Team Psychological Safety and Speaking-Up Pulse for Team Members and Their ManagersOne read of a team is not the team. This report says so first, prints what a team number would take, and gives you one thing to do on Monday.
Thirty-three short items about what has actually happened on your team in the last three months: what followed when somebody admitted a mistake, whether a question that showed a gap was asked and how it was received, whether a different view survived being said to the most senior person in the room, and whether a risk was raised before it landed. Every item is about the team, not about you. No right answers, nothing to know, nothing hypothetical. The report is built around one boxed if-then sentence, with a matching one addressed to whoever leads the team. Free to sit; the report is the product.
A team pulse that refuses to pretend one person is the team
The Team Psychological Safety and Speaking-Up Pulse is a twelve-minute situation report for anybody on a team and for whoever leads it. It records what has actually happened on that team in the last three months when somebody admitted a mistake, asked when unsure, disagreed upward or raised a risk early, and returns one specific change to make on Monday.
At one end of this market are free, unscored team surveys that hand back a mean and no uncertainty. At the other are licensed team scans sold per team, priced on request, whose output is an aggregate nobody on the team can read for themselves. Nothing in between publishes, to the person who answered, what their own read is worth and how many colleagues it would take before a team number was honest. This instrument does exactly that, and at ₹199 the sitting is free and the report is the product.
Psychological safety is a property of the team, not of a person, and every item here asks about the team over a declared window. That has a consequence most instruments hide: one person's answers read like facts about the team and are not yet. So the first sentence on the report, before any number, is that this is your read of this team and that one read is not the team. In the place a team score would have gone, in the largest type on the page, the report prints a refusal: how many raters it would take, at a declared and printed assumption, before a team mean is reliable enough to print, and the agreement those raters would also have to show.
Ten six-point ratings with a named end on each side and no midpoint to hide in, eight agreement statements about the team, six true-or-false descriptions of what has and has not happened, five checklists of concrete events, and four counts on a slider: how many times somebody said out loud they had got something wrong, how many times a decision changed because somebody pushed back, how many times a problem was raised before it hit a customer or a deadline, how many times somebody asked the question nobody else wanted to ask.
It is not the Speak-Up judgment test in this catalogue. That instrument presents a concern and asks what the reader would do with it; it is a judgment test and it describes the jobholder. This one asks what the team has actually done when somebody said something awkward; it measures a situation and it describes the team. The report says so on the page, and a low read here is never reported as a deficit in the reader.
The organising object is one boxed if-then sentence, alone on its panel: if a specific situation on this team, then a specific small behaviour, chosen from the facet your read puts furthest below the ordinary respondent. Implementation intentions of that shape carry a meta-analytic effect larger than most training a report could recommend, and their whole mechanism is specificity. A matching box is addressed to whoever leads the team, because half of what makes speaking up safe is not in any one member's hands, and the report splits every action into what you alone can do and what only the leader can do.
Below the box, four facets each drawn against an ordinary reference computed from a declared answer share on every item, authored high where the evidence says people answer high, so the reference is never the scale midpoint. Each carries its 68 and 95 per cent bands, a mark and a word, and the sentence that it describes the team. No total is printed across the four, because a team safe about mistakes and unsafe about disagreement is not the same as a team middling on both. It is a pulse: the report names the next sitting in about three months and prints, per facet, the movement a re-sitting would have to clear to count as real change rather than noise.
What you walk away with
One if-then sentence naming a specific situation on this team and a specific behaviour to replace what happens now, chosen from the facet your read puts furthest below the ordinary. Small enough to do on Monday, never a competency, never a verdict. A matching sentence is addressed to whoever leads the team.
In the place a team figure would have gone: no team figure, and the number of raters it would take before one was honest, at a declared reliability assumption printed with its published range and a table showing how much that assumption carries. The agreement statistic a team mean would also have to meet is printed beside it.
Admitting a mistake, asking when unsure, disagreeing upward and raising a risk early, each on its own scale against the ordinary respondent, with the 68 per cent band as a solid block and the 95 per cent band as an outlined box, sorted so the one furthest below the ordinary comes first. Every scale says it describes the team.
The facet most present on this team, paired with the cost of having it in abundance: a team where every risk is raised has learned to raise everything, and the signal thins. A report that names a cost cannot be read as praise.
Every action on the page is split in two: the part one member controls alone, and the part only whoever leads the team can change. A low read is a fact about the team's last three months, and the report never turns it into a deficit in the person who reported it.
An empty sitting scores exactly zero and describes nothing. One button pressed throughout is refused in words rather than scored. A facet too thin to carry a figure carries a placement word and its band, and the reliability table, with omega unrounded, says why.
Inside your report
Illustrative sample — your report is generated from your own responses.
One read has a reliability of 0.12 as an estimate of the team mean at the declared ICC(1) of 0.12. A mean of 18 raters would reach 0.70, and it would be printed only if they also agreed: r_wg(J) at or above 0.70.
| ICC(1) assumed | One read is worth | Raters needed |
|---|---|---|
| 0.05 (low end of the published range) | 0.05 | 45 |
| 0.12 (declared, and used above) | 0.12 | 18 |
| 0.25 (high end of the published range) | 0.25 | 7 |
If the most senior person in a meeting proposes something I think is wrong, then I will say in that meeting, before it ends, “I see it differently, because”, and give the one reason, rather than raise it afterwards.
If you propose a course of action in a meeting of this team, then before closing the point, ask one named person for the strongest case against it, and answer that case on its merits in the room.
Drawn from disagreeing upward, the facet with the largest negative departure on your read: −19 points against the ordinary respondent. One situation, one behaviour, small enough to do on Monday. Not a competency, not a verdict.
How to read it: the refusal occupies the exact place a team score would have taken, at the largest size on the page, because it is a rule the instrument applies on purpose. The box beneath it is the only thing most readers act on, so the whole page is built to earn it.
Only 3 of 8 items in this facet were answered, under the 4 a placement needs. Nothing is printed for it: no figure, no band, no placement.
| Facet | Your read | Ordinary | 68% | 95% | Placement |
|---|---|---|---|---|---|
| Admitting a mistake | 32 | 51 | 22 to 42 | 12 to 52 | Less safe than the ordinary read |
| Asking when unsure | 47 | 55 | 38 to 56 | 29 to 65 | Within the ordinary range |
| Disagreeing upward | 61 | 59 | 52 to 70 | 43 to 79 | Within the ordinary range |
| Raising a risk early | — | — | — | — | Withheld |
How to read it:each facet is one person’s read of the team against the ordinary respondent, never a team figure and never a percentile. No total is printed across the four, because a team safe about mistakes and unsafe about disagreement is not a team middling on both. A facet too thin to carry a figure is withheld in the place its scale would have gone.
Built for
- Anybody on a team who wants an honest account of whether the awkward thing can be said there, and one specific thing to do about it on Monday
- Team leads and managers who want to know which of the four facets their team's members would put furthest below the ordinary, and what only they can change
- People and culture teams who need a repeatable three-month pulse per person, with the measurement error printed and the team claim deferred until enough raters exist
- Anybody who has taken a team survey, been handed a mean, and wondered what their own answers were worth
Find out what this team does when somebody says the awkward thing, and what to do on Monday
33 items across five formats · about 12 minutes · the sitting is free and the report is the product: one boxed change, a matching one for whoever leads the team, four facets with their bands, and the printed cost of a team number. Report ₹199 in India inclusive of GST, or US$1.99 elsewhere. Sit it again in three months.
Take free · full report ₹199 (incl. GST)
Frequently asked questions
Neither is quite right, and the report says so first. Every item asks what has actually happened on your team in the last three months, so it describes the team, not you: no item has a right answer, nothing tests what you know, and nothing asks what you would do. But it is one person's read of the team, and psychological safety is a shared property. The report therefore opens with the sentence that this is your read of this team and that one read is not the team, and it prints in the place a team score would have gone exactly how many colleagues would have to sit it before a team number was honest.
That instrument presents a concern and asks what you would do with it: it is a judgment test of a person, and it describes the jobholder. This one asks what your team has actually done in the last three months when somebody admitted a mistake, asked when unsure, disagreed upward or raised a risk early: it measures a situation, and it describes the team. The two share a topic and nothing else. A low read here is never reported as a deficit in you, and the report says so on the page.
Because one rater cannot supply one. Under the standard model for aggregating individual reports to a group, one person's read has a reliability equal to the declared intraclass correlation, assumed at .12 from a published range of about .05 to .25 for climate constructs. A mean of 18 raters from the same team would reach the .70 reliability this instrument requires, and even then the team mean is printed only if the raters agree with each other, at a within-group agreement of .70 or above. The report prints all of that, with what the rater count becomes at .05 and .25, so you can see how much the assumption carries. Observed data from real teams replaces it.
It is a single boxed if-then sentence: if a specific situation on this team, then a specific small behaviour. It is drawn from the facet your read puts furthest below the ordinary respondent, and it names a replacement behaviour, never a suppression and never a competency, because specificity is the whole mechanism by which an intention of that form becomes an action. A matching sentence is addressed to whoever leads the team, since half of what makes speaking up safe is not in any one member's hands. Every action on the page is split into what you alone can do and what only the leader can.
The sitting is free and the report is the product: ₹199 in India inclusive of GST, or US$1.99 elsewhere, one time. Thirty-three items across five formats take about twelve minutes. It is a pulse: the report names the next sitting in about three months and prints, per facet, the movement a re-sitting would have to exceed to count as real change rather than measurement error. It also asks you to have the rest of the team sit it, because that is what a team figure needs. No total is printed across the four facets, no percentile appears anywhere, and you are ranked against nobody.
Each one takes a single capability, puts you inside the situations where it is actually tested, and scores your choices against published evidence — with a report designed for that capability alone, not a template. They span hiring, compliance, education, operations and personal skill.
Browse the catalogue →Methodology: Thirty-three original items across five formats: ten six-point ratings with a named end on each side and no midpoint, eight six-point agreement statements, six true-or-false descriptions of what has and has not happened, five select-every-line checklists of concrete events, and four counts on a slider. DECLARED RESPONSE INSTRUCTION: one instruction is declared for the whole instrument and it is a SITUATION-AND-BEHAVIOUR REPORT about the team over a declared recall window, the last three months. It is mixed with nothing: no item has a right answer, none tests what the reader knows, none asks what the reader would do in a hypothetical, and none asks the reader to judge a concern. The referent of every item is the team, not the reader. CONSTRUCT STATEMENT: this measures whether this team is one where a person can say the thing that is awkward, as shown by what has actually happened on it in the last three months across four facets: admitting a mistake, asking when unsure, disagreeing upward, and raising a risk early. It does not measure the reader's courage, character, judgment, communication skill, performance, personality, wellbeing or trust in any named person, and it does not measure the team's performance, engagement or satisfaction. HOW IT DIFFERS FROM THE SPEAK-UP INSTRUMENT IN THIS CATALOGUE: Speak-Up: Raising a Concern and Hearing One is a JUDGMENT test of a person; it presents concerns and asks what the reader would do with one, and it describes the jobholder. This instrument measures a SITUATION: what this team does when somebody says something awkward, and it describes the team. Different referent, different response instruction, different object on the report, and the report says so at the top of the page. SCORING DESIGN: a referent-shift single-rater climate score with a deferred aggregation claim. Every answered item is reduced to a position q in [0,1] toward the safe pole, honouring reverse keying: a six-point rating or agreement is (v - 1) / 5, a true-or-false description is 1 at the safe pole, a checklist is the share of lines whose ticked state matches the safe pattern, and a slider count is its position between the low and high ends of its range. Each facet's perception is 100 times the mean q over the items answered in it. The reference on every facet is the ordinary respondent computed from the declared marginal of every item, authored HIGH where the published evidence says people answer high, because climate self-reports skew toward the safe pole and a reference at the scale midpoint would tell every reader they are above the ordinary. The declared marginals are authored assumptions stated in the open and observed shares replace them once live data exists. Departure is perception minus reference; zero is the ordinary respondent. Reliability is McDonald's omega in its Spearman-Brown form from the answered item count and an assumed average inter-item correlation of .35, printed as an assumption. Each perception carries its standard error of measurement from a declared, modelled standard deviation, its 68 and 95 per cent bands, and a three-way placement against the ordinary respondent. THE AGGREGATION GATE, which is the design: psychological safety is a shared, team-level construct and a single sitting is one rater. The team figure is therefore REFUSED on every single sitting and the refusal occupies the place the team figure would have taken, at the top of the page in the largest type on it. Under a one-way random-effects model the reliability of a mean of k raters is ICC(1,k) = k ICC1 / (1 + (k - 1) ICC1), and the number of raters needed to reach a target reliability is k = ceil( target (1 - ICC1) / (ICC1 (1 - target)) ). At a declared ICC1 of .12 and a target of .70 that is ceil(17.11) = 18 raters; at .05 it is 45 and at .25 it is 7, and all three are printed so the reader can see how much the assumption carries. ICC1 is a declared assumption whose source range is the published climate literature, in which team-level ICC1 values typically fall between about .05 and .25, and observed data replaces it. A team mean would also have to show within-group agreement, r_wg(J) at or above .70 against a declared null distribution, or ICC(2,k), before it is a team mean; both conditions are printed. THE NULL AND ZERO: zero departure on any facet means the ordinary respondent; zero perception means the unsafe pole of that facet entirely. REFUSAL RULES: an unanswered item leaves the denominator and the reference together; an empty sitting scores exactly zero everywhere and is not reportable; fewer than twenty of thirty-three answered is not reportable; the flatness test runs on the raw presses of the eighteen six-point items and refuses the whole sitting when their standard deviation is under .35, because eight of those items are reverse-worded and one button pressed throughout produces a shape the scorer manufactured, and on a self-report no habit can be far below the ordinary so refusal is the only honest answer; a facet with fewer than eight answered items or omega under .70 on the unrounded value carries a placement word and no figure, with the refusal printed where the figure would have gone. No total is printed across the four facets, because a team safe about mistakes and unsafe about disagreement is not the same as a team middling on both. No percentile appears anywhere. A low read is never reported as a deficit in the reader: every action is split into what the reader alone can do and what only whoever leads the team can do. RE-MEASUREMENT: this is a pulse; the report names the next sitting at about three months and prints, per facet, the movement a re-sitting would have to exceed to count as real change rather than measurement error, 1.96 times the declared standard deviation times the square root of 2(1 - omega), from the Reliable Change Index, and distinguishes it from the smaller difference that would actually matter to the team. SOURCES: Edmondson, Psychological safety and learning behavior in work teams (Administrative Science Quarterly, 1999), for the construct as a shared belief of the team and its link to learning behaviour; Edmondson and Lei, Psychological safety: the history, renaissance, and future of an interpersonal construct (Annual Review of Organizational Psychology and Organizational Behavior, 2014), for the review of levels of analysis; Frazier, Fainshmidt, Klinger, Pezeshkan and Vracheva, Psychological safety: a meta-analytic review and extension (Personnel Psychology, 2017), and Newman, Donohue and Eva, Psychological safety: a systematic review of the literature (Human Resource Management Review, 2017), for antecedents and outcomes and for the finding that self-reports on this construct sit high; Chan, Functional relations among constructs in the same content domain at different levels of analysis: a typology of composition models (Journal of Applied Psychology, 1998), for referent-shift consensus composition and why an item about the team is not yet a team score; Klein and Kozlowski, From micro to meso: critical steps in conceptualizing and conducting multilevel research (Organizational Research Methods, 2000), for the conditions under which individual reports may be aggregated; Shrout and Fleiss, Intraclass correlations: uses in assessing rater reliability (Psychological Bulletin, 1979), and Bliese, Within-group agreement, non-independence, and reliability: implications for data aggregation and analysis (in Klein and Kozlowski, Multilevel Theory, Research, and Methods in Organizations, 2000), for ICC(1), ICC(1,k) and the number of raters a group mean needs; James, Demaree and Wolf, Estimating within-group interrater reliability with and without response bias (Journal of Applied Psychology, 1984), for r_wg and r_wg(J); LeBreton and Senter, Answers to 20 questions about interrater reliability and interrater agreement (Organizational Research Methods, 2008), for the .70 conventions and their limits; Morrison, Employee voice and silence (Annual Review of Organizational Psychology and Organizational Behavior, 2014), and Detert and Burris, Leadership behavior and employee voice: is the door really open? (Academy of Management Journal, 2007), for upward voice and its suppression; Van Dyck, Frese, Baer and Sonnentag, Organizational error management culture and its impact on performance (Journal of Applied Psychology, 2005), for error-management climate and the difference between a search for the cause and a search for a name; Gollwitzer and Sheeran, Implementation intentions and goal achievement: a meta-analysis of effects and processes (Advances in Experimental Social Psychology, 2006), for the if-then action; Jacobson and Truax, Clinical significance: a statistical approach to defining meaningful change in psychotherapy research (Journal of Consulting and Clinical Psychology, 1991), for the Reliable Change Index; McDonald, Test Theory: A Unified Treatment (1999), for omega; and Baumgartner and Steenkamp, Response styles in marketing research (Journal of Marketing Research, 2001), for balanced keying and the acquiescence risk. The construct is public and is cited and taught; no item from any published psychological-safety scale is reproduced or adapted, and no commercial team-scan product or branded instrument is named or used. All thirty-three items are original works written for this instrument. No competitor is named anywhere, and the instrument is not affiliated with or endorsed by any author or organisation named above.