Personality Tests, Myers-Briggs etc

[FlammableFlower]

As has been pointed out above, appointing someone is a lengthy process and shortlisting generally gets harder as the pool of applicants gets bigger. However, the use of these tests, frequently way outside the parameters of what they were originally conceived for, bothers me. Not only are they an incredibly blunt tool, but I suspect much misused. Great, you can divide your applicants into 16 personality types - but are you correct in assuming that you know which personality type will work for that particular job?* And then there's gaming the system - there are already websites where you can practise these tests. In all you can easily lose people who would have been good.

*Not that that isn't true of a more manual process - I'm just against the idea of reducing things to simple labels. For example the first postdoc position I applied for, I got immediately rejected. Odd, I thought I was well suited... 2 months later I was surprised to be called for an interview and got the position. It turns out that initially the small company made up of biologists had decided that they needed a biochemist for the synthesis and testing of their planned electrochemical labels. After their first round of interviews they realised that maybe they had been shortlisting people with the wrong skill set. Daft, you might say, but at least it was a human error and being of quite a technical and specialist nature one that was easily remedied. Chucking people off the shortlist because the computer (test) says "no" is quite different.

[Boustrophedon]

To be fair to Myers Briggs, at least I get consistent results when I do it. However, they're basically unemployable results

Employers need to cut this sh.t out. For my sake, if nothing else

You'll have to start cramming for your personality tests.

My Son does that, he regularly get zero for psychopathy, which should be a red flag to anyone with any sense.

[bagpuss]

I've paid quite large sums to use MBTI with my medical students over the years. Not because I felt it was much use, but to stimulate conversations and thinking about how different kinds of people might contribute to/be essential for a team, and to counteract the temptation to think that "how I operate is the CORRECT way, and any other way is therefore wrong". My own outcome was pretty stable.

I also had a free app on my phone, which did largely the same thing, but also placed you in a Hogwarts House.


BoaF wrote

I mean, we all know interviews don't really work, but how many people here have ever refused to turn up for one on the basis that it's quackery

Interviews can detect the (rather rare) highly deviant characters. I've been told "Team working? I don't know the meaning of the word! I just do my job and go home!"

Highly structured interviews are better than unstructured interviews. Multiple Mini Interviews are better still (where short interviews are held with a number of separate individuals) in terms of predictive validity for the role.

Where you have, say, 28,000 applicants for 8,000 places, almost any filter is helpful, but cognitive ability tests, prior educational achievement, and Situational Judgement Tests have all been shown to have predictive value. SJTs measure something different from prior educational achievements.

If I had to use a personal qualities test, I would try NEO PI-R, especially with regard to the trait of conscientiousness, which is generally the single biggest predictor of workplace performance.

Sorry, I don't have time to add references at the moment, but can if anyone is interested, and if time permits.

Yes please.

Certainly for technical roles, I think interviews are useful. Comparing people with somewhat relevant post-doctoral experience and seeing how they actually approach problems that are slightly outside their background is informative

While interviews are extremely imperfect, I'm not sure that there's another way to identify certain necessary characteristics for the type of people I need in my team. When recruiting, I can use the skills and experience on their CV/application to identify those who have the relevant skills and experience but what I can't get from any written application are 2 necessary things:

1. A feeling for how they think about and approach an analytical problem. We set people problems that require them to work something out from information we give them, or just to estimate something based on no information whatsoever. We're not particularly interested in what their final answer is, we're interested in how they approach the challenge, whether they sense check what they're doing, the sort of questions they ask themselves or the interviewer(s), etc.

2. Whether they can communicate verbally with a non-technical human - a part of the job which is at least as essential as being able to do the data-geekery.

I have yet to find any way other than interview to assess either of these. Trust me, if I could, I would - I hate interviews with a passion. I have met so many candidates over the years who are absolutely perfect on paper but who I simply could not put in front of a client and get them to present their work. In a bigger team, this wouldn't necessarily be a problem as we could have people doing the technical stuff and others doing the client-facing side, but we're a very small team and everyone has to be able to do all of it.

There were also the 2 very memorable candidates, both of whom were excellent on paper for the junior role we were recruiting for, but neither of whom was employable. One spoke barely a word of English and didn't even engage with me when I gave up asking questions in increasingly simple English and spoke to him in French - I'd have been prepared to offer him a job despite the lack of English, if he'd tried, given that I was at the time fluent in French and, as he was a junior, he wouldn't have needed to go in front of clients - this was in a previous job where not everyone needed to be client-facing. The other, on being asked why he'd left his previous job at a desirable employer after only a short time, told us without any hesitation that he'd come very close to punching his boss. He then went on to say that this was because, as a graduate trainee, he had disagreed with what he was being asked to do so had done what he thought he should do instead. Despite repeated warnings, he'd continued on this course and when his boss told him that if he didn't do what they'd asked him to do, then he would get a formal warning and the dismissal process would be started, he'd physically threatened him and very nearly punched him. He seemed to have zero awareness that this would have a negative effect on the chances of us employing him. He'd looked excellent on paper - the only question mark being his having left that employer after less than a year.


I can't imagine ever using a personality test, though. I'm not sure what I'd even be looking for from a personality test in terms of who I'd want to employ. I just want someone who has the technical and analytical nous to be able to do analysis, write any code necessary, and then the ability to communicate their findings to a client who might not have any analytical nous of their own. And someone who isn't actively hostile to me or anyone else in the team (yeah, that misogynist contractor didn't last long in our then majority female team, although he was my then boss's fault). Other than that, extra points for someone who likes baking and brings the results into the office (in non-COVID times when we're actually in the office)

[nezumi]

To be fair to Myers Briggs, at least I get consistent results when I do it. However, they're basically unemployable results

Employers need to cut this sh.t out. For my sake, if nothing else

You'll have to start cramming for your personality tests.

My Son does that, he regularly get zero for psychopathy, which should be a red flag to anyone with any sense.

This is brilliant!

Surely a good chunk of the problem with using these tests is that you absolutely can just pretend to bo somebody else to "pass" it. Maybe if the advert was more accurate to the position applied for and the vibe of the company then those who wouldn't be suited simply wouldn't apply.

This is particulary important in callcentre work like what I do right now. I am an introvert on a sales team. If they had used personailty tests I simply wouldn't have got the first interview, as it happens I got the highest score at interview that anyone has apparently ever had* and I perform pretty well**.

There really needs to be a better way to whittle people down to the actual best candidates than an arbitrary grouping based on an evidence-free questionnaire which is so vague as to be useless and is eminently cheatable. What is that? Pfffft, your guess is as good as mine.

* I have now worked for enough callcentres to know precisely what they'll ask, what my best examples are, what the follow ups will be and so on.
** I'm above average on the phones and sh*t hot at anything tedious, complicated or difficult. They give me all the fault callbacks and awkward system problems to handle. It's very basic but way more interesting than talking about bills. I am wasted in this job




Edit: bagpuss please let me in on what kind of data geekery you do?

[bagpuss]

Edit: bagpuss please let me in on what kind of data geekery you do?

I work for a marketing services company so it's a whole range of marketing-related/customer-related reporting and analytics. Anything from a simple report in Excel, Tableau dashboards monitoring a marketing campaign, customer profiling, segmentation, analysis of customer surveys, predictive modelling, estimating customer lifetime value, scoring potential prospects, identifying potential markets and, quite frankly, anything that might involve numbers that a client might be struggling to do themselves.

[Allo V Psycho]

I've paid quite large sums to use MBTI with my medical students over the years. Not because I felt it was much use, but to stimulate conversations and thinking about how different kinds of people might contribute to/be essential for a team, and to counteract the temptation to think that "how I operate is the CORRECT way, and any other way is therefore wrong". My own outcome was pretty stable.

I also had a free app on my phone, which did largely the same thing, but also placed you in a Hogwarts House.


BoaF wrote



Interviews can detect the (rather rare) highly deviant characters. I've been told "Team working? I don't know the meaning of the word! I just do my job and go home!"

Highly structured interviews are better than unstructured interviews. Multiple Mini Interviews are better still (where short interviews are held with a number of separate individuals) in terms of predictive validity for the role.

Where you have, say, 28,000 applicants for 8,000 places, almost any filter is helpful, but cognitive ability tests, prior educational achievement, and Situational Judgement Tests have all been shown to have predictive value. SJTs measure something different from prior educational achievements.

If I had to use a personal qualities test, I would try NEO PI-R, especially with regard to the trait of conscientiousness, which is generally the single biggest predictor of workplace performance.

Sorry, I don't have time to add references at the moment, but can if anyone is interested, and if time permits.

Yes please.

Certainly for technical roles, I think interviews are useful. Comparing people with somewhat relevant post-doctoral experience and seeing how they actually approach problems that are slightly outside their background is informative

While interviews are extremely imperfect, I'm not sure that there's another way to identify certain necessary characteristics for the type of people I need in my team. When recruiting, I can use the skills and experience on their CV/application to identify those who have the relevant skills and experience but what I can't get from any written application are 2 necessary things:

1. A feeling for how they think about and approach an analytical problem. We set people problems that require them to work something out from information we give them, or just to estimate something based on no information whatsoever. We're not particularly interested in what their final answer is, we're interested in how they approach the challenge, whether they sense check what they're doing, the sort of questions they ask themselves or the interviewer(s), etc.

2. Whether they can communicate verbally with a non-technical human - a part of the job which is at least as essential as being able to do the data-geekery.

I have yet to find any way other than interview to assess either of these. Trust me, if I could, I would - I hate interviews with a passion. I have met so many candidates over the years who are absolutely perfect on paper but who I simply could not put in front of a client and get them to present their work. In a bigger team, this wouldn't necessarily be a problem as we could have people doing the technical stuff and others doing the client-facing side, but we're a very small team and everyone has to be able to do all of it.

There were also the 2 very memorable candidates, both of whom were excellent on paper for the junior role we were recruiting for, but neither of whom was employable. One spoke barely a word of English and didn't even engage with me when I gave up asking questions in increasingly simple English and spoke to him in French - I'd have been prepared to offer him a job despite the lack of English, if he'd tried, given that I was at the time fluent in French and, as he was a junior, he wouldn't have needed to go in front of clients - this was in a previous job where not everyone needed to be client-facing. The other, on being asked why he'd left his previous job at a desirable employer after only a short time, told us without any hesitation that he'd come very close to punching his boss. He then went on to say that this was because, as a graduate trainee, he had disagreed with what he was being asked to do so had done what he thought he should do instead. Despite repeated warnings, he'd continued on this course and when his boss told him that if he didn't do what they'd asked him to do, then he would get a formal warning and the dismissal process would be started, he'd physically threatened him and very nearly punched him. He seemed to have zero awareness that this would have a negative effect on the chances of us employing him. He'd looked excellent on paper - the only question mark being his having left that employer after less than a year.

Indeed, one of the good ways of telling if some will be able to do the job in the workplace is a work-sample where possible. And I don't think you can fake communication skills - if you can fake them, you have them when you need them, so if you observe those in an interview, it's a start. But often selecting for interview is a problem.

And I DID have a candidate tell me they had left their last job after punching their line manager. And then at the end, with the non-scored feel-good exit question, "What is the thing you are most proud of in your career?" they returned to 'punching the line manager'. As the prospective line manager, I wasn't too thrilled.

Incidentally, while faking answers always seems possible, nonetheless, tests like SJTs still measure something useful. Perhaps being able to work out what is wanted in the test is a relatively high level skill...

[nezumi]

(Aside: I do get free free linked in learning from work and it has courses in Data Science so let's see... I must admit, I have starry eyes - it sounds like my kinda gig. I bet the learning curve is brutal though )

[Squeak]

(Aside: I do get free free linked in learning from work and it has courses in Data Science so let's see... I must admit, I have starry eyes - it sounds like my kinda gig. I bet the learning curve is brutal though )

Just a little thought, based on what you've said in this thread and elsewhere about work and career goals... KerryA1 and I are both data managers, which might be quite a good fit for the kinds of skills you have and are interested in acquiring, and that won't necessarily need a very specific educational background...

Data manager can mean anything from a highly technical job quality-assuring and tidying/documenting datasets, through to building databases or systems to manage those datasets, and into the almost entirely squishy human work that I do, wrangling the people with the technical skills to pull in the same direction. (My most important professional skill is the ability to translate different dialects of geek.)

Because it's so varied and relatively young as a career, there aren't particularly ossified pathways in. Most jobs largely consist of knotty problems, plus the ability to help scientists work out what they need to do to navigate a system that they don't quite understand but know they are obliged to use.

A few data science classes might be quite a good complement to your call centre skills, plus a smattering of domain knowledge for the field you're going into. (Mind you, I'm a terrestrial ecologist, working with ocean science data from a range of disciplines, so I have had to ask a *lot* of dumb questions.)

Anyway, just a thought...

[noggins]

The trick with the work related personality questions is to imagine you are answering them while working at a wonderful job doing interesting work for a great cause for top pay with lovely working conditions and super colleagues (etc). It’ll do wonders for your scores.

[Bird on a Fire]

As a certain kind of Marxist, I refuse to work for any company that would have me as an employee.

[Millennie Al]

There really needs to be a better way to whittle people down to the actual best candidates than an arbitrary grouping based on an evidence-free questionnaire which is so vague as to be useless and is eminently cheatable. What is that? Pfffft, your guess is as good as mine.

If different employers use different tests with different systematic biases, things are not too bad. Those using tests with bad biases employ unsuitable people and are more likely to go bust, while the wide variety of biases means that everyone gets a chance of doing well somewhere. If employers settle on a small number of possible tests, candidates could find themselves almost unemployable if they do badly on those tests.

In some fields it may be easier than others. In programming, for example, it seems that jobs attract applicants who appear qualified until you give them a trivial task when you discover that many are completely useless despite holding a relevant degree and maybe also having experience:
https://blog.codinghorror.com/why-cant- ... s-program/

[bolo]

As a certain kind of Marxist, I refuse to work for any company that would have me as an employee.

This remark makes you sound a little grouchy.

[nekomatic]

Since data science classes were mentioned, let me put a word in for the Weka courses from the University of Waikato. Weka itself is a bit funky but the courses IMHO are very good at grounding you in the things you need to fundamentally understand rather than whizzing straight to clever code. Plus they’re presented by a chummy Kiwi who plays the clarinet.

[Allo V Psycho]

Highly structured interviews are better than unstructured interviews. Multiple Mini Interviews are better still (where short interviews are held with a number of separate individuals) in terms of predictive validity for the role.

Where you have, say, 28,000 applicants for 8,000 places, almost any filter is helpful, but cognitive ability tests, prior educational achievement, and Situational Judgement Tests have all been shown to have predictive value. SJTs measure something different from prior educational achievements.

If I had to use a personal qualities test, I would try NEO PI-R, especially with regard to the trait of conscientiousness, which is generally the single biggest predictor of workplace performance.

Sorry, I don't have time to add references at the moment, but can if anyone is interested, and if time permits.

Yes please.

Here you go, jimbob, sorry for the delay. I've tried to provide articles with links, but of course, these are only a tiny selection from the literature, and biased towards my own interests.

Highly structured interviews are better than unstructured interviews (1). Multiple Mini Interviews are better still (where short interviews are held with a number of separate individuals) in terms of predictive validity for the role (2).

Where you have, say, 28,000 applicants for 8,000 places, almost any filter is helpful, but cognitive ability tests (3), prior educational achievement (4), and Situational Judgement Tests (5) have all been shown to have predictive value. SJTs measure something different from prior educational achievements.

If I had to use a personal qualities test, I would try NEO PI-R, especially with regard to the trait of conscientiousness (6), which is generally the single biggest predictor of workplace performance (7).

(1) Levashina, J., Hartwell, C.J., Morgeson, F.P. and Campion, M.A., 2014. The structured employment interview: Narrative and quantitative review of the research literature. Personnel Psychology, 67(1), pp.241-293.
https://d1wqtxts1xzle7.cloudfront.net/4 ... GGSLRBV4ZA

(2) Makransky, G., Havmose, P., Vang, M.L., Andersen, T.E. and Nielsen, T., 2017. The predictive validity of using admissions testing and multiple mini-interviews in undergraduate university admissions. Higher Education Research & Development, 36(5), pp.1003-1016.
(no direct link, but stick title into Google Scholar and you can download the PDF)

(3) Tiffin, P.A., et al., 2016. Predictive validity of the UKCAT for medical school undergraduate performance: a national prospective cohort study. BMC medicine, 14(1), pp.1-19. https://bmcmedicine.biomedcentral.com/a ... MC_TrendMD[/url]

(4) McManus, I.C., Woolf, K., Dacre, J., Paice, E. and Dewberry, C., 2013. The Academic Backbone: longitudinal continuities in educational achievement from secondary school and medical school to MRCP (UK) and the specialist register in UK medical students and doctors. BMC medicine, 11(1), pp.1-27. https://bmcmedicine.biomedcentral.com/a ... 015-11-242

(5) Patterson, F., Zibarras, L. and Ashworth, V., 2016. Situational judgement tests in medical education and training: Research, theory and practice: AMEE Guide No. 100. Medical teacher, 38(1), pp.3-17. https://openaccess.city.ac.uk/id/eprint/12506/1/

(6) Wood, R.E., Beckmann, N., Birney, D.P., Beckmann, J.F., Minbashian, A. and Chau, R., 2019. Situation contingent units of personality at work. Personality and Individual Differences, 136, pp.113-121.https://dro.dur.ac.uk/24452/1/24452.pdf

(7) Arnold J et al. Work Psychology: Understanding Human Behaviour in the Workplace. (No link because it's a textbook)

[nezumi]

Since data science classes were mentioned, let me put a word in for the Weka courses from the University of Waikato. Weka itself is a bit funky but the courses IMHO are very good at grounding you in the things you need to fundamentally understand rather than whizzing straight to clever code. Plus they’re presented by a chummy Kiwi who plays the clarinet.

On the list!

So far I'm half-way through an intro to excel 365 because it's directly applicable to my current role and might lead directly to promotion. I had a brief conversation with my boss' boss' boss last night so I'm working my a**e off - I want to finish the whole course by mid next week so I can approach another manager* for work and support. The company (or at least my bit of it) is on a massive upskilling/career development kick and the boss of my entire division is based in my office so it is a very good time to try and get off the phones!

It would be interesting to look at a big dataset of what people actually get in these personality quizzes, I can see several interesting questions to answer, not least, how accurate are these tests anyway? I guess, not particularly. Here's another interesting question data could answer on this subject; for the people who do these tests to get a particular job (and presumably pass as a particular personality type) how happy are they in the job they end up getting? Sort of a proxy for how well do these tests actually do in reality to find the best staff who will be happy in their jobs and stay for a long time. Not great I suspect

* The one that works mainly with spreadsheets

[sTeamTraen]

The first is that MBTI and friends are bollocks for learning about your personality.

Which ones aren't bollocks though? I know there's a few tests out there that have scientific backing, mostly for specific issues like autism

Maybe we need to distinguish between "psychological tests" in general and "personality tests" in particular. Psychologists have tests for every aspect of human behaviour, and while they do tend to look like each other, not many of them measure what a psychologist who would define as "personality".

I don't think autism is classed as something that is measured by (what most psychologists would call) personality tests; it's kind of orthogonal to the main dimensions of personality. Autistic people are more likely to score low on extraversion, for example, but ranges of personality scores (under the Big-Five or HEXACO model or whatever) are only one aspect of an autistic spectrum diagnosis.

[Allo V Psycho]

It would be interesting to look at a big dataset of what people actually get in these personality quizzes, I can see several interesting questions to answer, not least, how accurate are these tests anyway? I guess, not particularly.

Yes, indeed people do look at big data sets, and the original Skeptoid article mentioned in the OP comments on an informal review. There are many articles: this is an example:
Hughes, D.J. and Batey, M., 2017. Using personality questionnaires for selection.
http://ndl.ethernet.edu.et/bitstream/123456789/29749/1/89.Harold%20W.%20Goldstein.pdf#page=170
and indeed, Myers Briggs doesn't come out well. On the other hand, again 'conscientiousness' does have predictive value as usual.
Something worth adding is that even with statistically significant predictors like prior educational achievement, conscientiousness and situational judgement tests, the effect is quite small: there is a great deal of unexplained variance. It may be defensible to use such tools to reduce 1000 applicants to 100, say, but I personally would hesitate to use them to select one candidate from two.

Here's another interesting question data could answer on this subject; for the people who do these tests to get a particular job (and presumably pass as a particular personality type) how happy are they in the job they end up getting? Sort of a proxy for how well do these tests actually do in reality to find the best staff who will be happy in their jobs and stay for a long time. Not great I suspect

I don't know on this one - any info from scrutes gratefully received!

[Opti]

As a certain kind of Marxist, I refuse to work for any company that would have me as an employee.

I'm quite proud that I was on the Economic League blacklist after I led a couple of strikes at a US cargo airline I worked at for a while at LHR. In support, strangely, of a fairly high management colleague who was being very shabbily treated indeed.

[kerrya1]

(Aside: I do get free free linked in learning from work and it has courses in Data Science so let's see... I must admit, I have starry eyes - it sounds like my kinda gig. I bet the learning curve is brutal though )

Just a little thought, based on what you've said in this thread and elsewhere about work and career goals... KerryA1 and I are both data managers, which might be quite a good fit for the kinds of skills you have and are interested in acquiring, and that won't necessarily need a very specific educational background...

Data manager can mean anything from a highly technical job quality-assuring and tidying/documenting datasets, through to building databases or systems to manage those datasets, and into the almost entirely squishy human work that I do, wrangling the people with the technical skills to pull in the same direction. (My most important professional skill is the ability to translate different dialects of geek.)

Because it's so varied and relatively young as a career, there aren't particularly ossified pathways in. Most jobs largely consist of knotty problems, plus the ability to help scientists work out what they need to do to navigate a system that they don't quite understand but know they are obliged to use.

A few data science classes might be quite a good complement to your call centre skills, plus a smattering of domain knowledge for the field you're going into. (Mind you, I'm a terrestrial ecologist, working with ocean science data from a range of disciplines, so I have had to ask a *lot* of dumb questions.)

Anyway, just a thought...

This is true, I have alerts set-up for Data Science jobs and they are so varied that they really shouldn't be lumped together. If you are interested in the academic sector you'll do best if you have skills in data cleaning and analyses using a variety of methods. A growth area right now is in facilitating the use of sensitive, sometimes unconsented data, through the use of Data Safe Havens, related to this data linkage, anonymisation/deidentification, and secure data transfer skills are also useful. You'll earn less in academia but it'll probably be more varied and interesting/infuriating than most jobs in industry.

[nezumi]

This is true, I have alerts set-up for Data Science jobs and they are so varied that they really shouldn't be lumped together. If you are interested in the academic sector you'll do best if you have skills in data cleaning and analyses using a variety of methods. A growth area right now is in facilitating the use of sensitive, sometimes unconsented data, through the use of Data Safe Havens, related to this data linkage, anonymisation/deidentification, and secure data transfer skills are also useful. You'll earn less in academia but it'll probably be more varied and interesting/infuriating than most jobs in industry.

Definitely gonna be industry - I'm not the academic type

I haven't even thought about deciding on an ultimate path yet but it's definitely going to involve coding in some capacity, maybe a developer of some kind. Backend software development looks quite lucrative and achievable, I certainly have the pre-requisite personality

You're totally right though, there are just so many different roles and specialisms it's going to be a case of learn everything I can and sort of segue my way into an appropriate role in the company I already work for!

[kerrya1]

This is true, I have alerts set-up for Data Science jobs and they are so varied that they really shouldn't be lumped together. If you are interested in the academic sector you'll do best if you have skills in data cleaning and analyses using a variety of methods. A growth area right now is in facilitating the use of sensitive, sometimes unconsented data, through the use of Data Safe Havens, related to this data linkage, anonymisation/deidentification, and secure data transfer skills are also useful. You'll earn less in academia but it'll probably be more varied and interesting/infuriating than most jobs in industry.

Definitely gonna be industry - I'm not the academic type

I haven't even thought about deciding on an ultimate path yet but it's definitely going to involve coding in some capacity, maybe a developer of some kind. Backend software development looks quite lucrative and achievable, I certainly have the pre-requisite personality

You're totally right though, there are just so many different roles and specialisms it's going to be a case of learn everything I can and sort of segue my way into an appropriate role in the company I already work for!

I LOVE academia, and it drives me nuts but the variety and sheer randomness are what keep me interested and give me job satisfaction. Senior academics are, mostly, great fun to deal with and I get opportunities to talk to people in every possible discipline about their research, this week I'll be working on projects in Education, Neuroscience, Psychiatry, Physics, Law, and Divinity.