About half the people who take MBTI get a different result six weeks later. That's not measurement error. That's the test telling you it measured nothing.
The descriptions feel accurate anyway. This is the Barnum effect: vague, positive statements that almost anyone accepts as personally true. Horoscopes are written the same way. The feeling of recognition is not evidence the test worked.
DISC is marginally better. "Marginally better than astrology" is not a useful bar.
The Big Five
The research that holds up is the Big Five: openness, conscientiousness, extraversion, agreeableness, neuroticism. Six decades of testing across cultures and settings. Validity coefficients of 0.3 to 0.5 for predicting real outcomes. Not perfect, but real.
Conscientiousness has the strongest track record. A 1991 meta-analysis of 117 studies (N = 24,000) found it predicts job performance across every occupational group tested. No other trait comes close.
Base rates, not verdicts
High agreeableness, low neuroticism: you probably take critical feedback as information rather than attack. High openness, low conscientiousness: you generate more ideas than you finish. These are base rates for your own behaviour, describing patterns that show up often enough to be worth knowing.
MBTI gives you a label. Labels tell you what you are. Base rates tell you what tends to happen when you make decisions under pressure, when you receive feedback you didn't ask for, when a project drags on past the point you thought it would end.
That distinction is why one is useful and the other isn't.
What Kalibr measures
Kalibr draws on four published instruments. The output covers 10 dimensions: Philosophy Cohesion, Drive Alignment, Bonding Index, Adaptive Intelligence, Volatility Vector, Ambiguity Tolerance, Influence Style, Trust & Challenge, Temporal Orientation, and Energy Resilience. Grit and resilient coping are supplementary scores that feed several of those dimensions; they are not themselves dimensions.
Each traces to published research. Scoring is deterministic: same inputs, same outputs every time, no AI at the measurement stage.
The output is specific enough to use. Not "you're a planner" but "you over-weight precision and under-weight speed when the cost of delay is high." That specificity is the whole point.