Emotions as a food lie detector

| Michaela Nesvarova

Customers often decide which product to buy based on ‘feelings’. Daisuke Kaneko, PhD researcher at the UT, studied how emotions could be used to predict food choices – and how this differs across cultures. The doctoral candidate in the Human Media Interaction group is defending his thesis tomorrow.

What was the main goal of the research? Why is it important to measure food-evoked emotions?

Kaneko: ‘Research shows that 60-70% of new food products fail within one year on the market. When deciding which product to put on the market, companies use self-report measures, in which people indicate how much they like certain food after tasting it. However, it has been suggested that emotions are one of the key factors when selecting food products in a shop. Consumers are influenced by emotions evoked by tasting food. Based on the request of the food company I work for, I therefore wanted to explore different methods to measure people’s emotions evoked by food experience.’

Can you describe these methods?

‘In my research, I focused on measuring physiological responses, such as heart rate, pupil dilation, skin conductance and brain activity. We cannot control these reactions, and so, in a way, they are expected to be more honest than filling in a questionnaire. If these physiological measurements are sensitive enough to evaluate food-evoked emotions, we can better understand consumers’ food choices and preferences.’  

What are you able to assess based on these measurements?

‘There are many emotional responses that we can detect, such as feelings of happiness, nostalgia, calmness or disgust. For example, we did an experiment asking participants to drink regular drinks – like orange juice or coke. We could also see they subconsciously regulate how much they drink. If they expect to like the drink, they take a bigger sip. Combining physiological measurements with surveys can therefore help to make better choices regarding new products.’

In your doctoral research you also focused on cross-cultural differences. Why is that?

‘Indeed. We did experiments with Dutch people and with Thai people. We travelled to Thailand and recruited locals who have only been exposed to the Thai culture and in the Netherlands we only included Dutch people who’d never been out of Europe. We ran experiments with both groups and compared their responses. Thai people are less likely to give very low or very high ratings. Even if they find the food disgusting, they will not give it the lowest rating, they are more likely to stick to the middle-range rating. Dutch people, on the other hand, have a more extreme response style. If they don’t like it, they will fill that in. Because of these cultural differences, there is a response bias that can mislead the product evaluation and development. However, in term of heart rate and sip size, we found their ‘genuine’ response on food image stimuli and didn’t find any response bias between Dutch and Thai participants. We found out that heart rate and sip size worked as good indicators for measuring the participants’ emotions even under cross cultural context. It is a more honest and less biased response compared to self-report ratings.’

What are the main findings of your research?

‘I’d say there is a good potential for using physiological measurements to predict consumers’ future buying behaviors. However, there are also a lot of limitations to these methods, because you’d always have to physically travel to the specific country and run experiments there. That is not the standard practice.’

Will the findings be used in practice?

‘That is up to my company, but I assume they will be taken on board. For the future, I’m considering to study if we could use digital instruments to remotely measure heart rate and other physiological and behavioral responses. These days, there are some new instruments like wearable devices developed which can remotely measure physiological response to food stimuli. We need to find ways to do the measurements more easily.’