Skip to main content
Lumium
Research project: DaTeamVR

Can teams be measured objectively?

In the research project DaTeamVR, we work with the Technical University of Applied Sciences Augsburg to find out whether the data of a VR mission shows how a team works together.

Our first example is team resilience, meaning the question of how a team deals with setbacks.

Our goal

A fingerprint for every team.

We measure how teams work together in the VR experience. We analyze the data with methods from data science and with psychological measurement instruments. The aim is a fingerprint for every team that we can compare with those of other teams.

    • Laboratory situation
    • Authentic behavior

    A team in VR

    • Movement and interaction
    • Voice and conversation
    • Heartbeat
    • Questionnaires
    • Eye movementsin testing

    Data from VR

  1. Behavioral fingerprint

What this is about

What holds a team together is invisible.

  • TrustSafety

    It lies between people.

    Whether a team can rely on each other, whether someone openly admits a mistake and whether the team finds its way back together after a setback all show in how it works together. Trust, psychological safety and team resilience are hard to see from the outside.

  • Worktenseafter

    The questions come afterwards.

    Today, such states are measured with questionnaires, usually once and after the fact. They do not show when things got tense during the work or how the team found its way back together afterwards.

  • TrustSafetyCohesion

    A questionnaire shows the self-image.

    It records how a team rates itself. We need that as a benchmark, which is why questionnaires are part of the data in DaTeamVR. How a team actually works together under pressure, it shows only indirectly.

  • Team ATeam BTeam C

    Comparing requires the same benchmark.

    In everyday work, every team has different tasks, different clients and different pressure, so it is hard to put one team fairly next to another. A comparison only becomes objective when all teams get the same task under the same conditions and their collaboration is measured in the same way.

Why VR

Why VR makes this possible. In VR, every team gets the same task, and the data is generated while the team works.

  • A laboratory situation.

    Every team gets the same mission under the same conditions. Only this makes it possible to compare teams with each other at all.

  • Real team behavior.

    After a few minutes, VR becomes a real place for the participants. Under time pressure, a team shows its patterns and not its seminar behavior.

  • Same building blocks, its own sequence.

    Teams do not behave at random. What they say in the mission can be assigned to the same team processes, such as coordination or mutual support, automatically and with the help of a dictionary. The sequence in which a team uses them is different for every team.

  • Data, data, data.

    The VR headset itself records movement and conversation during the mission. In addition, there is a chest strap and short questionnaires, and nobody has to interrupt the mission for them.

What we measure

Five types of data, several metrics from each.

We take an exploratory approach and look for the data streams in which teams differ clearly, for example to recognize resilient teams.

  • Movement and interaction

    What the team does in VR, where it moves and who takes on how much of the task.

    Examples

    • Paths through the room
    • Who takes on how much, and how quickly that changes
    • Exploring the room or taking familiar paths
  • Voice and conversation

    What the team says and how it sounds. The recording is turned into a word-for-word transcript.

    Examples

    • Speaking shares
    • Team process per utterance: planning, acting, maintaining relationships
    • Phases in the conversation
    • Emotional expression in the voice
    • Turn-taking and pauses
  • Heartbeat

    How the heart beats, measured with a chest strap.

    Examples

    • Heart rate
    • Heart rate variability
    • Synchrony within the team
  • Questionnaires

    What the team reports itself, before and after each run. We check all other data against the questionnaires.

    Examples

    • Team resilience
    • Psychological safety
    • Cohesion
    • Strain
    • The critical moments of a run
  • Eye movements

    Some VR headsets record where someone is looking. We are currently testing this type of data.

Sources

  • Braun, Hogh and Kubowitsch (2026), Small Group Research, peer-reviewed. 15 teams
  • Huang, Zhang and Pol (2026), arXiv, preprint, not yet peer-reviewed. 15 teams
  • Hogh, Watermann, Braun and Kubowitsch (2026), Ein deutschsprachiges Wörterbuch zur Erfassung von Teamprozessen (a German-language dictionary for capturing team processes), conference poster
  • Hogh, Watermann, Braun, Brandt and Kubowitsch (2026), Kann man eine Teaminteraktion „LIWCen“? (Can a team interaction be analyzed with LIWC?), conference poster. 22 work teams
Multimodality

Only the combination of data streams makes the data reliable.

Each type of data shows only part of the picture on its own. Our assumption is that a finding becomes reliable when several data streams show the same thing at the same moment, for example voice, conversation and heartbeat. We test whether this holds by looking at the connections between them.

Fingerprint
Movement
Voice
Heart
Questionnaire
Eyes

Sample questions we ask of the data

  • Does a peak in the emotional expression of the voice coincide with a change of phase in the conversation?
  • Do heartbeat and questionnaire say the same thing about the team processes?
  • Does the data find the same critical moments that the team names in the questionnaire?
The benefit

Our goal:Objective measurements.Targeted measures.

In the end, there should be measurements that show how resilient a team is and how it works together, and that can be compared with other teams. From these, a team should be able to derive measures that improve it in exactly these areas.

The vision

A dashboard for collaboration.

The idea is a data-based measurement tool for collaboration. It shows a team where it stands and helps it draw the right conclusions.

The data is a mirror for the team, not a judgment of individual people. Guidance from our trainers remains part of the method.

This is a product vision!

A dashboard like this does not exist yet. The values in it are examples, not measurements, and many of the metrics behind them have been calculated but only a few have been validated.

DaTeamVR · Team A

Run 2 · 15 minutes

Sample data, not measurements

Team mental workload

Questionnaire after the run

5-7out of 10

  • Task6-8
  • Teamwork4-6
  • Balance5-7

Perceived strain

Questionnaire before the first mission

3-5out of 10

How straining the team finds its everyday collaboration, independent of the mission.

Team processes in conversation

Share of utterances, assigned automatically

Coordination24-31%Agreement9-14%Monitoring15-21%Support11-16%Relationship4-7%Planning3-6%

This teamOther teams

  • Similarity within the team0.5-0.6

    How similarly everyone uses the categories. A low value means that the roles are distributed.

  • Topic changes0.7-0.8

    How often the conversation jumps from one category to the next.

Course of the mission

Utterances per half minute and peaks in the emotional expression of the voice

  • Coordination
  • Agreement
  • Monitoring
  • Support
  • Peak in emotional expression
  • In the questionnaire: most stressful

View from VR

Secondary monitor, for the debrief

View from VR: two avatars in the room of the missionView from VR: several avatars side by sideView from VR: close-up of the missionView from VR: avatars in a group
01:30 / 15:00
01:30First agreementThe team gets organized and hands out the first tasks.

Key scenes

Take part

Two ways to join DaTeamVR.

Would you like to take part with your team? Then we will send you the details by email.

We only use your address to send you the information.

The team

Who works on DaTeamVR.

DaTeamVR is a joint project of Lumium and the Technical University of Applied Sciences Augsburg. Business psychology defines which team variables we measure and validates them, data science at the Technology Transfer Center looks for patterns in the data, and Lumium contributes the VR experience and the workshops. The project is funded by the Bavarian Ministry of Economic Affairs under the BayVFP program and runs until the end of 2027.

Project leads

The three project leads of DaTeamVR: Simone Kubowitsch, Lukas Karwan and Jianing Zhang

Lumium

  • Portrait of Michael Paff
    Michael PaffHead Programmer
  • Portrait of Katharina Meier
    Katharina MeierProject Management
  • Portrait of Stefanie Böttcher
    Stefanie BöttcherHead of Didactics · Process trainer
  • Portrait of Annika Härtel
    Annika HärtelProcess trainer
  • Portrait of Simon Scharnweber
    Simon ScharnweberVR tech trainer
  • Portrait of Emmelie Proboscht
    Emmelie ProboschtVR tech trainer

Business Psychology

  • Portrait of Lara Watermann
    Lara WatermannResearch, Business Psychology
  • Portrait of Natalie Hogh
    Natalie HoghResearch, Business Psychology
  • Portrait of Joshua Braun
    Joshua BraunResearch, Business Psychology (former)

Data Science

  • Portrait of Pooja Pol
    Pooja PolResearch, Data Science
  • Portrait of Qing Huang
    Qing HuangResearch, Data Science
Data protection

What happens to the data.

When data is anonymized

The research data is collected and analyzed anonymously. Recording takes place only in DaTeamVR workshops and only with consent; nothing is stored in a regular Lumium workshop. Anyone who does not want to fill in a questionnaire still experiences the full workshop.

Guide for participants

Who can link what to whom

Nobody can link answers or measurements to an individual person. The data is processed exclusively on servers of the Technical University of Applied Sciences Augsburg, without disclosure to third parties. The university reviewed and approved the concept for this in its own procedure.

What your company receives

Your team gets the workshop and then discusses its collaboration together with our trainers. Your company does not find out who said what or what was measured for whom.

Frequently asked questions

Questions about DaTeamVR.

Are individual people assessed?

No. We analyze teams, not individual people. The data is collected anonymously, and your company does not find out who said what or what was measured for whom either.

What happens to the data, and who sees it?

Recording takes place only in DaTeamVR workshops and only with consent; nothing is stored in a regular Lumium workshop. The data is anonymized and processed exclusively on servers of the Technical University of Applied Sciences Augsburg, without disclosure to third parties. The university reviewed and approved the concept for this in its own procedure.

Guide for participants
Is this already scientifically proven?

Not yet. DaTeamVR is basic research. First results have been published, for example on the synchrony of heart data in the journal Small Group Research (peer-reviewed) and on the phases in conversation as a preprint (not yet peer-reviewed). They show correlations in small samples and are not proof that a workshop is effective.

Where we stand
What does my team get out of it?

Your team gets a VR workshop at a reduced price, or free of charge as a non-profit organization. It experiences its collaboration under time pressure and then discusses it together with our trainers. And it helps make it possible for teams to see their collaboration in data in the future, not only in questionnaires.

Who can take part, and what does it cost?

Teams from companies with 5-9 people who work together in everyday life pay €1,299 net instead of the regular price from €2,899. Teams from non-profit organizations with 5-8 people take part free of charge. The details are on the two workshop pages.

To the workshops
How much effort does the measurement take?

The VR headset itself records movement and conversation during the mission. In addition, there is a chest strap that is barely noticeable after a short time, and short questionnaires before and after the runs.

Do you need VR experience, and does it make you feel sick?

Nobody needs VR experience. There is an onboarding before the mission, and what counts is collaboration, not technology. Modern headsets process head movements quickly enough, and the mission avoids scenes such as virtual roller coasters. In practice, we see no more complaints than with conventional physical exercises.

Where we stand

What already exists, and what is still a goal.

DaTeamVR is basic research and runs until the end of 2027. So that you can place every statement on this page, here is an overview of what exists today, what we have observed so far and what we still want to achieve.

Available today

What exists and is used in the workshops.

  • A VR mission in which every team gets the same task under the same conditions.

  • Recording of movement, actions, conversation, heartbeat and questionnaires, only in DaTeamVR workshops and only with consent.

  • A German-language dictionary that automatically assigns every utterance in the conversation to a team process, such as coordination or mutual support.

    • Hogh, Watermann, Braun and Kubowitsch (2026), Ein deutschsprachiges Wörterbuch zur Erfassung von Teamprozessen (a German-language dictionary for capturing team processes), conference poster
  • First data sets from 15 student teams and from 22 real work teams.

Observed so far

Correlations in small samples, analyzed exploratively.

  • A team's heart data is related to trust and psychological safety. Depending on the method of calculation, however, the same heartbeat points in opposite directions.

  • Phases that match the actions in the mission can be identified from the conversation. In the second run, the teams get into the task sooner.

  • In 22 work teams, an unequal distribution of speaking shares was related to lower performance.

    • Hogh, Watermann, Braun, Brandt and Kubowitsch (2026), Kann man eine Teaminteraktion „LIWCen“? (Can a team interaction be analyzed with LIWC?), conference poster. 22 work teams

Research goal

What we want to achieve and cannot do yet.

  • A behavioral fingerprint for each team that can be compared with those of other teams.

  • Estimating team states such as psychological safety and team resilience continuously from the data, instead of only asking afterwards.

  • Benchmarks from many teams that show a team where it stands.

  • A dashboard that gives a team feedback on its collaboration.

The observations are correlations in small samples, not causes. They are not proof that a workshop is effective.

News from the research

DaTeamVR on the blog. Findings, insights, dates.

All articles
Contact

Questions about the project?Or interested in a workshop?

Then write to us. We will agree on date, room and team size with you and answer your questions about the data.

Lukas Karwan
Lukas KarwanManaging Director

Your details are transmitted by email to our team and used only to handle your inquiry. Details in our privacy policy. Or write directly to dateam@lumium.de.

Technical University of Applied Sciences Augsburg

DaTeamVR is a joint project of Lumium and the Technical University of Applied Sciences Augsburg, funded by the Bavarian Ministry of Economic Affairs as part of the Bavarian Collaborative Research Program (BayVFP), funding line Digitalization.

To the BayVFP funding program