Are job interviews such a necessary evil?
They might not be as useful as you think
By Jack Fereday
Clammy hands, trumped up resumes, fake smiles… Job interviews aren’t just a bane for employers and candidates alike, they’re also riddled with errors. Perhaps it’s time we drew the curtain on this painful tradition.
In the history of labour, job interviews showed up late. For a long time, trades were passed on from father to son, or from master to apprentice, until the appearance of the first selection processes at the dawn of the industrial era. These were initially spontaneous — underperforming workers were swiftly replaced –, then formalised in 1921, when Thomas Edison introduced questionnaires to test the abilities of potential recruits. But nearly a century later, the job interview remains as unpopular as ever. The increasingly competitive job market leaves us little choice but to comply; but still, no-one wants to reveal their life story to a perfect stranger… To make matters worse, this odd convention might also be counter-productive: social psychology researchers have highlighted a number of biases that undermine its efficiency. So should the job interview finally be scrapped? And if so, how else can a company find the people it needs?
Choosing the right tie
‘Your choice of tie will likely influence the outcome. The problem is that this information is as irrelevant as it is influential’
“It’s the first impression that counts” says the nervous candidate as she puts on her freshly ironed power suit. “I can trust my first impression” says the employer as she marches down to the interview room. Unfortunately for them both, the first belief is as true as the second is false — and so their short exchange gets off to a shaky start, splutters in a series of misunderstandings, and ends in yet another missed opportunity.
A year before Thomas Edison implemented his first ability tests, the psychologist Edward Thorndike discovered what he called the “halo effect” — a potent tendency to make all kinds of unrelated assumptions about someone based on the first (and often flimsy) information we receive. Let’s say your new colleague happens to look like that actor that plays James Bond. Within minutes of meeting him, you also start to find him charming and witty… From now on, it’s as if you’re blinded by a radiant halo: when you look at him, you only see what you want to see.
Physical attractiveness might be the most blatant source of cognitive bias; but it’s not the only one. An array of more subtle factors are also at play, like those “subsidiary characteristics” analysed by Pierre Bourdieu, which help us identify a person’s social origin and “may function, in the form of tacit requirements, as real principles of selection or exclusion without ever being formally stated” (Distinction. A Social Critique of the Judgement of Taste, 1979). Such innocuous details as a person’s taste in clothes or music make up what the French sociologist called “cultural capital”, which in turn contributes towards their “economic capital”. So your choice of tie, just like the sport you choose to play on a weekend, will likely influence the outcome of the interview. The problem is that this information is as irrelevant as it is influential.
These flaws haven’t gone unnoticed by specialists: “The problem with interviews is worse than irrelevance: they can be harmful, undercutting the impact of other, more valuable information about interviewees”, says Jason Dana, Assistant Professor at Yale School of Management. “Interviewers typically form strong but unwarranted impressions about interviewees, often revealing more about themselves than the candidates” (“The Utter Uselessness of Job Interviews”, New York Times, 8/04/2017)
Where human judgement proves to be flawed, might Artificial Intelligence (AI) be the answer? Psychology is no longer a human prerogative: AI bots can now decipher and evaluate our facial expressions, tone of voice, word choice, and speech patterns. Unilever claims one such program has already led to an increase in ethnic and socioeconomic diversity: after using AI to screen all entry-level applicants for over a year, they say the number of universities they recruited from has risen from 840 to 2,600.
‘Bots can only evaluate behaviour using criteria submitted by humans, thus reproducing their bias’
But computers aren’t infallible. First of all, their widespread might one day influence our behaviour. If you know your speech is being recorded, converted into text, and analysed by an AI bot, you’ll be sure to use a certain number of key words. Similarly, you might want to simplify or exaggerate your facial expressions to conform to whatever input images it’s using as reference models — subtle personality traits, like the use of irony or a dry sense of humour, will only work against you… Then there are the limitations of machine learning: even when the bot interprets a person’s behaviour correctly, it can only evaluate it using criteria submitted by humans, thus reproducing their bias. Amazon learned this the hard way, when its AI recruitment program screened thousands of resumes… only to discriminate against female candidates. According to its engineers, the machine’s algorithms had learned from the vocabulary found in hundreds of past CVs, most of which belonged to men.
The golden age of psychometrics
To compensate for these shortcomings, AI can be used alongside a battery of other psychological tools, the most sophisticated of which are known as “psychometric tests”. These are supposed to identify all the main characteristics of an individual: their main personality traits, motivation, cognitive capacity, emotional intelligence, etc. “The goal of our interview process is to predict how candidates will perform once they join the team”, says Google’s former Vice President of Human Operations, Laszlo Bock. “We achieve that goal by doing what the science says: combining behavioral and situational structured interviews with assessments of cognitive ability, conscientiousness, and leadership.” (Work Rules!, 2015). To help their recruiters, he says Google has developed an internal tool called qDroid, which automatically generates a series of questions for each candidate, based on the job they’re applying for. Each one is then given a series of scores, which allows for an objective comparison.
‘The psychometric approach is the latest expression of a scientist legacy’
The tech giant isn’t the only one turning to science to improve their hiring process: most big American firms are now using such tests, as well as 70% of all British companies, according to Cambridge University’s Psychometrics Centre, created in the early 2000s. But the rise of psychometrics hasn’t been without controversy. The British public will remember the notorious case of Rev Paul Flowers, whose sudden rise to the head of Co-op Bank ended in a series of scandals involving sex, drugs, and gross incompetence. Members of the parliamentary committee erupted in laughter when they were later told that Flowers “had been picked as chairman because he had done well in psychometric tests”, despite his lack of experience in banking. The story illustrates the point made by a number of critics who claim such tests lack reliability, especially when it comes to vetting people for important positions. Others stress people’s reticence to submit to what they see as an invasive procedure, as well as their tendency to distort the results by faking the answers.
The legacy of scientism
But is it even possible to measure — never mind predict — a person’s character? Are we so easily penetrated by the tools of science? Over this important philosophical question, two distinct visions of man are at odds: the first was inspired by Charles Darwin’s work in biology in the Victorian era, and sees individuals as objects to be isolated, analysed, and classified, like any other object in nature. The psychometric approach — like phrenology (the study of a person’s character based on the shape of their skull) before it — is the latest expression of this scientist ideology.
Meanwhile, continental philosophy proposes a very different vision. In Being and Time (1927), the German philosopher Martin Heidegger deconstructed the modern conception of man as a distinct subject, which had been the dominant paradigm in Europe since René Descartes. Following Heidegger’s analysis, human experience is fundamentally structured by our relationship to the world around us — that’s to say, a direct relationship to objects, as well as other people. “By ‘Others’ we do not mean everyone else but me — those over against whom the ‘I’ stands out”, he explains. “They are rather those from whom, for the most part, one does not distinguish oneself — those among whom one is too.” As a science, psychology traditionally tends to isolate individuals in order to better know them better. But in our “everydayness”, Heidegger explains, man is almost completely absorbed in his interaction with others, with whom he identifies, and whose influence is part and parcel of his identity. Therefore, attempts at any a priori definition of a person are vain — only a context-based approach can unlock the secrets to their behaviour.
The pragmatic approach
‘We should stop focusing on people’s individual characteristics, and instead look at the way they interact’
If we were distinct subjects capable of introspection, as Descartes believed, interviewing us would be a doddle; but as embodied beings caught up in an environment, we’re tougher nuts to crack: our behaviour changes according to the situation and people we’re surrounded by. That’s why we should leave aside man’s “personality” — if such a thing exists! — and focus more directly on his interaction with others.
This is precisely what a team of researchers did in Palo Alto, California, in the 1960s. Led by the Austrian-born philosopher and psychologist Paul Watzlawick, the “Palo Alto School” sought to highlight the pragmatic effect of communication on human behaviour. In a “complementary” relationships, for instance, “dissimilar but fitted behaviors evoke each other. […] Each behaves in a manner which presupposes, while at the same time providing reasons for, the behavior of the other. Their definitions of the relationship fit” (Pragmatics of Human Communication, 1967). That explains why we can be quite dominant in our relationship with one person, whilst being completely submissive towards another. Transposed to the workplace, this paradigm shift means we should stop focusing on people’s individual characteristics, and instead look at the way they interact.
Matt Mullenweg, creator of WordPress and founding CEO of Automattic, came to this conclusion after wondering why so many hires didn’t work out: “As we considered the situation, it became clear that we were being influenced by aspects of an interview — such as someone’s manner of speaking or behavior in a restaurant — that have no bearing on how a candidate will actually perform”, he recalls in the Harvard Business Review. After selecting candidates from a batch of resumes, the company now offers them a paid trial: for periods of three to eight weeks, they get to work alongside their potential colleagues. “The more we thought about why some hires succeeded and some didn’t, the more we recognised that there is no substitute for working alongside someone in the trenches”, says Mullenweg. “Candidates do real tasks alongside the people they would actually be working with if they had the job.” This allows him to gauge their “soft skills”, which are crucial in the modern workplace, but also the way individuals interact within the company’s distinct work environment.
Mixing things up
‘A person is always more than the sum of their emotions, intelligence, or past’
This pragmatic approach seems to work: of the 101 people Automattic hired in 2013, only two didn’t work out. What if paid trials are two expensive or risky for the company? The idea is to assess people in a way that puts them in work-like situations. “Maybe it’s a trial; maybe it’s a presentation”, says Mullenweg. “Maybe it’s a short-term assignment that can be done in an afternoon”. Or, why not, a game: European-based companies such as Nationwide, PwC and Lidl are now using to escape games to find the right people. After all, if “to be playful and serious at the same time is possible, and defines the ideal mental condition” according to John Dewey, that’s because the player is fully engaged in the activity at hand. To succeed at an escape game, participants need to think but also communicate, thus revealing several key aspects of their personality (as long as they play the game, which isn’t always guaranteed).
Does that mean job interviews will soon be a thing of the past? Unfortunately, that would be jumping to conclusions: none of the above-mentioned companies have scrapped the practice entirely. Like Unilever, who uses AI to filter through resumes, most of them they still prefer to finish off the hiring process with an interview (or an anonymous exchange of texts in the case of Automattic). But when used in combination with these other tools, the interview serves more to help confirm their choice rather than to guide it. After all, talking face to face allows employers to see if the candidate can conform to rudimentary formalities, which can be crucial in certain customer-oriented prositions. It also allows them to express any queries they might have about the candidate’s motivation. For the latter, the interview can be an opportunity to find out more about the organisation they’re seeking to join, and make sure it’s compatible with their own projects — in which case the interview ceases to be an awkward interrogation to become a more fruitful exchange.
At the end of the day, every hiring process has a blind angle. But if no single method is enough, that’s due to the ever-fleeting nature of its object — and that’s a good thing. For if no game, test, or resume can grasp the multiple dimensions of an individual person, that’s because a person is always more than the sum of their emotions, intelligence, or past. If you can accept and perhaps even appreciate the unpredictable nature of the game, then perhaps the traditional interview can still play a part — but a small one at that, preferably towards the end… and please, keep it short.
- Take first impressions with a pinch of salt: we’re more biased than we think!
- The best way to assess a candidate is to see how they perform within a specific work environment
- Mix things up a little: interviews can be useful when combined with more hands-on exercises, such as trials and games
Originally published on Philonomist