This statistical briefing note explains the research behind the BBC1 Panorama programme 'Too poor to stay warm' broadcast on 21 March, 2016. The programme highlighted the problems of fuel poverty in the UK and was based on calculations produced by the PSE project team. This outlines how the results, discussed in the programme, were produced.

This statistical briefing note compares differences between Great Britain and Northern Ireland in items reported as ‘necessary’ using the harmonised UK Omnibus dataset. Heat maps and relative risk ratio calculations were carried out separately for all adult and child items and activities. No major differences are apparent between respondents.

The statistical briefing examines the implications of a programming error that resulted in higher than expected amounts of missing data in the ‘sort card’ sample of the June 2012 Northern Ireland Omnibus Necessities Module. It concludes that the missing data can be considered to be MAR (Missing At Random) and they are ignorable for analyses of the proportion of respondents considering an item or activity to be a necessity as long as both the missing data and the don’t knows are excluded from the numerator and denominator i.e. if the don’t knows are set to ‘missing’.

How can we tell whether any differences between different groups, such as men and women, in their attitudes to necessities is down to chance or not? This statistical note explains why the PSE team uses 'relative risk' rather than 'hypothesis testing'.