Smartwatches offer little insight into stress levels, researchers find

The Guardian

6 days ago

They are supposed to monitor you throughout the working day and help make sure that life is not getting on top of you.
But a new study has concluded that smartwatches can't accurately measure your stress levels – and may think you're overworked when really you're just excited.
Researchers found almost no relationship between the stress levels reported by the smartwatch and the levels that participants said they experienced. However, recorded fatigue levels had a very slight association with the smartwatch data, while sleep had a stronger correlation.
Eiko Fried, an author of the study, said the correlation between the smartwatch and self-reported stress scores was 'basically zero'.
He added: 'This is no surprise to us given that the watch measures heart rate and heart rate doesn't have that much to do with the emotion you're experiencing – it also goes up for sexual arousal or joyful experiences.'
He noted that his Garmin had previously told him he was stressed when he was working out in the gym and when excitedly talking to a friend he hadn't seen for a while at a wedding.
'The findings raise important questions about what wearable data can or can't tell us about mental states,' said Fried. 'Be careful and don't live by your smartwatch – these are consumer devices, not medical devices.'
Fried said although there is a lot academic work looking for physiological signals that can act as proxies for emotional states, most aren't precise enough. This is because there is an overlap between positive and negative feelings – for example, hair standing on end can signal anxiety as well as excitement.
Fried, an associate professor in the department of clinical psychology at Leiden University in the Netherlands, and his team tracked stress, fatigue and sleep for three months on 800 young adults wearing Garmin vivosmart 4 watches. They asked them to report four times a day on how stressed, fatigued or sleepy users were feeling before cross-referencing the data.
And the results, published in the Journal of Psychopathology and Clinical Science, found that none of the participants saw the stress scores on their watches meet the baseline for significant change when they recorded feeling stressed. And for a quarter of participants, their smartwatch told them they were stressed or unstressed when they self-reported feeling the opposite
The relationship with physical fatigue, described by Garmin as 'body battery' was 'quite a bit stronger than for stress but overall quite weak' said Fried. Garmin does not disclose the calculations it uses to work out the body battery score, though he suspected it was a combination of a pulse measurement and activity levels.
The relationship with sleep was stronger again, though Fried noted it measures sleep duration and tells us little about how well rested someone is, other than the fact there tends to be a relationship between how long you sleep and how well rested you feel.
There was a significant association between the Garmin and self-reported data for two-thirds of the sample for sleep. The researchers noted that in nearly all cases, if participants went from one day of self-reporting bad sleep quality, to another day with a good score, they could predict an increase in sleep duration on the Garmin of around two hours. 'This is a really noticeable effect,' they said.
The research is intended to feed into an early warning system for depression, in which wearable tech users receive data that will help them receive preventive treatments before an episode begins.
So far, there are promising signs that lower activity levels could be a predictor, though Fried has been unable to identify whether this is because of exercise's protective effect against depression or because people feel less energetic as their mental state deteriorates.
'Wearable data can offer valuable insights into people's emotions and experiences, but it's crucial to understand its potential and limitations,' said Margarita Panayiotou, a researcher at the University of Manchester, after reading the study.
'This research helps clarify what such data can reliably reveal and makes an important contribution to ongoing discussions about the role of technology in understanding wellbeing. It's important to remember that wearable data does not necessarily represent objective truth and should be interpreted alongside broader context, including individuals' perceptions and lived experiences.'
Garmin has been approached for comment.