Why Cambridge Analytica’s techniques could kill gamification


Thanks Alex! That is a really good and important point about distinguishing candidate experience from the value added. Also about Cambridge Analytica, the importance of informed consent cannot be ...

Read More Lara Montefiori
Add a comment

Could gamification turn out to be an over-hyped, excessively complicated flash in the pan?

The recent scandal surrounding Facebook and data collection by Cambridge Analytica has brought attention to the scale of data that is automatically collected and used online to build psychological profiles. Much of this is done automatically and without the users’ explicit knowledge.

HR departments must always balance the interests of people with business and budgetary limitations. Psychometric testing is a valuable resource for any HR department’s core activities, but the gamification of psychometrics has three major barriers that greatly limit its practicality and utility.

Barriers to starting

The reality of using psychometrics to test people in the workplace means the most cost- and time-efficient option will likely be prioritised. Psychometrics always take some time to explain. Participants need to understand what they are doing and why. Everyone needs to know how their data will be used, who will have access to it, and in many cases participants need to be debriefed.

Adding additional layers of complexity with a mobile game is likely to make psychometric testing more difficult, not more efficient.

The automatic and unobtrusive alternatives to gamification, brought to recent public attention by Cambridge Analytica, show how there are far more effective and practical ways to adapt psychometrics. These automatic methods make gamification look like a clunky option that may quickly become a relic.

Barriers to completion

Another significant barrier to using psychometrics is that adding unnecessary complexity reduces participant completion rates. The less effort on behalf of the participant, the more participants are likely to finish.

It can be challenging enough to convince people to participate in a math test. Is gamification really going to encourage busy, working adults to test their math skills by installing an app on their phone that puts mathematical problems on a pirate ship or in an Aztec temple?

If the same task can be done simply by clicking ‘I agree to share my data’ and have an algorithm do all the work, the simplest option is typically the most effective.

Barriers to long-term viability

Gamfiying psychometrics ultimately risks being an amusing distraction that fails to imbue mobile psychometric applications with the intrinsic or extrinsic motivators in other applications which are commercially successful. Social media companies like Facebook, Instagram and Snapchat have already found the magic dust to sprinkle over their apps that keep people coming back and seeking online social rewards. Video game companies have developed games which keep people paying and returning for hundreds or thousands of hours.

Gamification in psychometrics seems to be attempting to straddle the science and the technology, while missing the boat on both counts. Gamification seems to be an interesting idea that is likely an artefact of the time, but not necessarily the best use of either the science or the technology.

The verdict

When people complete psychometrics, they tend to opt for the fastest option. And when people want to play games, they opt for the most fun option. Gamification does not seem to hit the mark as either an efficient psychometric or an enjoyable game.

Social media companies seem to have hit the nail on the head when it comes to gamification; they have made virtual social interaction an addictive online platform. The recent revelations from the Facebook and Cambridge Analytica scandal shows how quickly and efficiently psychometric profiles can be compiled, while the same is not true for gamification.

Of course, issues of consent must be addressed openly, stated clearly and meet ethical standards for an HR department. And when they are, most people will choose the automatic option and click, ‘I agree to share my data.’

Ian MacRae is co-author of High Potential: How to spot, manage and develop talented people at work


Interesting perspective, thank you Ian. I would agree that simply gamifying an existing process or test does little to improve the experience for the candidate or the organisation - there has to be more value to it. At pymetrics we are often put into the "gamification" bucket, but for us the games have always been a side note - simply the most effective way of getting the behavioural data that we need in order to build and use our AI models. This then allows for a standardised, objective, unbiased and engaging way to assess people rather than relying on either (a) psychometrics that can be biased and/or boring, or (b) collecting random data points from multiple online sources that were never designed for assessing people. A significant lesson that we should learn from the Cambridge Analytica story is that we should use data for the purpose that was intended when consent was given. Alex Cresswell, MD EMEA, pymetrics


There's a whole problem with the definition of 'game' and the gamification movement as a whole. I think this article is a great representation of the band wagon that everyone is jumping on. Gamification shouldn't be the flag bearer for behavioural and motivational psychology, or techniques that are used to affect some type of behavioural change. It's a trending term that's used to encapsulate a whole lot of different techniques that existed before 'gamification' became a thing. To be completely honest I believe what this article is touching on is an issue of privacy and consent. I'm not sure how CA used gamification or if at all. If I consent to being tested and the tools are valid and reliable then it shouldn't be a problem. I may not like the tool, or wish it was different, but if it works in what it does and is defensible then it should be fine. On the other hand, if employers are looking at my smart fridge data to figure out how much sugar intake I have, what kind of foods I like to eat, how long I hold the fridge door open etc. to predict my fit with their lunch menu without me knowing or consenting to it then it becomes a problem. I mean that's a silly example but it seems to me that we're at that point of no return where somehow all data seems to be used in some way, and then sold to third parties to help them target me. Not sure how gamification fits into this at all. ps. Also agree with Alex. With game based tests being different to gamification, a process. The term game based is perhaps a misnomer as games are often a single part of an incredible engineering, analytic and psychometric effort to build essentially better assessments.


You will be pleased to read the papers we have written about some of the points you made but, while those are being peer-reviewed, let me set some records straight on whether the above-mentioned are indeed barriers that limit my assessment's practicality and utility. In order of appearance to keep it simple... 1)"Barriers to starting": there is minimal need to explaining to people how to approach game-based assessment because the tasks are highly intuitive and rely very little on language. This actually removes (not adds) levels of difficulty and, yes, it does make the assessment more efficient. Quick reality check: people must be informed of how their data is used , how to access it and all that - it is their right, and, perhaps surprisingly, people do value that degree of control quite a lot. 2)"Barriers to completion": our clients have seen a steep increase of completion rates when they introduced our assessment, it is safe to say that drop-out rates are virtually non-existing, and yes, the 15-odd-thousands live candidates who completed our post-assessment surveys do believe that pirates are pretty cool. I am not quite so sure how many people would be happy to click on 'I agree to share my data' ESPECIALLY after the Cambridge Analytica tomfoolery, but I like to base my opinion on fact so I am open to see how that goes down. 3)"Barriers to long-term viability": we don't miss the boat (or pirate ship) on both counts, we actually make the best of both worlds. The use of game technology reduces the tedium of traditional psychometrics without having to make the experience attractive enough to people to cause the level of obsession that makes someone return to a game for thousands of hours. The motivation for an assessment, no matter how you look at it, will always be the job at stake. 4)"The Verdict": in my experience with candidates, validity is valued massively over speed, and, in my experience with gamers, fun is not necessarily the only or the strongest selling point of a game. Rather than hitting the nail on the head, I would refer to a last nail in the coffin, and good luck climbing up the grave Cambridge Analytica dug for social media scraping. I would normally close this comment thanking you for the mention, though, to the best of my knowledge, no Aztec ever dwelled on Easter Island (which is where my totems are from...) , and I would never want to steal somebody else's spotlight ;)


Thanks Alex! That is a really good and important point about distinguishing candidate experience from the value added. Also about Cambridge Analytica, the importance of informed consent cannot be stressed enough.


Laura, I would be very interested to read any papers on this topic that are published in the future. In the meantime, a more detailed and thoroughly referenced discussion of the topic is in the High Potential (MacRae & Furnham, 2018) book. The point is there is yet to be any good scientific evidence demonstrating incremental validity over other techniques - therefore in the interest of parsimony it has not been demonstrated they are a superior or necessary alternative. It looks like this has sparked some good discussion on the topic, and I really look forward to seeing future research done on this topic!

Change the CAPTCHA codeSpeak the CAPTCHA code

All comments are moderated and may take a while to appear.