sex, social grade, geographic region, smoking status and alcohol use (Alcohol Use Disorder Identification Test, AUDIT).Among past-year smokers, the use of nicotine products (nicotine replacement therapy, electronic cigarettes, heated tobacco products and nicotine pouches) was assessed.
Specific wording for each measure is provided in the study protocol, registered on the Open Science Framework (OSF) at https://doi.org/10.17605/OSF.IO/2Z3TE.

Analyses
We estimated unweighted statistics and 95% confidence intervals for age, sex, social grade and region.We also estimated weighted prevalence statistics and 95% confidence intervals for each measure for Great Britain, England, Scotland and Wales separately, except nicotine product use, which was assessed for the overall Great Britain sample only due to small samples from Scotland and Wales.

Results
In March 2022 in Great Britain, 2,607 and 2,064 adults aged 16 and older participated in respective telephone and face-to- face surveys for the STS/ATS.
The weighting procedure applied to both is designed to produce samples that match the population on these dimensions, and consequently the weighted estimates were very similar (Table 1).Weighted estimates in the telephone and face-to-face surveys differed by around one percentage point or less for response categories of daily smoker, non-daily smoker, pipe/cigar smoker, and stopped last year (Table 2), which meant key estimates of current cigarette smoking (tel: 14.44%%; f2f: 15.91%), current smoking (tel: 16.47%; f2f 16.70%:), pastyear smoking (tel: 18.45%; f2f:18.21%)and quit rate in past-year smokers (tel: 10.8%.f2f: 9.0%) were all similar between the surveys.There were differences in the estimates for never smoking and stopped more than a year ago between the surveys but the combined estimate for never or long-term ex-smoking was similar.