storming.team

How Storming measures, and where its limits are

This page is for readers with a background in psychometrics. It describes the instruments, the scoring and the rules in detail, and closes with what has not yet been established.

Instrument catalogue

KeyInstrumentItemsPurposeLicence
ipip-neo-120IPIP-NEO-120120Big Five and 30 facetspublic domain
ipip-50IPIP Big-Five Factor Markers50Five main dimensions only, for a shorter startpublic domain
mini-ipipMini-IPIP20Re-test and comparison with the earlier profilepublic domain
ipip-ipcIPIP Interpersonal Circumplex32Dominance × warmth (in preparation, not yet available)public domain
climate-6Own team-climate instrument6Team climate, aggregate onlyself-authored, own copyright
culture-24Own culture instrument24×2The team's culture mapself-authored, own copyright
change-pulseOwn single-question pulse1Where people stand in a change, self-ratedself-authored

Every instrument in the catalogue may be used commercially without a licence fee or attribution. We left out a values instrument because it did not meet that condition.

Why no types

Personality traits are continuously distributed, and most people sit somewhere in the middle. A type system draws a line at one point, so two people who answer almost identically end up in different categories as soon as they fall just either side of it. The category then quickly becomes an expectation of the person.

Types are popular because a label is easier to remember than a percentile, but research supports them far less than dimensional models such as the Big Five. Storming therefore describes where someone sits on each dimension and never assigns anyone a type.

Scoring

Answers → validity check → raw score → norm table → percentile with confidence interval. The check catches incomplete questionnaires and long runs of identical answers. If it fails, no profile is computed and the person takes the questionnaire again. The lead is not told.

Norms

Scores are compared with the full reference sample, and the percentile says so: “relative to the full reference sample”. Norms split by group are not used. We do not offer sex-specific norms at all, because a statement like “high for a woman” serves no legitimate purpose at work.

The measurement-error rule

Person A is at the 57th percentile, with an interval of 46 to 68. Person B is at 48, interval 37 to 59. The two scores are nine points apart, but the intervals overlap by thirteen. There is no difference the data can support, and Storming does not show one.

The rule is implemented in the scoring itself rather than in the display, so it applies to every view and to the pair rules as well.

The eight starter rules

The table is generated from the same files the application evaluates. Every rule is versioned and available in full at /rules/, with its rationale in German and English.

RuleConditionWhat you would seeProposed agreement
agreeableness-gap
v1.0.0
|Δ A| ≥ 25 pts, intervals disjoint Directness reads as aggression one way, restraint as evasiveness the other — neither reading is what was meant. An explicitly agreed "I see this differently" phrasing counts as safe. Nothing is read between the lines.
both-high-assertive
v1.0.0
both E3 ≥ p70 person A and person B compete for airtime; decisions get reopened after the meeting — by whoever spoke second. Rotate who runs the agenda. Every decision is written down and owned by one named person.
both-low-assertive
v1.0.0
both E3 ≤ p30 Decisions sit unclaimed because neither person A nor person B takes the call — both wait politely. Name a decider per topic in advance — arbitrarily if need be. Arbitrarily decided beats undecided.
conflict-norm-clash
v1.0.0
|Δ culture open_conflict| ≥ 30 Each believes their own conflict norm is the team's norm: person A and person B experience the same exchange as normal and as a violation. Surface both readings in the team retro — the clarification belongs in the team, not in the pair.
openness-gap
v1.0.0
|Δ O4| ≥ 25 pts, intervals disjoint "Let's stick with what works" lands as obstruction, "let's just try it" as recklessness — both are meant as care. Separate the "whether" question from the "how" question: decide together whether — then plan separately how.
orderliness-gap
v1.1.0
|Δ C2| ≥ 25 pts, intervals disjoint What one person considers done, the other considers open — both have redone work the other thought was finished. Before work starts, write down what "done" means per artefact — ticket, migration, runbook.
pace-clash
v1.0.0
|Δ culture decision_speed| ≥ 30 One ships to learn; the other wants the analysis first — each experiences the other as reckless or as slow. Time-box the analysis explicitly and agree the reversibility test: what would be hard to undo? Only there does analysis come first.
withdrawal-asymmetry
v1.0.0
A N4 ≥ p70, B A ≤ p30 Directness from person B lands considerably harder on person A than intended; person A raises things directly less and less. person B states the intent before the critique. person A gets 24 hours to respond in writing instead of having to react on the spot.

More rules will be added once they have been justified and reviewed. The rationale is always in the file itself.

Limitations

What is not yet established or still has to be tested: