Methods
Monthly household surveys
Each month involves a new representative sample (16 and over) of ~450 respondents.
Due to the pandemic, surveys are conducted by telephone (rather than face-to-face), and until December 2021, among adults aged 18 and over.
From April 2020 (switch to telephone), Ipsos started recording ~<5% cases as missing for social grade. To derive weight variable (which uses social grade), Ipsos assumed missing = social grade C1. To ensure analyses using weight and social grade reflect target weighed profile, social grade was recoded to reflect this assumption. Different assumptions for missing social grade cases were explored in a parallel telephone and face-to-face (with no missing social grade) wave, and this assumption found least distorting.
Kock et al., (2021). Protocol for expansion of an existing national monthly survey of smoking behaviour and alcohol use in England to Scotland and Wales: The Smoking and Alcohol Toolkit Study. Wellcome Open Research, 6(67), 67.