I’ve been deep diving into all new forms of assessment tooling, primarily for recruiting purposes. I’ve been asked many times to share my experiences and knowledge about them, so I’ve written a White Paper (Link at the end of this post) about it for you to download. Since the white paper is a 22-page read and I’ve learned most recruiters settle for the sound bites, I’ll summarise my findings in this article.
Why modern assessment tooling?
Let me start with the why of my research. The reasons for this are two-fold:
First of all, we seem to select really terribly. How often is someone not doing the job you hired him (or her) for like expected? Very often. In most cases, we say the expectations were ‘totally unrealistic’, but were they really? Usually, they were unrealistic for this person, but probably not in general. So we need a better way to select people and guess what, the problem is that a resume doesn’t really give us the information we need to be able to select well. It doesn’t tell us anything about the quality of the work of the applicant or the circumstances of delivering that quality. And the interview? We only confirm our initial bias based on the CV. So we needed something better, assessments seem to be able to do the trick.
But questionnaires are over the hill. Why? First of all, because the more I’ve researched them, the more I found their scientific base is actually not that strong as many people claim. Second of all because they’ve only been tested and validated in ‘working’ environments, not in applying/recruitment situations. And guess what: people act differently. Research shows this time and time again. Third of all, a question can simply be answered differently if you think you will benefit from them, so you can manipulate them easily.
Games were my first love in this sector and if you look at the stage of development, they have developed the furthest. I’d say they are in adolescence when you look at maturity. Within the games, there are several different types of games. Psychometric (that measure behavior), Cognitive (that measure potential skills, that also result in natural behaviors) and Situational Judgement Games. All are useful in certain situations, depending on what it is you want to measure. Games can measure just about anything, except for ambition and preferred ways of working. Different suppliers have different quality and different strengths and weaknesses. A browser-based game delivers a better candidate experience for example but is relying on a stable internet connection. Modular games can be shorter depending on the job, delivering a better candidate experience, but because they cannot measure certain traits in more settings, scientifically they are less robust. I’m not saying they are not valid, but it’s like the safety of a Volvo versus a Seat. A Seat is perfectly safe, but a Volvo is known for being super strong in a crash.
The great thing about game-based assessments is the fact they are really hard to manipulate. In all the games I’ve played in only a few instances I actually knew what they were measuring. In those cases, you might be able to manipulate a little, like when you have a chance to retry a game. You can derive from that they are measuring your endurance and tenacity and try a few more times than you might have normally. But the potential for manipulation is very limited, so you get honest results.
Linguistic assessments derive your personality from how you write things. What words do you use? How do you frame your sentences? There are about 70 years of scientific research behind this, meaning it’s in childhood phase.
The great things about this technology are that it takes very little time and you can do ask two things at once. You can ask questions you would like to ask anyway and read the actual answer, and meanwhile, you have an algorithm build a personality profile that gives you an unbiased report of the persons Big Five and other personality traits.
The main problem I’ve encountered with this technology is that it doesn’t compensate (yet) for professional writers. I write a lot so my writing style changed over the years to more suit my audience than my personality. Writing with headers, bullet points for example. Things I never used to do. This makes me more structured and the linguistic assessments see me as very structured. That’s a mistake. I also tend to not put myself in the middle as much, making me a little more introverted than I am. Professional writers might include lawmakers, policy advisors, journalists, scientists. So the use is currently limited as far as I’m concerned with high volume jobs.
The solutions may be in either using a chatbot (since in conversations you communicate naturally) or using social media postings.
Social Media assessments are in their infancy, to be honest. I’ve only seen one tool that delivers pretty credible results. But all that actually give in-depth analyses seem to fail time and time again. Even though many seem to use linguistics and adjust by other behaviors, these adjustments are very often man-made. And men is full of bias. To give one specific example, my liking of Linkin Park made me unhappy and discontent with the world according to one test that was transparent about their algorithm. Clearly, there is some form of bias at play here. Many social media tools also ignored all Dutch pages I liked, so they ignored most data because they didn’t classify that yet.
I believe these tests have great potential but need many, many years of research before I would suggest using them.
There is a lot of argument about the fact if facial expressions, mainly micro expressions, have a scientific base. In short, they do and they don’t. There is no peer-reviewed article out there. That’s true. The reason for this is because as soon as it’s submitted people start yelling Nazi-science. This technology has no link to the nazi-science of identifying jews based on looks, but nobody wants to be associated with it. However, there are plenty of scientists that have done really good research into this. And this year an article did get published that we can derive 4 out of the 5 Big Five traits based on eye movement alone.
When it comes to the candidate experience, that’s really great. With a few minutes of video, you can build a profile that you cannot fake. Because you do not control your micro-expressions. You can control expressions (like raising an eyebrow or blinking), but not the microexpression (like in my case my eyebrow vibrating in 0,4 seconds).
What I love about this technology it has the potential to identify traits many other tools cannot identify, mainly in communication styles. I love this tooling when for example you want to assess a team’s diversity of thought. For example, do we have enough action-oriented people or analytical people or human-centric people in a team? Or are we all the same and is that why we keep failing in the same part of the process?
DNA based assessments are just conceived. About 5.000 people have done one. Scary stuff right? Your genes cannot tell us anything about your behavior, so they cannot be used for recruiting. They can tell us things about your potential and future. Because we eventually end up being our most natural selves. Usually, that’s why most of us start looking like our parents. We revert to our natural selves. It is possible to identify if someone might have a talent for something, just not if this talent has been nurtured and developed. Someone with great potential for music that has never played an instrument will not be able to but will learn much quicker.
Not useful for recruitment. Or at least not for identifying current qualities or traits. If you’re hiring for the future it does have great potential. Since we all tend to revert back to our most natural selves, hiring a 40-year-old with a DNA profile you know the behavior he or she will be displaying in a decade or so. Scary as hell, especially if you think about Crispr, a way to edit our genes and turn them on or off. Really scary, but keep an eye on in the future.
Like DNA it’s only just conceived. Great potential. Unable to fake. Could be used in highly sensitive positions. It’s possible to see one’s brain react to certain stimuli and know if someone poses a risk under certain circumstances. For example, all major swindlers have certain area’s of the brain lighting up under certain stimuli. It doesn’t mean you will steal from your company if that happens, it does mean you run a much bigger risk that people who do not have this trait. Maybe not make them CFO of the company?
Would you like to read more about a specific assessment technology, let me know in the comments and I’ll write a more in-depth piece about that technology?
If this has your interest: this is just a very short summary of the White Paper I’ve written on the subject. Download it for free here: https://rdaily.co/2KP6YZz