Trends and seasonality in food buying

As individuals and as a population, the food we buy and eat is not static, but changes over time, and varies with the time of year.  Many things, prominent among which are cost and availability, influence these changes.  It is likely that changes in food prices will affect lower and higher income households differently.  The COVID-19 pandemic has also had a substantial effect on choices in buying food.  Data from the Kantar survey has been used by BioSS researchers, collaborating with SRUC and Rowett scientists, to explore what influences variation in food buying behaviour.  Income is found to be a dominant factor.  Decisions impacting on meat and dairy products, but also those associated with products rich in fat or sugar, are found to have the largest impact on food-derived greenhouse gas emissions.

Dietary patterns are most often studied through surveys of intake over a few days, or food frequency questionnaires, and so can offer little information on longer term, longitudinal changes within individuals.  The Kantar consumer behaviour survey, although originally developed to provide commercial intelligence, does provide this, by capturing buying patterns in households over many months.  These data, however, come with limitations: data are reported at a household rather than an individual level, and food purchased is a proxy for, but not the same as, food eaten.

In collaboration with Rowett and SRUC scientists, BioSS researchers explored the patterns in aggregated Kantar data for 4-week periods between 2013 and 2020, for three household income tertiles.  Principal component analysis was used to draw out the sources of variation in what was very high dimensional food group and nutritional composition data.  The analysis made it apparent that income group had the biggest influence on behaviour, followed by a specific trend seen over the years.  Seasonality, while significant, had the least effect. Interactions between these factors could be detected, though not for income group and seasonality, but were much smaller in magnitude. 

The environmental impacts of such changes were examined by calculating the greenhouse gas emissions associated with product choices, and different food groups were assessed in terms of the extent to which they were driving the observed variation.  Results were as expected, with meat and dairy products clearly having the biggest effect, and foods high in fat and sugar also explaining much of the variation.

A presentation of project outcomes was delivered to five members of the Nutrition Science and Policy team at Food Standards Scotland (FSS), including public health nutrition and science advisors; project findings and ideas for policy improvement were discussed and co-constructed for further investigation in both the previous and current RESAS programmes.  The project output provided an analysis and evaluation of both long-term Scottish dietary patterns through an equity lens and of the implications of a policy change on their nutrient quality and greenhouse gas (GHG) emissions, with recommendations for improvement of dietary advice.  

This work was done in collaboration with Vanessa Rungapamestry and Wisdom Dogbe at The Rowett Institute and was funded under the Scottish Government's Strategic Research Programme for environment, agriculture and food.

Graham Horgan

For more information contact: