TeamChat

What you see

Above the board is the question "Which reply handles this best?". Under it are two panels and a status line.

• The thread (top), headed by a channel name such as #release-plan or #client-orbit, and by where you are, such as "Thread 1 of 4". Each message appears as a bubble, with the sender's name above it whenever the sender changes. On a wider screen, your last reply in the same thread stays on screen above the new messages, on the right. On a narrow screen, such as most phones, only the new messages are shown.
• Your replies (bottom): three or four buttons, each showing a reply and its number. Nothing else is on them: no tone, no hint about credit.
• The status line, such as "Message 2 of 8".

You choose a reply or press its number key, and the reply is sent at once and cannot be taken back. The next two messages then arrive, either in the same thread or, when that thread's part of the set is over, in a new channel. There is no clock on any decision and the session can be paused. When every reply is sent, the status line says so, for example "All 8 replies sent", and you select Check answer to see how the set went.

There are ten threads, and each one follows a single problem through four exchanges of two messages:

Each set uses different threads, picked at random, and takes a run of exchanges in order from each one, so the story moves on with every reply you send. Every reply is written for its own message, and the replies are shown in a random order.

What changes as levels rise

Higher levels mean more threads, more replies inside each thread and four replies to choose from instead of three. At levels 1 and 2, one of the two no-credit replies is left out at random, so the three you see are one full-credit, one half-credit and one no-credit reply. The difficulty is endurance and consistency: making the thirtieth judgement as carefully as the first.

What it measures

TeamChat is built on the situational judgement test, which Motowidlo, Dunnette and Carter described as a low-fidelity simulation: a realistic work situation, several possible responses, and a choice. Meta-analyses find SJTs predict job performance (ρ = .34 in the first large meta-analysis; about .26 in a 2022 revision) and add prediction beyond both cognitive ability and personality.

What an SJT captures depends on the item. Scores correlate substantially with cognitive ability (ρ = .46), and knowledge-style instructions ("what is most effective?") lean further that way, while "what would you do?" instructions lean towards personality. Lievens and Motowidlo argue that much of what SJTs measure is general domain knowledge: knowing which kinds of behaviour tend to work at work, such as acknowledging someone's concern, dealing with problems early, and escalating with a proposal rather than a complaint. SJTs aimed at teamwork and leadership, the territory of TeamChat, have shown relatively high validities.

The guide for this task sums up the instruction well: the longest or most agreeable reply is often not the best one; judge what moves the problem forward and keeps trust.

How it is scored here

Your result for a set is the credit you earned as a share of the most you could have earned. A set worth 98% or more counts as right, a set that earned no credit at all counts as wrong, and anything in between counts as partly right. The level follows that: it goes up one after a right set, down one after a wrong set, and stays where it is after a partly right one.

In the screen above, reply 3, "Only if the login fix passes its tests. I’ll know by four and tell you then.", earns full credit. Reply 1, "Probably, yes. I’ll let you know if anything changes.", earns half credit, and replies 2 and 4 earn nothing. The key's reason for reply 3: "She has to decide today. Saying what the date depends on, and when you will know, gives her an honest basis for that decision."

Here is what the report shows:

• Agreement with the answer key: your credit as a share of the most you could have earned. It is shown for completeness only and labelled that way.
• Reply tone you picked most: every reply in the key also carries a tone tag (direct, diplomatic, analytical or supportive). You do not see the tags while you choose. This figure says which tag your replies carried most often; read it as a record, not a diagnosis.
• Median time per scenario: how long you typically took to decide.

In the review of a set, each message shows the reply you sent and its credit, the full-credit reply, and the key's reason for it. Sessions from before the written key gave credit to replies at random, so treat any figures from them as noise.

So what is TeamChat for, while the key is reviewed? It is a realistic rehearsal space. The situations are exactly the kind SJTs and HR interviews use, each thread follows one problem as it develops, and the rule that a sent reply cannot be recalled trains you to decide before you commit. Use it to practise the method below, then check your reasoning against the key's reasons in the review.

Where you will meet it

SJTs are among the most widely used selection tools. In a global survey of 1,406 HR professionals, 43% said their organisations used them. You may meet them as a stand-alone section, as items inside a "behavioural" block, or as video clips with response options.

In India, the clearest official example is defence officer selection: the Services Selection Board's Situation Reaction Test gives candidates 60 situations to respond to in 30 minutes. In private-sector campus drives, behavioural sections are more often personality questionnaires than SJTs, so check your invite. Interviews use the same logic out loud: "What would you do if a teammate missed a deadline that affected your work?" is an SJT item with no options on the page.

The method

1. Read the whole thread

   Who is messaging: a teammate, a client, your manager, a new hire? What did they actually say, and what have they already admitted or asked? A teammate who has already apologised needs something different from one who has not noticed the problem.

2. Name the two goals

   Almost every workplace situation has an outcome goal (the delivery, the release, the fix) and a relationship goal (trust with this person). The best reply serves both. A reply that wins one and wrecks the other is rarely keyed best.

3. Strike out the harmful replies

   Remove anything that blames publicly, hides a problem that others need to know about, avoids the issue, or pushes a decision upwards without reason. These are the replies SJT keys usually rate lowest.

4. Prefer acknowledge plus a next step

   Among what is left, the strongest reply usually does two things: shows the person you have understood them, and moves the problem forward with something concrete. If you lack key information, a clarifying question is the next step.

5. Check whose call it is

   Escalate when the decision is genuinely above your role or affects people who are not in the conversation, and say what you propose. Do not escalate what you can and should handle.

6. Say your reason in one sentence, then send

   If you cannot justify the choice in a sentence, you have not finished judging. Once you can, commit: the reply cannot be taken back.

The worked examples below are real messages and replies from the task, in one of the orders you might see them. The credit for each reply and the reason for the full-credit reply are quoted from the written key. The notes on the other replies are our own reading.

Worked example 1: the late handover

Show the reasoning

Full credit: 2, "A little. I need the API part by noon. Could you send that first?" The key's reason: "Priya has owned the delay and asked if it blocks you. Saying exactly what you need, and by when, keeps the work moving without blame."

Half credit: 4, "No worries, these things happen. Send it whenever you can." Kind, and good for the relationship, but it does not answer her question. Whatever you are blocked on stays blocked.

No credit: 1, "Let’s go through why it slipped first, so this doesn’t happen to us again." She has already owned it. A review in the middle of the problem adds blame and fixes nothing today.

No credit: 3, "Don’t worry. I’ll tell the team my part may slip too." It lets the delay spread to your own work without trying to limit it.

Worked example 2: the unhappy client

Show the reasoning

Full credit: 3, "Sorry for the worry. What matters most on Friday? I’ll send a plan today." The key's reason: "An unhappy client needs to feel heard and to know what happens next. Asking what Friday is for lets you plan something you can deliver." The next exchange in this thread shows why the question matters: Friday is a demo to investors, and only the sign-up flow needs to work.

Half credit: 1, "Let me check the status with the team and come back to you." Honest, and it promises a next step, but it does not acknowledge her worry, gives no time, and does not find out what she needs.

No credit: 2, "Please don’t worry at all, it will definitely be with you on Friday as promised." Agreeable, but it promises a date nobody has checked. If Friday slips again, trust is gone.

No credit: 4, "Delays happen in software. It will be ready when it’s ready." It dismisses a client who is already frustrated.

Worked example 3: the heated argument

Show the reasoning

Full credit: 4, "You both care about this. Can we take it to a quick call, the three of us?" The key's reason: "The argument has turned personal in a public channel. Moving it to a short call lowers the heat and lets both be heard."

Half credit: 2, "Marco, what went wrong last time? That might help us decide." It looks for the real objection, which is useful, but asking in the group chat may keep the argument going in front of everyone.

No credit: 1, "Lena, that’s not a helpful thing to say in here." Calling out one colleague in public takes a side and raises the heat.

No credit: 3, "I’ll stay out of this one. I’m sure you’ll both sort it out between yourselves." It leaves a personal argument running in the team channel.

Tips and traps

Levelling up

As the number of threads and replies grows, the challenge is consistency. Keep your method identical from the first reply to the last, and use the pause if you notice yourself choosing by feel. Each thread follows one problem as it develops, and on a wider screen your last reply in the thread stays above the new messages, so check that your next reply fits with what you already said. The effort pays off beyond this task: SJT performance responds to preparation, with coaching raising scores by about half a standard deviation in one high-stakes admissions setting, because what SJTs reward is learnable knowledge about what works.

1. Week 1
   • Read the method, then complete one TeamChat session slowly
   • For every reply, write the outcome goal and the relationship goal in a few words
   • Note which replies you struck out first and why
2. Week 2
   • Two sessions; before sending, say your one-sentence reason aloud
   • For three threads, write the reply you would send in your own words, then compare it with the full-credit reply in the review
   • Compare your choices with the worked examples above
3. Week 3
   • One longer session without using the pause
   • Turn two TeamChat situations into STAR interview answers from your own experience
   • Ask a friend to challenge your reasoning on three replies

Check yourself

5 practice questions

1. Why does TeamChat not count towards your scores yet?
2. What happens when you choose a reply?
3. In #release-plan, Ravi writes “The login bug still isn’t fixed. I’m on it.” Then your manager Dana asks “Where are we on the release? I need an update for three.” Which reply earns full credit in the key?

and 2 more

Create a free account to answerSign in

Sources

1. Motowidlo, S. J., Dunnette, M. D., & Carter, G. W. (1990). An alternative selection procedure: The low-fidelity simulation. Journal of Applied Psychology, 75(6), 640–647. doi.org/10.1037/0021-9010.75.6.640
2. McDaniel, M. A., Morgeson, F. P., Finnegan, E. B., Campion, M. A., & Braverman, E. P. (2001). Use of situational judgment tests to predict job performance: A clarification of the literature. Journal of Applied Psychology, 86(4), 730–740. doi.org/10.1037/0021-9010.86.4.730
3. McDaniel, M. A., Hartman, N. S., Whetzel, D. L., & Grubb, W. L. (2007). Situational judgment tests, response instructions, and validity: A meta-analysis. Personnel Psychology, 60(1), 63–91. doi.org/10.1111/j.1744-6570.2007.00065.x
4. Sackett, P. R., Zhang, C., Berry, C. M., & Lievens, F. (2022). Revisiting meta-analytic estimates of validity in personnel selection. Journal of Applied Psychology, 107(11), 2040–2068. doi.org/10.1037/apl0000994
5. Lievens, F., & Motowidlo, S. J. (2016). Situational judgment tests: From measures of situational judgment to measures of general domain knowledge. Industrial and Organizational Psychology, 9(1), 3–22. doi.org/10.1017/iop.2015.71
6. Christian, M. S., Edwards, B. D., & Bradley, J. C. (2010). Situational judgment tests: Constructs assessed and a meta-analysis of their criterion-related validities. Personnel Psychology, 63(1), 83–117. doi.org/10.1111/j.1744-6570.2009.01163.x
7. Berry, C. M., Lievens, F., Zhang, C., & Sackett, P. R. (2024). Insights from an updated personnel selection meta-analytic matrix. Journal of Applied Psychology. doi.org/10.1037/apl0001203
8. Services Selection Board. Wikipedia, psychological tests. en.wikipedia.org
9. Lievens, F., Buyse, T., Sackett, P. R., & Connelly, B. S. (2012). The effects of coaching on situational judgment tests in high-stakes selection. International Journal of Selection and Assessment, 20(3), 272–282. doi.org/10.1111/j.1468-2389.2012.00599.x

Put it into practice: TeamChatSign in to try it