Consumer Insights Tracker: Chapter 3 Methodology

This report draws on data from two key tracking surveys, commissioned by the FSA via Ipsos UK’s online omnibus:  

1. COVID-19 consumer tracker survey – this monthly survey was established in April 2020 to track behaviour and attitudes towards food purchasing and consumption during the COVID-19 pandemic. Data collection continued on a monthly cycle until October 2021 (see Annex 1: Fieldwork dates and sample sizes – COVID-19 consumer tracker survey for details for each wave of the survey), and publication reports are published on the FSA website. From November 2021, some survey questions from the COVID-19 tracker survey were retired whilst others continued in the FSA’s new ‘Consumer Insights Tracker’. This was established in November 2021 and replaced the COVID-19 consumer tracker survey.  
2. Consumer Insights Tracker – this survey was established in November 2021, to monitor consumer behaviours and attitudes in relation to food availability, affordability and confidence in the food industry and the FSA (see Annex 2: Fieldwork dates and sample sizes – Consumer Insights Tracker survey for details of each wave). Many of the questions in this tracking survey were previously monitored in the COVID-19 survey (as outlined above), allowing the timeseries to continue (see Annex 4: Questionnaires for the full list of survey questions). The survey questions were developed jointly between FSA’s social science team and Ipsos UK. This report summarises findings from 13 waves of the Survey from March 2022 to March 2023 (including March waves in both years) and where relevant also draws upon data across all waves of the COVID-19 consumer tracker.  

The sample size for each wave of the COVID-19 and Consumer Insights Trackers is approximately 2,000 adults (aged 16-75), living in England, Wales and Northern Ireland. The sampling approach is based on quota sampling from an online panel and data is weighted to be representative of the adult population aged 16-75 living in England, Wales and Northern Ireland on key demographics: age, gender, working status and social grade. An elimination rule is applied to the online omnibus so that participants are not invited to complete consecutive surveys. For example, in a monthly survey, participants would be eliminated from the sample if they had completed the survey in the month prior. The full datasets for each survey, including demographic breaks, are available via the FSA data catalogue.  

3.1 Data analysis

Several demographic groups of particular interest for the FSA and its audiences were selected for additional analysis. These included: 

• households with children present and households with no children 
• household size (1, 2, 3 and 4+) 
• gender (male and female) 
• age (grouped into age bands: 16-24, 25-34, 35-44, 45-54, 55-75) 
• employment status (in work and not in work) 
• long-term health conditions and no long-term health conditions 

Participants were included in these demographic groups based on the answers given to specific questions in the online survey. In the case of employment status and health conditions, answer options were assigned to a broader category or ‘net’. Participants categorised as ‘working’ were employed or self-employed, those ‘not in work’ were unemployed, retired, studying or looking after the home and family. Participants with long-term health conditions reported having a physical and/or mental health condition or illness lasting or expecting to last for 12 months or more. 

To identify any differences between these demographic groups, analysis was conducted using March 2023 data, as well as comparing the responses of these groups in March 2022 and March 2023.  

Statistically significant differences at the 5% level were ascertained by the two-sample independent t-test. This testing determines whether a difference is ‘real’ or if it has occurred by chance because not everyone in the population has been surveyed. Throughout this report, in instances ‘significantly’ is used, this indicates that a difference is statistically significant at the 5% level. 

3.2 Note on interpreting the data

Results should be interpreted with care. All surveys are subject to a range of potential limitations including sample imbalances which are not easily identified and corrected through weighting. Surveys are also subject to participants’ interpretation of survey questions and response options. The data included within this report relies on participants’ self-reported behaviours. Errors could occur due to imperfect recollection, or participants’ tendency to over-report behaviours which are perceived as being desirable and under-report undesirable behaviours. Some survey question options were randomised and presented in a different order to help overcome order bias, which is the tendency for survey participants to pick the first option rather than reading all the options and choosing an honest answer. 

The project was quality assured by both the FSA’s Advisory Committee for Social Science (ACSS) and the FSA, and Ipsos UK quality assured the research process at key stages. This included reviewing survey wording to ensure questions would be understood by participants, and that questions met the research objectives. The survey script was checked by members of the Ipsos research team to ensure the questions were being asked as agreed. Sampling and data collection were subject to standard quality assurance protocols, including automated checks on completed questionnaires. Finally, data checks were carried out on data tables and reports. 

To limit the length of this report, demographic commentary has only been included in the section focusing on household food insecurity. Unless stated otherwise, where comparisons are made in the text between different demographic groups or between survey waves, only those differences found to be statistically significant at the 5% level are reported.  

3.3 Acknowledgements

Ipsos UK would like to thank the Food Standards Agency’s Social science team for their assistance throughout this project and in the writing of this report. We would also like to thank all those who took the time to participate in the survey.