with No Comments

Post No.: 0543interview


Furrywisepuppy says:


Since most people are biased in thinking they’re great judges of character and honesty – job interviewers think that if they just spend a few minutes with someone, they can predict exactly what sort of employee they’ll be. This is despite lots of data highlighting that someone’s performance in a traditional job interview is on average entirely non-predictive of their performance on the job if they get it. (People aren’t even always good judges of themselves e.g. judging their own levels of inebriation, dehydration, gluttony, hypoxia or sleep deprivation!)


Job candidates are often subjected to psychometric tests, but these are poor predictors of future job performance too. The problem with typical personality or psychometric tests is that they’re self-reported i.e. people fill in the questionnaires with their own opinions of themselves thus there’s room for bias (e.g. people can misjudge themselves, their answers can depend on how they felt that particular day, or they’ll give responses that they think the employer is looking for). Something being systematised, computerised or quantified doesn’t automatically make it objective. These tests are also often utilised without employers considering whether it actually matters for the job whether someone is, say, introverted or extroverted?


The best tests are those that involve doing tasks that are exactly like what the candidates are expected to do in the job they’re applying for rather than are proxy tests, but this understandably isn’t always practical. But when it’s difficult to answer whether someone will be suitable – interviewers will resort to answering ‘substitution’ questions, such as judging their tattoos, the way they dress or any minor spelling errors on their application forms – which are all usually quicker and easier judgements but based on shallow and often irrelevant cues. Unfortunately, shallow is how much of the human world works. (Read Post No.: 0277 for more about attribute substitution.)


Substitution is such an intuitive and everyday heuristic that people don’t realise that they’re using it and don’t tend to think that it’s prone to errors of judgement. The way someone dresses might indicate the amount of effort they’ve put into preparing for the job interview – but is answering the question ‘how much effort did he/she put into his/her outfit today?’ truly the same thing as answering the question ‘how much effort will he/she put into his/her work if he/she got the job?’ Some people spend way too much time looking in the mirror every day that they logically cannot be spending that much time on other people or on, indeed, their work! (Now what one is wearing can make one feel different – so wearing a suit as an interviewee can make one stand or sit more upright, which feeds back into making one feel more confident. But this is about the interviewer judging by appearances. This applies to criminal court contexts too – is someone less guilty just because they can afford a nice suit?)


Such heuristics are used because interviewers are trying to predict a future performance in some job based on a present performance in some interview, but the predictive success of such heuristics aren’t reliable. It happens in business (e.g. judging an entrepreneur’s handshake to somehow judge his/her business acumen), in politics (e.g. judging a candidate’s age to somehow judge his/her leadership skills), in dating (e.g. judging a man’s height to judge his ability to protect or judging a woman’s dress length to see if she’s ‘asking for it’(!)) and so on.


Both job interviewing and dating are particularly notorious when it comes to using the substitution heuristic. Can we truly reliably determine enough of someone’s character by, say, what car they aspire to own? This also incentivises people to concentrate on their superficial appearances because that’s what they’re being judged on (e.g. looking like one is working is more important than actually getting any work done!)


The unreliability is also partly down to confirmation biases after making stereotyped judgements upon seeing someone’s CV/résumé, which can lead to biased questions that’ll confirm one’s existing beliefs, like ‘are you a strong leader?’ (which already implies ‘strength’ and ‘leadership’), rather than neutral, non-leading questions. It’s therefore better to ask every single candidate the exact same set of worded questions and in the same tone and order in an interview, and only asking these questions.


People are usually subconsciously influenced by people’s names on job application or university admission forms, due to implicit or explicit racist or other prejudices. This is why it’s preferable to judge these applications without knowing the applicants’ names (or addresses).


Interviewers should really be ultimately judging the content of CVs/résumés and business plans, not the paper thickness or print quality they’re presented on. (It’s increasingly electronic nowadays but not completely.) In just about any context – don’t judge the presentation over the content. In these domains where people frequently judge superficial proxy attributes as substitutions for the questions they actually (or ought to actually) care about answering, there are usually lots of hit-and-miss errors of judgement – yet people will typically continue to lazily rely on them and won’t even learn from any judgement errors made from personal experience. Too many take it for granted that ‘of course you judge the paper quality’ when they need to take a step back and question ‘is the paper quality really telling me much about the person?’ and perhaps ‘is using a thicker paper when a thinner one would’ve sufficed good for the environment anyway?!’


Knowing how somebody performed in similar situations in the past is the best predictor of their future behaviour. So it’s useful to use past behaviour trends to predict future behaviour trends if the measure is standardised and directly comparable – such as how they pay their bills on time, their grades over time and other things that are hard to fake or fluke unlike one-off events like an interview or a one-off exam. Look for trends over the long-term. (That’s why you shouldn’t marry someone too soon after first dating them!)


Someone who interviewed well but has no relevant track record will be a more uncertain candidate than someone who interviewed poorly but has an excellent relevant track record – the former may have left the best personal impression on you (‘what you see is all there is’) but the latter will more likely be the better candidate. Also, with scant information about the first person’s past, the problem of the law of small numbers enters too i.e. his/her great interview was more likely down to a one-off fluke, hence one must account for the regression to the mean i.e. someone who performs above average one time is more likely to perform below average the next. (This is just like a large number of independent product reviews are more reliable than a small number of reviews. A second and third independent opinion could be useful before making any major decision.) One might still feel an urge to favour the first person though – following our own personal intuitions feels more natural and pleasant than acting against them and following statistical logic. Being rational thus isn’t easy or natural for us to do.


The perennial problem with relying on the historical track records of candidates however is how do the inexperienced get a chance? Absent of a comparable historical job experience – proxies like their punctuality or attitude towards continued education (‘upskilling’) are better than superficial proxies, or totally overlooking them.


Overall, to argue that the reason why a candidate failed to secure a job is because of his/her interview skills – rather than his/her relevant intelligence, talents and work ethic – points to a fundamental flaw in the process! This questions the appropriateness of traditional interviews in many contexts where they’re relied upon. Some people can be poor at interviewing or spelling but excellent at those very jobs they’re applying for – perhaps because they’re introverts, have dyslexia or simply aren’t very good at making up lies on the spot to get through an unexpected, curveball interview question about a hypothetical scenario! Lies can of course also be prepared for the expected interview questions. Candidates will say things that they think the interviewer will want to hear, even if they’ll need to cherry-pick, exaggerate or fabricate stories. Are we seeking the best verbal liar, manipulator, poseur or fuzzy blowhard? These traits might be important for the job post at stake, or they might not. The typical interview process can particularly fail for autistic people – for the types of jobs they apply for, they can probably do the job but just can’t pass the interview stage.


Interviewers will obviously be checking to see if the candidates meet the minimum criteria they’re looking for but, unless they like wasting people’s time, this should already have been done at the application stage and a candidate shouldn’t have made the interview stage if they didn’t meet that criteria. Therefore the rest is relatively subjective and about uncertain predictions of the future.


Yet it’s unconventional for interviewers to get reviewed over how good their predictions were, based on how those they passed through have subsequently performed in their roles! The employees tend to get solely blamed if they do poorly rather than those who gave them their employment when they should’ve realised they were unsuitable for the roles! It really goes right to the top because of who gave the interviewers the role of interviewing job applicants for the company(!) (It’s like although the players should shoulder some responsibilities – the manager of a professional sports club, or really the board, should get sacked too if the wrong managers or players are hired and the team performs badly.) The buck always stops at the top and scapegoats should never be made out of anyone below. Woof!


Many job interviewers aren’t trained for interviewing, thus may treat interviews as more like social gatherings to see if they personally like the candidates rather than if their skills, experience and temperaments are suitable for the roles they’ll be asked to perform. Candidates may be asked questions and asked to do tests and exercises that the interviewers think are ‘ingenious’ but are irrelevant or are cod psychology (e.g. ‘if you were an animal, what species would you be?’) Interviewees may consequently think the company is unprofessional and isn’t right for them.


Interviewers should therefore be highly trained at interviewing. An untrained or badly trained interviewer is more likely to rely on prejudices, fall for superficial charm and overweight first impressions (when someone displaying some nervousness isn’t necessarily a bad thing because it could indicate that they genuinely want the job). Most people, in any context, think they’ll be able to smell it when anyone tries to lie or wing it – but logically they’ll not realise how much has gotten passed them precisely because it has gotten passed them.


There could also potentially be a situation where interviewers don’t want to hire someone who’s ‘too good’, hence a conflict of interest. Knowledge above and beyond requirements might be seen as a threat to interviewers because they don’t want to hire someone who might take or overtake their position in the organisation one day, make them seem unintelligent in comparison, or make them potentially subordinate to the new recruit one day.


Woof! In conclusion, unless a job happens to involve verbally answering questions like ‘name an example of when you worked well with others?’ – it’s better to test someone actually doing that job, or similar, in a team, if you want to see how they’ll perform in that job. Test their actions more than their words. Auditioning candidates with relevant simulations of the types of things they’ll be expected to do, or trial periods if possible, are better approaches. Relatively modern interview methods like hackathons and even virtual or augmented reality assessments might prove promising. Testing candidates on the job also helps them to understand what they’re exactly applying for to see if the role is right for them too.


Comment on this post by replying to this tweet:


Share this post