Filming
So, how many kitchens were filmed and how much data was collected?
A whopping 101 kitchens were involved in the project! From cozy household setups to bustling food businesses, a diverse array of 70 households and 31 food establishments gave permission for filming to take place. Recruitment was led by recruitment agency Fieldsauce, with additional support from Acumen (who specialised in the recruitment of food businesses). Existing contacts were primarily used, with face-to-face recruitment used to boost numbers where needed. Quotas were set to ensure a diverse range of households and businesses were included in the study, although these were softened as it became increasingly difficult to recruit businesses.
In each kitchen, filming occurred over five days, including a weekend. But these weren’t your ordinary cameras! They were motion-sensitive, activated by movement around the kitchen space; rolling from the moment individuals entered the kitchen, and kept rolling for an extra 30 seconds after the kitchen was empty, just to make sure they didn't miss a beat.
Footage was coded by analysts from Lifestream, who assigned data labels to time-stamped filmed footage. Each label corresponded to food safety behaviours and their context. In total, 163 labels were developed over the course of the project, which were tagged to 5,464 videos, with a total of 292,216 data labels applied to over 650 hours of footage across household and food business kitchens!
Did people change their kitchen behaviour knowing they were on camera?
To reduce the likelihood of capturing “camera-aware” moments, only three of the five filmed days were used for analysis. Plus, we did not analyse the first day of footage, to ease participants into the filming process; allowing them to become more acquainted with the camera. In interviews, participants told us that they were so engrossed in their culinary routines they forgot the cameras were even there.
Ethics
What happened prior to filming?
Before the cameras even started rolling, the research team conducted a thorough ethics assessment. Every aspect of the study was accounted for.
A Privacy Impact Assessment (PIA) was paramount, which recorded:
- what data would be captured, how it would be held, handled, and when it would deleted
- the legal basis for data collection
- the range of assurances to protect participants
- how data breaches would be handled should they arise
- processes to identify and account for risks, big and small
Of course, consent was crucial – written consent was obtained from all participants known to be in the filming areas. Participants were clearly told not to discuss near the cameras anything they didn’t want recorded. Participants were made aware, during the consent process, that anything criminal or harmful stepped past the boundaries of participant privacy and would be reported if anyone were in immediate danger.
Nudity was also a concern. What if people were so engrossed in their daily routine that they forgot to wear clothes around the cameras? Well, to mitigate that (admittedly small) risk, automatic obfuscation technology was applied to screen nudity, and cameras could be switched off by participants or remotely if they needed extra privacy. Participants were not explicitly informed about the blurring technology, but they were made aware that the camera would record them when they entered the kitchen, and footage would be stored safely and only be seen by researchers. Any incidental filming of behaviours not relevant to the project was deleted within 48 hours.
What ethical issues arose once filming had started?
Whenever an issue reared its head, the KL2 research team – featuring the contractors mentioned at the start of this piece and of course the FSA – huddled together, brainstormed solutions, and made the necessary adjustments. As required, further risk assessments or amendments to the original PIA would be made to reflect any new or emerging challenges. These included for example:
- collecting data in schools
- obtaining consent to film in kitchens for 7 days where more filming was required
- using memory cards to store footage in cameras where Wi-Fi quality was weak
- capturing conversation
To deal with the audio data dilemma, it was decided that participants could provide separate consent for the collection of anything that was spoken aloud, which was not a requirement of participation in the study. Observing behaviour was the goal and while dialogue did help understand motivation, it was not wholly necessary.
What about data handling?
Protecting an individual’s privacy regarding data handling and access was an additional consideration.
There was a concern that businesses might have felt that they were being watched for regulatory purposes, so data access was severely restricted to assuage concerns. Participants could ask for timestamped footage to be deleted, if required. Plus, no feedback was given to participants about the quality of their food safety practices. The study’s purpose was for research: not for regulation or enforcement.
What additional assurances were applied to businesses?
It was also important to protect kitchen staff in businesses during the study. Consent needed to be obtained from all staff who might be filmed, and to prevent any feelings of obligation to participate via the business owner. Further, there was a risk of accidentally filming customers or filming workers while on their break and trying to relax away from the cameras.
To address these issues, several measures were set in place.
A take-home information sheet was provided for all staff members. Separate consent was obtained for each staff member and from the business owner; this process was entirely confidential, with business owners not knowing which staff did or did not give consent. Only businesses who provided full consent from all staff were invited to proceed in the study. This administrative process did create significant issues with recruitment: less than 3% of FBOs that were initially contacted to take part completed the research (1 in 32 businesses).
Cameras were situated to minimise accidental filming of public or private staff areas. This also required adaptation throughout the fieldwork. In cases where it was discovered during filming that members of the public were visible to the cameras (for example, through a serving hatch), blurring was used to address the problem, and footage was deleted where necessary.
Crucially, business owners were kept well away from the footage, most of it being streamed directly to our contractor’s servers (the rest was securely stored on memory cards – this was only necessary where a business’s Wi-Fi was insufficient for streaming), while interviews with owners did not directly reference any specific instances of unhygienic behaviour. This prevented any attempts to single out staff.
Plus, all staff all got a gift certificate or financial incentive for their participation!
Conclusion
And there you have it, folks – a behind-the-scenes look at Kitchen Life 2 and the ethical tightrope the team walked. Every step of the way was carefully orchestrated to shed light on kitchen behaviour while respecting the privacy and dignity of all involved. So, next time you're whipping up a culinary masterpiece in your kitchen, just remember: you never know who might be watching, but rest assured, they've got your ethics covered.
Want to know more?
If you’d like to find out more about the project, or its ethical considerations, please do take a look at the Technical Report and the Behavioural Reports, or get in touch with either [email protected] or [email protected].
Author Bios
Thomas Mills is a Senior Social Science Research Officer at the Food Standards Agency. He joined the FSA’s Social Science team in October 2023, having previously worked for two years at the Ministry of Defence on a range of social research projects, including Wraparound Childcare and the Haythornthwaite Review.
Helen Heard is also a Senior Social Science Research Officer at the Food Standards Agency. She joined the FSA’s Social Science team in February 2020, and has recently led on the FSA’s Monthly Consumer Insights Tracker Survey as well as the Kitchen Life 2 behavioural research project.
Acknowledgements:
Kitchen Life 2 was delivered by lead contractor Basis Social, with academic support from Leeds University Business School and a variety of sub-contractors including Lifestream (responsible for technology, filming, coding and data delivery), analysis from Analytical People and recruitment by Fieldsauce.