Asked often enough, people's reports of their own day can drift

A lot of everyday health research relies on asking people, again and again, what they have just been doing. The method is called ecological momentary assessment. It usually assumes that answering the same questions many times doesn't change the answers. A study published on October 29 in the Journal of Medical Internet Research tested that assumption.

How it was checked

The authors, Toledo, Schneider, Smyth, Junghaenel and Stone, used data from 315 adults in the United States, recruited from a national research panel between January and July 2023.

For seven days, each person wore a small sensor on the thigh that continuously recorded how much they moved and how much they sat. At the same time, their smartphone asked them five times a day about their activity and sitting.

There were two versions of the questions. In one, people reported on just the past five minutes. In the other, they recalled the past two hours.

The sensor gave a steady reference. If self-reports changed over the week in a way the sensor didn't, the change was likely in the reporting, not in the behaviour.

What drifted

With the two-hour recall, self-reports drifted away from the sensor over the seven days. Reported sitting rose relative to what the sensor recorded, and reported activity fell, with the gap following a curve rather than a straight line.

With the five-minute version, the authors found no sign of that drift. Self-reports and the sensor moved together across the week.

The difference between the two versions was itself statistically significant, for both sitting and activity.

Why it matters

The authors describe the drift as modest. But it challenges a common assumption, and it matters most when a study looks at change within a person over time. If the reporting itself shifts across a week, some of what looks like a real change could come from the act of being asked.

The fix they point to is design. Short, near-moment questions seemed to hold steady. Longer recall windows did not.

What it doesn't cover

This was a secondary analysis of one panel of adults over seven days. It looked at activity and sitting, not other things people are often asked about, such as mood or food. And a thigh sensor has its own limits as a reference, even if it doesn't change its behaviour from day to day.

Where Soma fits

This study gets at something we care about. A measurement can change because the thing measured changed, or because the way it is measured shifted. Telling the two apart needs a reference that stays the same.

Soma is a general wellness product, not a medical device. It doesn't diagnose anything, and it doesn't track activity or sitting. Soma Axon, the pod we are still building, uses the same light, cameras and stance every week, and shows each value with its margin of error, so a change over time is easier to trust. We wrote more about that in The same instrument, every week.

Sources, as of October 30, 2026: Toledo, Schneider, Smyth, Junghaenel and Stone, Journal of Medical Internet Research. This post describes research and isn't medical advice.