Collecting data: what does it tell us about wellbeing?

What should research ask after an experience?

After an audio exercise, an interactive story or an immersive scene, someone may be willing to describe how it felt. Researchers may also want questionnaires, behavioral records, or physiological or biological measures. More records do not automatically mean a clearer understanding of the experience. First, each measure needs a question it can answer.

The BeWell report published on 9 October 2026 offers a concrete example. It describes collecting and processing several kinds of data in a fully remote trial and examines whether that approach is feasible. The authors explicitly state that this paper does not report Healthy Minds Program (HMP) app efficacy or the trial’s primary clinical outcomes. [1]

Read the denominator and definition beside the 84% retention rate

The paper counted 971 of 1,156 participants, about 84%, as retained at week 17 because they completed at least one measure then. That does not mean 84% completed every procedure or experienced improved mood. Table 1 offers a different figure: 678 of 796 submitted week-17 video logs, 85.18%, met its high-quality criteria. That denominator is submitted video logs, rather than all 1,156 participants. [1]

The figures answer different questions: whether someone still provided some research data, and whether a particular kind of collected data met analysis standards. For new digital wellbeing studies, we propose reporting invitations, submissions, analyzable records and outcome-analysis samples separately, with time points and reasons for missingness. A single attractive percentage should not stand in for that process.

Analyzable data still need an account of what they measure

The study screened video using lighting, face-tracking and speech-quality criteria, with other checks for memory tasks and biological samples. It also states that Teddy, the video-collection tool, has not been independently validated. [1] Passing a quality check makes a record suitable for further analysis under the specified procedure. It does not by itself establish that an expression, vocal feature or biological marker accurately represents a person’s change in wellbeing.

We suggest keeping three questions distinct: can data be obtained reliably; does a measure support its intended interpretation; and does an experience produce meaningful change relative to a comparison condition? The answers may support one another, but each needs its own evidence. A clear video still requires evidence about what its measurements mean. A changing questionnaire score also needs its comparison and time frame.

For interactive narrative, sound or VR, a study could begin by specifying the intended change and choosing measures that fit it. A question about feeling understood needs an explicit way to ask about that experience; a question about everyday action needs separate observation. These are research-design proposals from this site, not findings validated in those experiences.

Those who stayed had already passed through an enrollment process

Before randomization, participants had to complete questionnaires and a remote clinical interview, submit at least one biological sample, and attempt behavioral tasks. The paper records 52,052 screener starts and 1,156 eventual enrollees. Greater symptoms at screening were associated with lower odds of completing baseline procedures. Among enrolled participants, baseline symptom severity did not significantly predict later completion across data types or questionnaire quality; insufficient variability precluded the corresponding quality analyses for behavioral and biological data. Those findings do not establish equally accessible participation at every level of symptoms. [1]

High retention therefore deserves to be considered alongside the requirements for entering the study. The sample did not match US population proportions across racial and ethnic groups. English proficiency, smartphone access and US residence were also required. Feasibility under these conditions cannot simply be extended to every country, language or living situation. [1]

We propose including reasons for nonenrollment and noncompletion in an evaluation, while allowing people to withhold private explanations. A new global project could work with local participants to examine language, devices, task burden and support before reusing a procedure. This is a future design proposal, not cross-country validation already conducted by this site.

Remote research has real support conditions

The study used staff, reminders, instructions, sample logistics and participant compensation; it did not simply provide an app and wait for data. The authors describe difficulties switching from computer questionnaires to smartphone video tasks and the possible burden of a lengthy assessment battery. [1] Procedures feasible with these resources are not automatically equally feasible in everyday use without staff support.

A digital wellbeing laboratory can treat measurement itself as a design task: ask a clear question with a manageable set of activities, explain collection purposes, and provide suitable options for withdrawal and feedback. Each requested data type should have a reason beyond technical availability. These are this site’s research proposals, rather than improvements tested by the paper.

BeWell contributes an account of organizing remote multimodal assessment, what data could be obtained, and where its limits lie. Questions about effects and mechanisms still require corresponding outcome analyses. Describing feasibility accurately gives subsequent digital wellbeing evidence a sound starting point.

This article contains AI-generated content. AI researched, synthesized, wrote and edited both languages using the cited public original source. This site did not conduct the trial, hands-on product testing or original interviews. The operating team develops Six Realms and provides Heart to Heart and Sleep, creating relevant commercial interests. The research discussed here did not assess these products and does not establish their effects. This article does not provide a personal diagnosis or treatment plan.

Luoboli · Investigating evidence and experience