Pay and Reward Decision Assessment for Managers and HR TeamsEvery raise you approve is a rule somebody else will quote back to you.
Sixteen pay and reward dilemmas a manager actually faces, and twenty-two more exercises about what you would want in front of you first. Five readings, each corrected against a typical respondent and each carrying its band. The report writes every finding twice, once to you and once to whoever runs reward, and the second column never reaches your copy.
Judgement on pay, measured as judgement, with the maths on the page
The Pay and Reward Decision Assessment for Managers and HR Teams is a thirty-minute behavioural-tendency assessment of how a person decides individual pay: whether evidence arrives before the number, whether like cases are treated alike, whether each decision is made as the precedent it becomes, whether it can be explained, and whether the boundary holds under pressure.
Most pay training teaches a policy and most pay decisions are made without one in the room: a counter-offer on a Thursday, a market figure quoted with nothing behind it, a manager three weeks before the cycle saying his person will leave, a bonus promised by somebody who has since left. This instrument puts those situations in front of the reader and asks what they are most likely to do, with four moves on every scenario that a capable manager would defend. The difference between the moves is degree, not sense, and the errors are spread across doing the wrong active thing and deferring, so no fixed habit pays.
The scoring is construct-weighted graded situational judgement, with the part most instruments skip: every option carries a sparse credit vector that says which of the five constructs it is evidence about, declared in the item and read by the scorer. Every option credits one construct, sometimes two, never three, and the report prints the count. If every option counted towards every construct the five readings would correlate at about .9 and the report would say one thing five times; the structure here gives an expected correlation of about .30, printed as an assumption that observed data will replace. Every figure is corrected against a declared typical respondent, drawn from an authored answer prior on every exercise, and every figure carries its 68 and 95 per cent band, drawn and written in plain words.
The report is the Two-Column Translation. Each of the five findings is written twice from one placement: a sentence to the person who sat it, and a sentence to whoever runs reward in their organisation, in the form of something they can do. Because both come from the same row they cannot drift apart, and because the second column is printed only on a manager copy the candidate never reads a description of themselves written for somebody else. Under the findings sit the credit bars, one if-then sentence drawn from the reader's own widest gap, a reliability table that prints every refusal, and the size of change on a second sitting that would count as real.
The other twenty-two exercises ask what a decision-maker would want in front of them before answering, which claims about how pay decisions behave they would rely on, how they would read a small pay table or a market spread, and what would settle each kind of request. The keying draws on the equity and organisational-justice literature, on pay-secrecy and pay-communication research, on anchoring and first offers in negotiation, on the compensation-structure work on ranges and compression, and on the situational-judgement validity literature; the methodology note names every source.
This is a legally sensitive subject and the instrument treats it that way. No country's law on pay, pay transparency, equal pay, minimum wage or reporting is named or keyed anywhere; every dilemma is built on principles that hold in any setting. It is a judgement assessment and not legal or compensation advice, it audits nobody's pay practices, and it says nothing about whether any organisation's pay is fair. The report says all of that first, before any number, and again at the end.
What you walk away with
Whether offers, quoted rates and current salaries move you before their source and scope do. Read from the counter-offer, the market claim, the recruiter's ask and the candidate's anchor.
Whether the same work at the same standard lands on the same terms, and whether a known gap waits for a complaint. Read from the two analysts, the compressed role and the confidentiality request.
Whether each exception, promise and quiet payment is decided as the rule it becomes. Read from the departed manager's bonus, the unfunded promotion and the off-record retention payment.
Whether a no arrives with the range, the reason and what would move it, without touching anyone else's number. Read from the team-meeting question, the capped performer and the leaked peer figure.
Whether a threat to leave, a senior favour, a hard month or a request for secrecy moves the basis pay is decided on. Read from the retention request, the leader's ask and the hardship case.
Every reading as a sentence to you and, on the manager copy only, a sentence to whoever runs reward: what to give you before the conversation, what to ask before approving. One structure, two columns, no drift.
Inside your report
Illustrative sample — your report is generated from your own responses.
You settle a market-rate argument with a number before you have heard what the number is being asked to do.
Give her the range and the constraint before the conversation, not the answer after it.
You held the basis for pay steady under a resignation threat, a senior request and a hardship case, by naming the basis rather than refusing the person.
She will say no to a senior request on the same grounds she says no to anyone else. Back her when the request comes to you instead.
Both columns come from one data structure, so they cannot drift apart. The candidate's copy carries the left column and a note that the right one exists; the right column is printed only for somebody with manager access to the sitting, and contains nothing the left does not.
Zero is a respondent choosing the way people typically choose, computed from the declared answer prior on every exercise. A reading with too few exercises behind it, or an assumed reliability under .70, carries a word and no point, and the reason is printed in the reliability table.
Built for
- Line managers who approve, propose or explain individual pay decisions and want to see the rules they are writing
- HR business partners and reward teams deciding how to brief managers before the pay cycle, and what evidence to ask for
- Organisations building a manager-capability programme around pay, promotion and retention decisions
- Anyone stepping into a first role with a pay budget, before the first counter-offer lands on their desk
Find out which pay rules you are writing without meaning to
38 exercises across five formats · about 30 minutes · five readings against a declared typical respondent, every figure with its band, and the report written twice from one set of answers.
₹899 (incl. GST) · assessment and full report, nothing further to pay
Frequently asked questions
No. It names no country's law on pay, pay transparency, equal pay, minimum wage or reporting, and no organisation's policy or grading structure. Every dilemma is built on principles that hold anywhere: evidence, consistency, precedent, explanation and boundary. It is a judgement assessment, not legal or compensation advice, it audits nobody's pay practices, and it says nothing about whether any organisation's pay is fair.
One overall figure against a typical respondent with its 68 and 95 per cent bands, its contributors printed beneath it. Five findings, each a placement or a banded figure, each written as a sentence to you with what it gets you and what it costs, and the scenario it was drawn from. One boxed if-then sentence from your widest gap. A reliability table that prints every refusal, what zero means, and the size of change a second sitting would need to show.
Only somebody with manager access to the sitting in the organisation that assigned it. Every finding is written twice from one placement: a sentence to you, and a sentence to whoever runs reward about how to brief and support you on pay decisions. Your copy carries your column and a note that the other exists. The second column contains nothing yours does not: no risk flag, no inferred trait, no note on how you answered.
Every scenario option carries a value from zero to three and a sparse credit vector saying which construct it is evidence about, declared in the item and read by the scorer. Each construct's figure is the credit earned over the credit available on the exercises that bore on it. Zero is a respondent choosing the way people typically choose, computed from an authored answer prior on every exercise, and every figure carries its band from an assumed reliability printed unrounded.
A construct carries a figure only when at least eight effective exercises bore on it and the assumed reliability is at least .70. Below that a figure would look precise and be noise, so the report prints a three-way placement in words and says in the reliability table why the figure was withheld. A sitting with fewer than fourteen exercises or six scenarios answered is refused an overall figure altogether.
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-eight original exercises across five formats: sixteen pay and reward scenarios with four graded moves each; six select-every-that-applies exercises asking what you would want in front of you before answering; six true-or-false claims about how pay decisions actually behave; five judgement-of-evidence exercises that read a small pay table or a market spread; and five match-the-following exercises pairing a stated reason for a pay request with the thing that would settle it. One response instruction is declared for the whole instrument and it is a BEHAVIOURAL-TENDENCY instruction: what you are most likely to do, want in front of you, act on, or rely on, never what a textbook says should be done. Construct statement: this measures the judgement a manager or reward professional brings to individual pay and reward decisions, in five parts: getting evidence before committing to a number, treating like cases alike, deciding each case as the precedent it becomes, explaining a decision so that the person can check it, and holding a boundary under pressure, sympathy or a request for secrecy. It does not measure knowledge of any country's law on pay, pay transparency, equal pay, minimum wage or reporting; it is not legal advice and not compensation advice; it does not audit any organisation's pay practices; and it says nothing about whether any organisation's pay, or the respondent's own pay, is fair. No law, regulator, statute or reporting threshold is named or keyed anywhere; every dilemma is built on principles that hold in any setting. Scoring design is C1, construct-weighted graded situational judgement. Every scenario option carries an effectiveness value from zero to three and a sparse construct-credit vector: every option credits the scenario's primary construct at weight one, and at most two options on a scenario credit one secondary construct at weight one half. Sixty-four option credits across the sixteen scenarios: thirty-two name a single construct and thirty-two name two; no option names three. The credit vectors are declared in each scenario's content.construct_credit and read by the builder; nothing is inferred. For each construct, the score is the sum of credited value on answered exercises divided by the maximum attainable on those same exercises, and the twenty-two non-scenario exercises each credit one construct at weight one. The sparsity is the design: if every option credited every construct, the five readings would correlate at about .9 and the report would say one thing five times. Under independent construct abilities the credit structure alone produces an expected correlation between constructs of about .04; allowing for a shared judgement factor the expected figure is about .30, with a plausible range of .04 to .44; observed data will replace this assumption. Every figure is chance-corrected against a declared null: on every exercise an answer prior is authored, the share of ordinary respondents expected to choose each option, select each option, key each claim, or match each pair, and zero on the reported scale is what a respondent drawing from those priors would score. Refusal rules: an unanswered exercise leaves the numerator, the denominator and the chance term together; an empty sitting scores exactly zero; a sitting with fewer than fourteen exercises answered, or fewer than six scenarios answered, is refused a headline figure and the refusal is printed where the figure would be; a construct carries a number only when it has at least eight effective exercises answered and an assumed omega of at least .70, and otherwise carries a three-way placement with the refusal printed. Omega is assumed from item count and a stated inter-item correlation until live data exist, and is printed unrounded. No percentile appears anywhere because there is no norm group. The report is the Two-Column Translation: every finding is written twice from one data structure, once to the person who sat it and once to whoever runs reward in their organisation, and the second column is never shown on the candidate's copy. Every item is an original work; the instrument is not affiliated with, derived from or endorsed by any commercial instrument, certification body or professional association. Sources drawn on: Adams, equity theory, on inputs and outcomes judged against a referent; Leventhal, on the procedural rules of consistency, accuracy, bias suppression and correctability; Colquitt, on the distributive, procedural, interpersonal and informational dimensions of organisational justice; Folger and Konovsky, on procedural and distributive justice in reactions to pay-rise decisions; Shaw, Wild and Colquitt, meta-analysis of explanations and their effect on acceptance of decisions; Tversky and Kahneman, on anchoring and adjustment; Galinsky and Mussweiler, on first offers as anchors in negotiation; Colella, Paetzold, Zardkoohi and Wesson, on the costs and benefits of pay secrecy; Card, Mas, Moretti and Saez, on the effect of learning peers' pay on satisfaction; Bamberger and Belogolovsky, on pay secrecy and individual performance; Milkovich, Newman and Gerhart, on pay structures, ranges and compression; Bazerman and Moore, on consistency, escalation and precedent in managerial judgement; McDaniel, Hartman, Whetzel and Grubb, on situational judgement validity and behavioural-tendency response instructions; Lievens, Peeters and Schollaert, on situational judgement construct validity; and Gollwitzer and Sheeran, on implementation intentions, which the report's one if-then sentence follows.