September 21, 2016

Stop Making Huge Hiring Mistakes – Learn How to Use Pre-Hire Assessments the Right Way

Want help with your hiring? It's easy. Enter your information below, and we'll quickly reach out to discuss your hiring needs.


Last month, we published a list of our top 10 pre-hire personality and behavioral assessments. In that list, we included the EQ-i 2.0, an emotional intelligence assessment that was recommended to us by workplace psychologist Christine Allen, Ph.D., who is also the vice president of Insight Business Works and president of the New York Psychological Association’s Division of Organizational, Consulting, and Work Psychology.

In other words: Allen knows what she’s talking about when it comes to personality and behavioral assessments.

After we posted the article, Allen reached out to us about the DiSC, one of the assessments that made the list.

“I use the DiSC or MBTI after hiring to help figure out how someone will work together with others,” she wrote in an email. “They are not ‘normed’ assessments, so in my view, technically shouldn’t be part of pre-hiring assessment. Just my two cents!”

We thought that was a pretty darn valuable two cents, and we saw an opportunity we couldn’t pass up. We asked the esteemed psychologist if she wouldn’t mind answering some questions about assessments for us – and she thankfully agreed.

Below is a transcript of our email-based Q&A, lightly edited for style and clarity. Can you explain the difference between “normed” assessments and assessments that are not “normed”? Why is this difference important to HR people? How does EEOC compliance fit into the picture (if at all)?

Christine Allen: Normed assessments allow you to compare one candidate’s scores against another’s to determine which person may have more or less of a particular personality trait. Normative scores are typically on a percentage scale from 1 to 100. So if one candidate received a score of 58 on “extraversion” – say, on the WorkPlace Big Five assessment from Centacs – this score means that they were more extraverted than 58 percent of the sample population.

On the other hand, if someone answers a “quadrant-style” assessment (like the MBTI/Meyers-Briggs, DiSC, True Colors, or PeopleMap) and comes out with a very clear preference for one dimension, such as “extraversion” on the MBTI, this simply reflects how this individual sees themselves. Another person could answer the same questions and score similarly and you would not truly be able to compare their scores. For example, it would not show whether you are more or less extraverted than someone else in the population because there is no normed group that your scores are being compared to. This score reflects more or less how we perceive ourselves, which usually has some validity (also called “face validity”). For most of these quadrant-model assessments, there are four basic types with approximately 15-16 combinations; they don’t allow for people who are really more “mid-range” on traits.

swimNormed assessments have “validity scales” that help to detect abnormal answering patterns, such as trying to make a good impression or answering randomly. They must have good test-retest reliability and, most importantly, have been shown to predict job performance to be useful for selection.

While quadrant-model self-assessments can, in my opinion, provide tremendous value for self-discovery, team building, coaching, enhancing communication, and numerous other developmental applications, they are not good choices for selection or hiring due to limited predictive validity, low test-retest reliability, lack of norming and an internal consistency (“lie detector”) measure, etc. I use the PeopleMap and the MBTI frequently for team building and communication training. The Society for Industrial Organizational Psychology (SIOP) and the Society for Human Resource Management (SHRM) recommend against using non-normative assessments for selection and hiring.

Plus, such assessments are not EEOC-compliant, which means that the company is vulnerable to potential lawsuits. A company is more likely to be successfully sued for discrimination if it uses an assessment that is not properly validated or EEOC-compliant.

RC: Can you give some common examples of both normed and non-normed assessments used by HR people?

CA: Some commonly used normed assessments include the Hogan (which has some validation specifically for hiring and selection, i.e., that takes into effect the good impression candidates are trying to make); the WorkPlace Big Five; the CPI260 Coaching Report for Leaders; the EQ-i 2.0; the Caliper; the Harrison Assessments; etc.

Non-normed assessments include the DiSC, the MBTI, True Colors, and the PeopleMap (which I particularly like). This doesn’t mean they have no validity and reliability. For example, these instruments have been found to be fairly reliable (meaning that if you take them now and then again a year from now, the results are likely to be pretty similar) and valid (meaning that their scores have been correlated with other instruments assessing generally the same qualities). And they have “face validity,” meaning that people generally feel that the results accurately describe them. But they are quadrant models emphasizing “type” over “traits.” People have much more complexity than these models allow. Traits are viewed as categorical or bi-modal – e.g., you are either “extraverted” or “introverted” – whereas research shows many people are “ambiverted” – somewhere more in the middle of the distribution.

In quadrant-model assessments, there are only four dimensions of personality, not five. They leave out the very important dimension of “need for stability/adjustment/emotionality,” which generally assesses how reactive we are to stress.

On this dimension, a high or low score is not in and of itself good or bad – e.g., more reactive people are often the “canaries in the coal mine,” which is good in certain roles, especially customer service. You may not want a person who thinks everything is always rosy (lower emotionality or reactivity) in certain positions (think: chief risk officer), whereas in others this may be a great characteristic (e.g., sales rep). It needs to be considered in light of the particular job and in combination with other characteristics.

bikeRC: When should HR people use normed assessments?

CA: In general, it is wise to only used normative assessments for selection and hiring. They allow you to meaningfully compare candidates’ profiles and also to look at candidates in a more nuanced way, rather than in the broad-brush categories of the quadrant-model self-assessments.

For instance, if a candidate is reactive to stress, highly driven and ambitious, and also very humble and reserved, that tells you a lot more than the fact that they are INTJ on the Meyers-Briggs does. It is more complex and nuanced.

RC: Are there differences between the types of assessments in terms of the kind of information they can give you about a candidate?

CA: Yes. There are many assessments that look at the “Big Five” of personality (emotionality or need for stability, extraversion-introversion, originality or openness to new experience, accommodation or adaptability, and conscientiousness/consolidation). To use these assessments, consider who are or have been your top performers (someone you would want to clone!) and then rate the incumbent on these Big Five traits and their sub-traits. The person conducting the candidate assessment would interview the boss and others in the role to determine these “cloneable qualities.” Then, when you assess a candidate with a Big Five instrument, you observe to what degree the candidate may possess the ideal qualities.

Personality matters for work. For instance, you would not want to hire someone for a position of air traffic controller who scores lower on conscientiousness and perhaps who tends to be overly optimistic and positive on the emotionality trait. On the other hand, someone who is less optimistic, less warm and friendly, not very ambitious, and who is not open to new ideas may not be a great fit for a job in sales.

Other kinds of assessments may look more specifically at leadership strengths or emotional intelligence, for example. I tend to use a combination of assessments  – e.g., a Big Five assessment such as the WorkPlace Big Five or the Hogan, a leadership assessment like the CPI260’s Coaching Report for Leaders, and an emotional intelligence assessment such the EQ-i 2.0 or the the Mayer-Salovey-Caruso Emotional Intelligence Test, also known as the MSCEIT. When I use a combination of assessments and a structured interview of the candidate, I can often see patterns of both strengths and potential areas for derailment.

RC: Are there any other common mistakes/misconceptions regarding assessments that you want to address?

CA: Well-executed individual psychological assessments can increase the odds of a good hiring decision by around 40 percent over standard methods. There are still no guarantees of course, but it can be a good investment.

A recent study from University of South Carolina’s Darla Moore School of Business found that the leading cause of executive failure for both internal and external hires is behavioral. For example, the personality of the executive (particularly in terms of selfishness and narcissism) may have been a bad fit for the senior team or organization. Bad hiring decisions can be hugely expensive. The more senior the employee, the more expensive severance, bonuses, golden parachutes, recruitment, interviewing, and onboarding become.

leavesTo highlight the risk, 80 percent of employee turnover is caused by bad hires, according to the Harvard Business Review. Incredibly, studies show that almost 40 percent of new CEOs and senior executives fail within the first 18 months. The vast majority of these failures are not due to a lack of technical competence, expertise, or skill. In fact, David Dotlich and Peter C. Cairo, in their book Why CEOs Fail: The 11 Behaviors That Can Derail Your Climb to the Top and How to Manage Them, argue that most CEO failures have to do with hubris, ego, and a lack of emotional intelligence.

Only about 30 percent of companies use personality assessment or any kind of individual psychological assessment before hiring senior executives, whether the candidates are internal or external. The most valid measures of candidate characteristics, such as personality testing, structured behavioral interviewing, cognitive testing, and emotional intelligence assessments, are used much too infrequently to evaluate candidates for board and C-suite positions.

Another key benefit is that these comprehensive assessments can provide a new executive with detailed, developmental feedback for success in the new position. This feedback can be used to provide onboarding executive coaching, which can help prevent derailment, which often occurs after about a six-month honeymoon phase. So the assessments not only help to prevent bad hires, but also ensure that even great hires learn more about their areas of weakness as leaders, helping to avoid potential problems down the road.

Finally I do think (though not everyone agrees!) that non-normative, quadrant-model assessments such as DiSC, MBTI/Meyers-Briggs, and the PeopleMap are very useful for team building and improving team work, communication, conflict management, etc., on an existing team. These self-assessments – though of course only providing a piece of the picture – can help employees to become less defensive about areas of natural weakness or vulnerability (called the “Achilles’ Heel” in PeopleMap language) and also to recognize differences between people in terms of style, how we like to communicate, etc., which help people to learn to modify their own styles rather than to simply expect that others should be more like them (human nature!).

To paraphrase, Tony Alessandra: “The Golden Rule” says to treat others the way we want to be treated; “the Platinum Rule” says to treat others the way they want to treated!

Read more in Assessment Tools

Matthew Kosinski is the former managing editor of