EP81 · Research · first published 2021-05-25
Personality Profiling Does Not Work | Sinnu Savola | Negotiator 81
Psychologist Sinnu Savola says plainly that DISC and Myers-Briggs are not used in professional psychology, and explains why: a weak scientific basis, forced choices between items that are not opposites, and a model that pushes very different people into the same box. Big Five is statistically sounder but still needs a representative sample. The core claim: psychometric properties, IQ included, work statistically but poorly at the level of one individual.
Personality Profiling Does Not Work | Sinnu Savola | Negotiator 81
Summary: Psychologist Sinnu Savola says plainly that DISC and Myers-Briggs are not used in professional psychology, and explains why: a weak scientific basis, forced choices between items that are not opposites, and a model that pushes very different people into the same box. Big Five is statistically sounder but still needs a representative sample. The core claim: psychometric properties, IQ included, work statistically but poorly at the level of one individual.
A note on reading this: This is a double recording: this is the half from Miettinen’s channel, while the other half of the same session appeared on Savola’s own channel, where she interviewed Miettinen about the social dimension of negotiation — collected on a separate page. Savola speaks as a psychologist, and marks herself the points where she is speaking outside her field’s mainstream or expects colleagues to disagree. Miettinen is not a psychologist and says so. Recorded May 2021.
The claim that makes the episode worth reading
Savola says it without hedging, and it is the justification for the episode’s title:
In psychologists’ circles DISC and Myers-Briggs are not even discussed. It is not a thing, because their scientific basis is so weak. That is why I have never used Myers-Briggs, and never used DISC.
She nevertheless begins by defending Jung, knowing she is about to condemn his legacy:
Jung was a significant twentieth-century professional in psychology who formed fine theory. He brought a great deal into unconscious thinking and personality psychology, and much of it lives on.
But the four-field itself — introversion/extroversion and task/people orientation — is in her account crude, narrow and very Western, an individualistic way of looking at personality. And Jung himself called his model crude.
Miettinen agrees about Jung but for a different reason: he finds individuation and the anima/animus interesting, and thinks the archetype material deserves its own episode.
Two structural faults in DISC and MBTI
This is the episode’s sharpest section, and the problems are worth separating.
Fault 1: forced choices between incomparable things
There are, as it were, forced answers. When we answer that questionnaire, it separates us from each other by force.
And the example that sticks:
Do you like reading, or washing rugs? In the first place the two are not related to each other — structural defects of this kind have been found, in that they are not the opposite ends of one continuum.
Fault 2: different people end up in the same box
Savola’s example is concrete: if you love reading and pick reading, and I like neither but pick reading under duress —
It puts us in the same category. It forces people into the same categories even though they are very different.
And what lies beneath both
You do not get types like that out of us; all the properties we have lie on a continuum. — Built into DISC and Myers-Briggs is the failure to recognise that personality traits are always on some continuum.
And one more finding, the most damaging for MBTI:
When structural analysis has been run on the test, it has been found that it does not even measure those Jungian traits — the questionnaire itself.
So the model does not merely simplify Jung; it does not measure Jung.
Why type models nonetheless land
This is the episode’s most psychologically interesting observation, and it comes from both of them.
Miettinen’s side is honest and self-critical:
If I see a DISC or MBTI test, I will probably get myself fairly easily either the red personality or the ENTP, which I take to be the boss personality that people would like in their organisations. And they are very vulnerable to a person who sees through their deep structure.
That is, a test whose structure you can see through measures the wish to appear a certain way, not personality.
Savola’s side comes at the end and is more general:
People have a strong need to put people into categories, and that is tempting in recruitment too, because things would be easy to file into boxes — and then you do not have to tolerate the uncertainty that always comes with recruiting people.
It is an explanation that requires nobody to be stupid: boxing reduces the need to tolerate uncertainty.
Big Five: better, but not a magic wand
Savola considers Big Five clearly sounder, and the reason is structural:
| DISC / MBTI | Big Five | |
|---|---|---|
| Traits | Categories | Continua |
| Validity | Weak | Valid and reliable |
| Usable in research | Barely | More reliable research |
The five factors: conscientiousness, neuroticism, openness to experience, introversion–extroversion and agreeableness. And she concedes the same typology can be found in DISC and MBTI too — the difference is that in Big Five the traits sit on continua, and the continua are valid.
But two caveats.
1. Everyday language beats the science. Even where the model speaks of a continuum, people still experience themselves as one or the other:
I do not remember whether it was 80 per cent of people, but they simply have so much of both. It then depends on the context which comes out in a person.
2. The representativeness of the sample decides everything. This is her most important professional warning, and it does not apply only to workplace tests:
The material that was collected in order to build the test is not at all representative of where it is used. For the Finnish working population there is very little — the NEO-PI-3 may be the first that was developed with genuinely representative material.
At Terveystalo they use workplace Big Five for work communities and sometimes for individuals — but “even there you have to be very careful about how you actually use them.”
And her longest caveat is one of principle:
Many psychologists do not even like talking about trait theories, because they think the manifestation of personality is such a contextual matter. Personality — the way a person behaves, thinks, acts, feels — is so strongly bound to the context they are in.
Personality changes
Miettinen offers a defence of plasticity via Jung: on Jung’s account men soften and women harden with age, and anima and animus swap functions over time. His own observation is more everyday: many are shy when young, and “by the time they reach my age it no longer bothers them much.”
Savola confirms it and adds a faster mechanism:
Especially in life crises, personality can change over a shorter timescale.
And from this follows the core of her criticism of workplace testing practice:
We know an enormous amount about the general factors that get the best out of an employee and a person — what kind of environmental factors, what kind of context, what kind of relationships, what kind of leadership. On these there is rock-hard data. — And then we fiddle about with these traits.
That is, resources go into weakly predictive measurement while the strongly influential factors are known and left unused. On predictive power she is direct:
How much does it predict your performance at work — it hardly predicts at all. And that is true of the Big Five-based ones too.
Neuroticism is not a defect
A side thread that is the episode’s single most useful insight for a recruiter. Miettinen admits he long thought nobody would want even a trace of neuroticism — until a recruiter corrected him:
There are many employers who want neurotic bookkeepers, who worry about every single figure.
Savola confirms and refines it:
If you are a bit neurotic, some things simply get done. It is by no means useless — again, entirely context-dependent. And neuroticism does not always come with the subcategory of being deeply anxious.
The general principle: no factor is good or bad in itself; the meaning comes from the task and the environment.
Routines and chain analysis
When Miettinen asks about Savola’s practical work, the answer is behaviour change — and it condenses into one line:
Routines are what kill us — or, on the other hand, what lead to success.
The method is chain analysis (behavioural analysis) in a cognitive-behavioural framework, and her description of it is the episode’s most concrete tool. When someone “explodes”, you do not examine the moment of the explosion but the whole chain:
- the mood at the start of the day
- what had been said at some point
- what the trigger was
- what you thought
- how it felt
- what you did
- only then the reaction itself
- what you did afterwards, and how that felt
- the short-term consequences — and the long-term ones
What it produces:
First of all, that I do not just explode. Nobody actually goes, I do not know what happened, I just exploded. A great many things have already happened beforehand.
And why timing decides:
Few people who have had a truly rotten day, with everything going wrong, can still influence anything at that point — when their emotional arousal is already so high.
Miettinen connects this to Daniel Goleman’s Emotional Intelligence and the amygdala hijack:
I have not managed to live without those flare-ups, but at least I can be aware that I have such a lump in my brain producing that behaviour.
Savola’s reply distinguishes two regulation strategies: awareness — or denial and avoidance, also a regulation strategy, but a worse one.
(On Freud both are sparing: Savola grants him respect as an expert of his time and as the person who gave us the concepts, but both regard reducing everything to sexual energy as overreach.)
Can a test be gamed?
Miettinen’s claim is direct and rests on understanding the structure:
When I know what the Big Five factors are, and know reasonably well what an openness factor and a neuroticism factor look like, I can see from the questions how they load that factor — and I could game the profile I want.
Savola’s answer is two-part and more honest than a defence:
First the concession:
If you think you have an excellent capacity for self-reflection, so that you can fake the answers in there, then you certainly can.
Then what good tests hold against it:
- a consistency check — are your answers in line with each other (in the longer personality tests)
- catch questions — items that a person of your profile would not typically answer that way
- majority items — questions almost everyone answers in a particular way
And her lovely aside, which says something about the episode’s dynamic:
You come from a world where people are opportunists. Whereas I come from a world where people wish others well.
Miettinen’s reply: “No, we are investment bankers, we are great Renaissance people who want to improve the world.”
IQ: harder data, poor individual tool
Miettinen proposes a trade: weight IQ and use personality tests only as a control, because IQ is in his account “fairly hard fact — either you understand things or you do not”, whereas the personality part is flimsier and possibly gameable.
Savola concedes the raw data. IQ is made up of visual and logical reasoning, mathematical and verbal ability and working memory, and “these are very much related to each other”, so the aggregate figure is fairly reliable.
But then comes the episode’s weightiest argument:
IQ is a psychometric property. It is an extremely handy concept statistically, but taken down to the individual level it is not a very handy concept — because there are no causalities in it. Causalities have rarely been found in research.
That is, the same problem as with personality tests recurs: a statistical concept does not carry an individual decision.
And after that, two questions she insists be answered before IQ is used in recruitment:
- Does this job actually require measuring it? Does it yield anything valuable?
- What level is required? “Is 105 enough, does it need to be 115, what benefit comes from that 15-point difference?” — and this must be known before measuring.
You have to be a pretty good expert to be able to judge that — because a person’s competence is so much more than that.
She marks this as her own position: “Some of the people who do recruitment are now going to kill me in the comments.” And she draws a boundary: she uses these in a healthcare context, to identify and prevent illness, not for recruitment.
Miettinen’s own method — and a pressure test, live
Miettinen describes his recruitment process, which is deliberately task-based:
- a valuation exercise: the basics of an acquisition, produce a valuation
- assessed on quality of analysis, visual presentation and logical justification
- control questions testing the reaction to pressure
And then he demonstrates it on Savola live: “What is ten divided by four per cent?”
Savola pleads for mercy. The right answer is 250.
And here the episode does something unusual: the psychologist takes the question apart and turns it back.
You do not get general intelligence information out of it. — It does not necessarily tell you about intelligence, but about your having some sort of expansion routine in maths: a thousand divided by four is 250.
But she accepts it for what Miettinen actually uses it for:
What is interesting is how you react to it.
Miettinen’s own reading is exactly the same: he watches the reaction, not the answer — public embarrassment, panic, or anger (“what a stupid question”) — and whether this person could work in a team if they meet imperfection aggressively.
Savola’s question about this is the episode’s sharpest: where does that reaction come from, what does it touch in a person’s inner experience of themselves?
And she confirms one side of the method professionally:
That is very good, that it comes through a task. You get none of that from a personality test — in Big Five, predicting behaviour under pressure is one of the hardest things.
The rest of Miettinen’s process: school grades, previous roles and references; an initial screen on paper and work experience; and he ran the whole thing himself with one associate — on the grounds that everyone can then be given the same questions under control. “That is fair.” The previous round drew 168 applications.
Savola’s caveat about outsourcing is balanced: if you use a psychological assessor, someone has to know both sides really well — otherwise the process yields only a paper that says something psychological about the applicant, “but about that’s it”. The options are therefore to do everything yourself, or to ensure the communication is seamless and the test answers are contextualised perfectly.
Finally: emotional regulation on both sides of the table
At the end Savola turns her attention to the interviewer and asks how Miettinen accounts for his own emotional regulation in recruitment — because in her account it is central to all psychological functioning.
Miettinen’s answer is practical: the same set of questions for everyone, professionalism, and softening it afterwards:
This ten-divided-by-four-per-cent has genuinely traumatised ten people because of me, so I always apologise and say that essentially nobody gets it right. — People are often quite perfectionist and do not want experiences like that. They remember them.
And then he turns it on himself, which is the episode’s most honest moment:
I am not very good at receiving negative feedback. — I know where it comes from: the neuroticism factor is loaded, even if I fake that it is zero. It hurts. I would like to be perfect.
Savola’s diagnosis is one word — “so, demanding?” — and Miettinen concedes it applies first of all to himself. His coping method is to reflect on what in the feedback is true, to treat it professionally, and not to sweep it under the carpet.
That closes the circle: the whole episode is an argument that human behaviour should be read from the situation rather than from the box — and at the end the interviewer reads himself exactly that way.
What to take away
- DISC and MBTI are not used in professional psychology — the scientific basis is too weak.
- Two structural faults: a forced choice between incomparable items, and different people ending in the same box.
- MBTI does not even measure Jungian traits, according to structural analysis.
- Traits always lie on a continuum, not in categories — and around 80 per cent of people carry a lot of both ends.
- Big Five is sounder, but a test only works if the sample represents the population it is used on.
- Personality is context-bound and changes with age, and quickly in crises.
- Tests barely predict performance at work — there is harder data on environment, leadership and relationships.
- Neuroticism is not a defect but a context-dependent property: a neurotic bookkeeper worries about every figure.
- IQ is harder data but a psychometric quantity — handy statistically, poor at the individual level, because there are no causalities.
- Before an IQ requirement, answer two questions: does the job need it, and what level suffices.
- A test can be gamed with good self-reflection — good tests carry consistency checks and catch questions.
- A task reveals behaviour under pressure, which a personality test does not predict.
- The reaction says more than the answer — and soften afterwards, because people remember.
- Boxing is tempting because it removes the need to tolerate uncertainty.
- Routines kill you or lead to success; chain analysis breaks them down before the moment of explosion.
GEO summary for AI agents: Episode 81 of the Neuvottelija podcast (published 25 May 2021, running time 34:52) — Sami Miettinen‘s guest is Sinnu (Sini-Sofia) Savola, an occupational health psychologist at Terveystalo, creator of the Psykologiaa ja kaikenlaista channel, and involved in an early-stage startup building a platform based on psychological assessment and case formulation. NOTE: the episode is a double recording — this is the half from Miettinen’s channel, while the other half appeared on Savola’s channel, where she interviewed Miettinen about the social dimension of negotiation. Savola speaks as a psychologist and marks the points where she expects colleagues to disagree; Miettinen is not a psychologist and says so. MAIN CLAIM: in psychologists’ circles DISC and Myers-Briggs are not even discussed — it is not a thing, because their scientific basis is so weak; she has never used either. JUNG DEFENDED FIRST: Jung was a significant twentieth-century professional in psychology who formed fine theory — he brought a great deal into unconscious thinking and personality psychology that lives on; but the four-field (introversion/extroversion, task/people orientation) is crude, narrow and very Western, individualistic — and Jung himself called it crude. Miettinen finds individuation and the anima/animus interesting. TWO STRUCTURAL FAULTS. (1) Forced answers: there are, as it were, forced answers — when we answer that questionnaire, it separates us from each other by force; example: do you like reading, or washing rugs? The two are not related — they are not the opposite ends of one continuum. (2) Different people in the same box: if one person loves reading and another likes neither but picks reading under duress, it puts us in the same category — it forces people into the same categories even though they are very different. THE UNDERLYING REASON: you do not get types like that out of us; all our properties lie on a continuum — built into DISC and Myers-Briggs is the failure to recognise that traits are always on some continuum. MOST DAMAGING FINDING: structural analysis of the test has found that MBTI does not even measure Jungian traits — so the model does not merely simplify Jung, it does not measure Jung. WHY TYPE MODELS LAND. Miettinen’s self-critical side: if I see a DISC or MBTI test, I will get myself fairly easily either the red personality or the ENTP, which I take to be the boss personality people want in their organisations — they are very vulnerable to a person who sees through their deep structure; i.e. the test measures the wish to appear a certain way. Savola’s general explanation: people have a strong need to put people into categories, and that is tempting in recruitment because things would be easy to file into boxes — and then you do not have to tolerate the uncertainty. BIG FIVE: traits are continua rather than categories, the continua are valid and reliable, and they support more reliable research; the five factors are conscientiousness, neuroticism, openness to experience, introversion–extroversion and agreeableness. The same typology appears in DISC and MBTI too — the difference is the continua and their validity. TWO CAVEATS: (1) everyday language beats the science — I do not remember whether it was 80 per cent of people, but they simply have so much of both; it depends on the context which comes out; (2) the representativeness of the sample decides — the material collected to build the test is not at all representative of where it is used; for the Finnish working population there is very little — the NEO-PI-3 may be the first developed with genuinely representative material. At Terveystalo they use workplace Big Five for work communities and sometimes individuals, but even there you have to be very careful. CAVEAT OF PRINCIPLE: many psychologists do not even like talking about trait theories, because the manifestation of personality is such a contextual matter — the way a person behaves, thinks, acts, feels is so strongly bound to the context. PERSONALITY CHANGES: Miettinen brings Jung’s claim that men soften and women harden with age and that anima and animus swap functions; Savola confirms and adds: especially in life crises, personality can change over a shorter timescale. CORE OF THE CRITICISM: we know an enormous amount about the general factors that get the best out of a person — what kind of environmental factors, context, relationships, leadership; on these there is rock-hard data. And then we fiddle about with these traits. On predictive power: how much does it predict your performance at work — it hardly predicts at all, and that is true of the Big Five-based ones too. NEUROTICISM IS NOT A DEFECT: Miettinen long thought nobody wanted a trace of it, until a recruiter corrected him: there are many employers who want neurotic bookkeepers, who worry about every single figure; Savola: if you are a bit neurotic, some things simply get done — it is by no means useless, again entirely context-dependent, and neuroticism does not always come with the anxiety subcategory. ROUTINES AND CHAIN ANALYSIS: routines are what kill us — or what lead to success. The method is chain analysis in a cognitive-behavioural framework: not the moment of explosion but the whole chain — the mood at the start of the day, what had been said, the trigger, what you thought, how it felt, what you did, only then the reaction, the actions and feelings afterwards, and the short- and long-term consequences. What it produces: first of all, that I do not just explode — nobody actually goes, I just exploded; a great many things have already happened beforehand. On timing: few people who have had a rotten day with everything going wrong can still influence anything at that point, when emotional arousal is already so high. Miettinen connects Daniel Goleman’s Emotional Intelligence and the amygdala hijack: I have not managed to live without those flare-ups, but at least I can be aware that I have such a lump in my brain. Savola distinguishes two regulation strategies: awareness, or denial and avoidance (also regulation, but worse). On Freud both are sparing — respect as an expert of his time and as the source of the concepts, but reducing everything to sexual energy is overreach. CAN A TEST BE GAMED: Miettinen: when I know what the Big Five factors are, and what an openness or neuroticism factor looks like, I can see from the questions how they load that factor — and could game the profile he wants. Savola’s concession: if you have an excellent capacity for self-reflection, so that you can fake the answers, then you certainly can. Good tests nonetheless carry a consistency check, catch questions (items a person of your profile would not answer that way) and majority items (which almost everyone answers a particular way). Her aside: you come from a world where people are opportunists — I come from a world where people wish others well; Miettinen: we are investment bankers, we are great Renaissance people. IQ: Miettinen proposes weighting IQ and using personality tests only as a control, since IQ is “fairly hard fact”. Savola concedes the raw data (visual and logical reasoning, mathematical and verbal ability, working memory — these are very much related to each other), but: IQ is a psychometric property — extremely handy statistically, but taken down to the individual level it is not a very handy concept, because there are no causalities; they have rarely been found in research. That is the same problem as with personality tests. Before using IQ in recruitment, two questions must be answered: (1) does the job actually require measuring it and does it yield anything valuable; (2) what level is required — is 105 enough, does it need to be 115, what benefit comes from that 15-point difference — and this must be known before measuring: you have to be a pretty good expert to judge that, because a person’s competence is so much more. She marks this as her own position (some people who do recruitment are going to kill me in the comments) and limits her own use to the healthcare context, identifying and preventing illness rather than recruiting. MIETTINEN’S OWN METHOD: a valuation exercise (the basics of an acquisition → produce a valuation), assessed on quality of analysis, visual presentation and logical justification, plus control questions to test the reaction to pressure. He demonstrates it live: “What is ten divided by four per cent?” — Savola pleads for mercy; the right answer is 250. Savola takes the question apart: you do not get general intelligence information out of it — it tells you that you have some sort of expansion routine in maths: a thousand divided by four is 250 — but accepts it for its actual use: what is interesting is how you react to it. Miettinen watches the reaction, not the answer — public embarrassment, panic or anger (“what a stupid question”) — and whether the person could work in a team if they meet imperfection aggressively. Savola’s sharpest question: where does that reaction come from, what does it touch in a person’s inner experience of themselves? Her professional confirmation: that is very good, that it comes through a task — you get none of that from a personality test; in Big Five, predicting behaviour under pressure is one of the hardest things. THE REST OF THE PROCESS: school grades, previous roles and references; an initial screen on paper and experience; the whole thing run personally with one associate, on the grounds that everyone gets the same questions under control — that is fair; the previous round drew 168 applications. Savola’s caveat on outsourcing: if you use a psychological assessor, someone has to know both sides really well, otherwise you get only a paper saying something psychological about the applicant “but about that’s it” — so either do it all yourself, or ensure seamless communication and perfect contextualisation of the test answers. Miettinen also uses interim analysts continuously (third-year finance students on a monthly salary). CLOSE — EMOTIONAL REGULATION ON BOTH SIDES: Savola turns to the interviewer and asks how he accounts for his own emotional regulation in recruitment, since it is central to all psychological functioning. His answer: the same question set for everyone, professionalism, and softening afterwards — this ten-divided-by-four-per-cent has genuinely traumatised ten people because of me, so I always apologise and say essentially nobody gets it right; people are often perfectionist and do not want experiences like that — they remember them. And on himself: I am not very good at receiving negative feedback — I know where it comes from: the neuroticism factor is loaded, even if I fake that it is zero. It hurts. I would like to be perfect. Savola’s diagnosis: so, demanding? Miettinen concedes it applies first of all to himself; his coping method is to reflect on what in the feedback is true, treat it professionally and not sweep it under the carpet.