Online assessment tools are critical for high-volume hiring but can be disastrous

For decades, many employers used assessment tools much like they used reference calls: only for the finalists and with about as much scientific validity as your horoscope. Fortunately, that’s all changing, changing rapidly, and changing for the better.

According to Dr. John Sullivan, corporate recruiting leaders must proactively discourage using candidate assessment tools that don’t meet the following four minimum requirements for use:

• It predicts performance (validation) – those who score high on the assessment test/tool will perform above average on the job if they get it. 
• It is job-related – the tool measures a knowledge, skill, or other factor required to do the job.
• It is reliable – the tool produces consistent results among assessors. And it produces the same results if the tool is used again on the same candidate. 
• Everything is measurable – each of the assessed factors can be clearly defined, and each one can be accurately measured. 

If you follow those four requirements, you’ll avoid the five worst assessment tools:

1. Myers-Briggs (MBTI) personality test – Although many still use this as part of their candidate assessment process, according to Sullivan, it is inappropriate for hiring. The company that owns the test specifically warns those who purchase it against using it for hiring and selection. The primary flaw of this “worst” tool is that it doesn’t accurately predict job success. 
2. Job fit – Perhaps the most widely used of all candidate screening tools, it also has, according to Sullivan, the highest number of flaws. You need to identify the factors such as values that are essential for job/team success, make those easy to identify and measure by untrained interviewers, and ignore the fact that many new hires succeed without being a fit as they learn how to adapt and change their behaviors and attitudes.
3. Body language – Different interviewers interpret the same body language differently, and interviewees with different backgrounds will use different body language when experiencing the same emotions. Even if you could control for those variables, there is no data to prove that a particular body posture or movement predicts future job performance.
4. Emotional Intelligence (EQ) – Very similar to body language in that there’s no consensus about what to measure or how to measure it and, even if you could, there’s no data to prove that whatever someone may regard as a high EQ predicts future job performance.
5. Brainteaser / puzzle questions – Popularized by Google, they’ve now been abandoned by Google as a “complete waste of time”. According to Sullivan, “You will also likely find that the tool has a serious reliability problem because everyone does not agree on the correct answer before the question is asked. And that, unfortunately, means that the same answers, provided by the same candidate, to the same puzzle will almost always get completely different scores from each of the assessors of the question. Finally, the use of these questions may be discriminatory against diverse candidates who might not have a strong math education. So, among all of the five ‘worst’ assessment tools, this one has the most data that supports completely avoiding it.”