Exploring the link between obesity and food purchasing behaviour in the West Midlands: Report commissioned by the

Department of Health West Midlands

Professor Andrew Fearne

Centre for Value Chain Research

KentBusinessSchool

University of Kent

Canterbury

Kent CT2 7PE

November2010

CONTENTS

Page

Executive Summary 3

  1. Background 4
  1. Objectives 4
  1. Methodology 4

3.1 Stage 1: Identification of target area 5

3.2 Stage 2: Attitudes, perceptions and purchase drivers 5

3.3 Stage 3: Barriers and enablers and the potential impact of targeted interventions 5

4.Results 5

4.1 Stage 1 – dunnhumby data 5

4.2 Stage 2 – Focus groups 6

4.3 Stage 3 – Surveys 9

  1. Conclusions and Recommendations13

Appendices14

Executive Summary

  • This research project is designed to assist the Department of Health West Midlands (DHWM) by providing new insights into the food purchasing behaviour of supermarket shoppers in the West Midlands and how this impacts on perceptions of diet-related health, the incidence of obesity and the barriers to behaviour change in the region. In the process, it explores the methodological value of using supermarket panel data for the identification of target segments and potential interventions to tackle obesity.
  • The research methodology adopted comprised three stages – identification of the target area (through the analysis of supermarket panel data), exploration of the target group’s attitudes towards food purchasing, preparation and consumption (through a series of focus groups), and quantification of the barriers and enablers to ‘healthy’ eating and the potential impact of targeted interventions(through a surveyadministered face-to-face with young women in the target group and on-line with a panel of shoppers representative of the wider population of supermarket shoppers).
  • The analysis of the supermarket panel data revealed that the two target groups – low income families and single parents on low incomes - consistently over-index with respect to the ‘unhealthy’ food products and consistently under-index with respect to‘healthy’ food products, relative to all supermarket shoppers in the Midlands region. The analysis also identified the Coventry (Ricoh Arena) store had the highest number of shoppers and second highest level sales amongst the two target segments. This store and the surrounding area was therefore chosen as the focal point for the study.
  • The focus groups revealed that food is perceived primarily as fuel and many of the discussants made reference to childhoods during which convenience foods were the norm and they received little training in the preparation of meals from scratch. There was a universal recognition that ‘unhealthy’ foods were to be avoided but an overwhelming view that people cannot help themselves and the system (manufacturers and retailers) conspire to make it so much easier for them to make ‘unhealthy’ choices and so much harder to choose healthier options. The consensus was that any intervention designed to stimulate behaviour change amongst this target group would have to address four fundamental barriers: cost (healthy foods were widely perceived as more expensive), time (in buying ingredients and preparing healthy meals) a lack of cooking skills (fuelled by the habitual purchase of convenience foods) and the perceived lack of taste of healthy options.
  • The results from the survey indicate that the target group (all of whom had BMIs over 30 and were classified as low income and/or single parents) were more inclined to make less healthy food choices than the control group. They also supported the hypothesis that the perceived cost of ‘healthy’ foods is likely to be the most significant barrier to purchase amongst the target group, a third of whom said they would like to eat more healthily but find it really difficult to change their behaviour - a state of inertia compounded for some families by the perceived difficulty associated with cooking healthily (or the perceived convenience of unhealthy food) and the resistance to healthy food (or preference for unhealthy options) amongst other family members and particularly amongst children. When asked how likely their food purchasing behaviour would change (become ‘healthier’) as a result of different interventions, the two options that attracted the highest probability were ‘discounts on healthy foods’ and a ‘tax on unhealthy products’. However, the significant interest in cooking demonstrations and recipe ideas for healthy foods amongst respondents from the target group suggests that for many of them, reducing the most significant barrier (perceived cost/price premium) would still leave them with two substantive hurdles to overcome – what to cook and how to cook.
  • Changing the behaviour of the target groupis likely to remain a substantial challenge, particularly in the current economic climate. The results of this study suggest that a substantial proportion (around a third) of the target group are willing to change but find it ‘really difficult’ due to the dual barriers of perceived cost/price premium associated with healthy food and a lack of self-confidence/capability in the preparation of healthy foods that their families will enjoy. Interventions designed to tackle these two substantial barriers are likely to meet with success, as awareness of the problems associated with unhealthy diets and interest in finding solutions is growing. However, achieving an enduring change in behaviour takes time and resources. Both are running low, hence the compelling need for targeted interventions, focused on the most vulnerable. The analysis of supermarket panel data was an important part of this study, providing a benchmark of behaviour – actual not claimed – and a focal (geographical) area for the primary research. It’s use will be even more important in the evaluation of any interventions that this research will hopefully prompt.

  1. Background

Obesity is a problem of increasing significance in the West Midlands, which, according to 2010 figures from the Association of Public Health Observatories has the highest incidence rates in the UK with 29 per cent of adults being obese. The same figures also show that the UK is the ‘fattest’ EU member state.The causal factors are apparently simple, related largely to lifestyle and diet, yet changing the lifestyle and food consumption behaviour of individuals is a complex challenge there is little evidence to date that the battle against obesity is being won. A recent report by the Department of Health[1]highlighted the importance of designing intervention strategies that are relevant to distinct (regional) communities, recognising that behaviour change is complex and interventions must be effectively targeted at distinct consumer segments. This research project is designed to assist the Department of Health West Midlands (DHWM) by providing new insights into the food purchasing behaviour of supermarket shoppers in the West Midlands and how this impacts on perceptions of diet-related health, the incidence of obesity and the barriers to behaviour change in the region. In the process, it explores the methodological value of using supermarket panel data for the identification of target segments and differential interventions to tackle obesity in the West Midlands.

  1. Objectives

Before investing resources into the development of a range of intervention measures, DHWM is seeking a clear understanding of food purchasing behaviour and how this relates to the incidence of obesity, consumer attitudes towards diet and health and the implications for different interventions designed to stimulate and facilitate behaviour change amongst the most vulnerable households in the region[2]– single parents and low income households.

Specifically, the objectives of this research are to:

  • Determine, through the analysis of supermarket panel data, the penetration of ‘unhealthy’ food products amongst distinct (vulnerable) consumer segments in the West Midlands
  • Explore, through qualitative research, the drivers of ‘unhealthy’ food purchasing behaviour amongst distinct (vulnerable) consumer segments and their attitudes towards different interventions designed to change their food purchasing behaviour (i.e. improve their diets)
  • Quantify, through survey research, the potential impact of different interventions on the food purchasing behaviour of different shopper segments
  • Provide DHWM with an evidence base from which they can develop targeted intervention strategies for reducing the incidence of obesity in the region
  • Establish the methodological value of combining supermarket panel data with other data sources (qualitative and quantitative) to identify distinct target segments with distinctly different needs in terms of interventions to tackle obesity
  1. Methodology

Previous research in this area has been predominantly qualitative, designed to explore awareness of and attitudes towards obesity – its causes and how to deal with it – amongst different social groups (e.g. individuals and families from different socio-economic and ethnic backgrounds) and the perceived impact of interventions on claimed behaviour. This research builds on previous researchby focusing onobserved purchasing behaviour and reported attitudes towards, perceptions of and intentions with respect to healthy foods and the purchasing/consumption thereof (the barriers and enablers to improving the diets of distinct consumer segments).

Theresearch methodology adopted comprisedthreestages – identification of the target area, exploration of consumer (shopper) attitudes perceptions and purchase drivers, and quantification of the barriers and enablers to ‘healthy’ eating and the potential impact of targeted interventions.

3.1 Stage 1 –Identification of target area

The first stage involved the analysis of dunnhumby (supermarket panel) data for the West Midlandsregion[3] for the 52weeks ending January 31st, 2010. A list of indicative ‘unhealthy’ products was identified, in discussion with DHWM and the National Social Marketing Centre (NSMC), and the dunnhumby data was analysed to identify thosestores in which the penetration of these products amongst the target segments was the highest. These stores were categorized as ‘unhealthy’ hotspots amongst our target groups - single parents and low income familiesin the West Midlands region. From this list of potential target areas (see Appendix A), one (Ricoh Arena, Coventry) was selected, in discussion with staff from DHWM and theNSMC, for the recruitment of participants in stage 2 and survey respondents in stage 3.

3.2 Stage 2 – Exploration of attitudes, perceptions and purchase drivers

In order to explore the attitudes, perceptions and purchase drivers associated with ‘unhealthy’ foods a series of (six) focus groups were conducted with people recruited from the target supermarket (Coventry Ricoh Arena) with the highest penetration thereof. Six groups were conducted with young mothers (single, married and living with a partner) who shop regularly at the store and had a Body Mass Index (BMI) greater than 30(classified as obese). In addition, two control groups were conducted with young mothers, all of whom had a BMI well below 30[4]. A discussion guide was developed (see Appendix C), in consultation with NSMC and DHWM, and the focus groups were video recorded for the purpose of transcription and content analysis, which was undertaken in NVivo. The groups were conducted over a period of two weeks (March 2nd – March 10th).

3.3 Stage 3 – Quantification ofbarriers and enablers and the potential impact of targeted interventions

Having gained an understanding of the attitudes, perceptions and purchase drivers of the target groups – single parent and low income households, the third stage of the project soughtto quantify the likely impact of targeted interventions designed to address the key barriers and enablers to buying healthier foods identified in stage 2. This involved the administration of a survey to two independent cohorts: a) an on-line panel of around 200 supermarket shoppers, with a broad range of socio-demographic characteristics and BMIs, and b) a sample of around 200 shoppers with BMIs over 30, interviewed face-to-face outside the Tesco (Coventry Ricoh Arena) store.

  1. Results

The dunnhumby data used for the identification of the target area (stage 1) is presented in Appendix A. The details of the focus groups are presented in Appendices B-D. The tabulated survey results are presented in Appendix E. What follows is a summary of the key findings.

4.1 Identification of target area

In order to identify the target area for this study, an analysis was undertaken of the dunnhumby (supermarket panel) data for the Midlands (TV advertising) region. This involved three steps: a) identification of categories of foods that could be considered representative or indicative of what could be classified as ‘Healthy’ and ‘Unhealthy’ foods, b) the geo-demographic profiling of supermarket buyers of these two groups, and c) the identification of the store with the highest footfall of people buying ‘Unhealthy’ foods, for the purpose of recruitment – for the focus groups and the face-to-face interviews.

The categorization of healthy and unhealthy foods was done in consultation with staff from DHWM and resulted in the identification of fourteen ‘unhealthy’ categories (351 products) and twelve ‘healthy’ categories (216 products) – see Table 1, Appendix A. These categories are not proposed as being in any way exhaustive but are believed to capture a substantial proportion of those products that dieticians would regard as unambiguously ‘healthy’ – rich in vitamins, low in salt, fat and sugar – and those that would be widely regarded as ‘unhealthy’ – low in vitamins and high in salt, fat and sugar.

Having identified these two distinct categories, the data was analysed to identify the extent to which the two target groups indentifiable within the dunnhumnby data (Cameo geo-demographic segmentation) as ‘poorer families and single parent households’ and ‘poorer council tenants – many single parents’, (which together account for around 20% of the UK population) within the Midlands region under/over index. The data presented in Tables 2 to 5 (Appendix A) and the shopper profiles presented in figures 1 and 2 (Appendix A) show that these two shopper segments consistently over-index with respect to the ‘unhealthy’ SKUs and consistently under-index with respect to the ‘unhealthy’ SKUs, relative to all supermarket shoppers in the Midlands region.

The final step in this first stage of the primary research was the identification of a focal area/store in which to undertake the recruitment, interviews and pilot intervention. This was achieved by examining the store level sales of ‘unhealthy’ SKUS (Table 6, Appendix A), which revealed that the Coventry (Ricoh Arena) store had the highest number of shoppers and second highest level sales amongst the two target segments. This store and the surrounding area was therefore chosen as the focal point for the study.

4.2 Exploration of attitudes, perceptions and purchase drivers

In order to explore the attitudes, perceptions and purchase drivers with respect to food in general and healthy/unhealthy food in particular, six focus groups were recruited from the Ricoh Arena shopping complex and conducted over two days (March 2nd and 10th 2010). The criteria for recruitment was that the participants were under 30yrs with children (aged less than 11yrs) at home and with a BMI over 30. Two of the groups were run as a control group and comprised young mothers with BMI below 30. The details of the focus group participants are presented in Appendix B.

A discussion guide (see Appendix C) was designed, covering aspects of food purchasing behaviour, perceptions of healthy/unhealthy foods and the barriers/enablers to changing behaviour – towards a healthier diet. All six groups were recorded and the transcribed and the detailed verbatim from each of the groups are presented in Appendix D. What follows is a summary of the key findings.

4.2.1 Attitudes towards food shopping

The consensus from the four main groups (comprising people from the target segments with BMIs over 30) was that food shopping was a stressful activity, particularly when accompanied by children and very rarely enjoyed. The level of involvement in food shopping or food preparation was not formally assessed, but it was evident from the majority of comments that food was perceived primarily as fuel and many of the discussants made reference to childhoods during which convenience foods were the norm and they receive little training in the preparation of meals from scratch:

Box 1 – Attitudes towards food shopping

“I dread it – I walk in for a few bits and come out with a trolley full!”

“If you go with the kids at the weekend it’s a nightmare – I prefer to go during the week when the kids areat school”

“I have to take my son with me half the time and all he does is scream. He picks up everything he wants then when he gets home he decides he does not like it. He is quite big so I have to check everything so it takes even longer”

“It’s alright but when I go to the big Tesco, it might sound silly, but I find myself a bit lost sometimes. It takes too long and I find I buy stuff that I don’t really need”

“I try and plan meals for the week ahead...but I still don’t stick to it”

“I go up and down all the aisles and come back with a load of stuff. I try and rotate things in the fridge – I am getting better at it but I used to throw out lots of stuff. I am trying to manage my money better.”