You Get What You Measure
I have one of those watches that tells me my heart rate and my step count for the day. When I look at the heart rate, I have no doubt that it is counting the pulses of blood in my veins. I feel confident that the watch is measuring what it tells me that it’s measuring. The step count leaves me far less sure. That number on my watch is based on a specific pattern of vibration that occurs when I walk. But I notice that my step count is quite high over long road trips in which I know I’m not walking much. I also occasionally wake up to a step count of 25 (and no I’m not a sleepwalker).
When we try to measure psychological traits, it’s a lot like the step count. My watch tries to detect actual steps, but occasionally it misattributes my movement, causing my step count to be off from the real number. Likewise, we can attribute many actions to multiple traits. Is she crossing her arms because she’s mad? Cold? Nervous? Needs a hug?
This is problematic because many leaders want to understand the psychological makeup of their team members; so how do we ensure that we are measuring what we think we’re measuring? Below are a few tips to keep in mind when you want to create a measure of performance.
Look for Distinct Behaviors
While self-assessments can help us see a person’s thoughts, very often we measure teams by judging an individual’s behavior. As my example of the crossed arms show, there are some behaviors that are hard to attribute to one distinct trait. A person who thinks of a unique solution to a problem has demonstrated Creativity. Creative ideas are not solely the domain of those with high IQ and no amount of determination can guarantee creativity, so this behavior can safely be attributed to Creativity. Behaviors that are distinct to the trait you are measuring are the best measures.
Find Patterns, Not Instances
Sometimes you need to measure traits that have very few distinct behaviors. For those traits, it’s important to identify patterns of behavior or multiple behaviors that point to a specific trait. If a woman crosses her arms, you can’t say that she’s angry and not nervous. But if she crosses her arms, looks down, stumbles over her words, and laughs at comments that aren’t funny, it’s safer to attribute that to nervousness than anger. If your test allows for the measure of multiple behaviors, that can often provide a better judgment.
Avoid Overlapping Traits
If you put your psychological variables into the circles of a Venn diagram, how much would they overlap? Honesty, communication skills, and determination are pretty distinct from one another. Communication skills, teamwork, and agreeableness, on the other hand, have such large degrees of overlap that three evaluators could attribute the same behavior to three different traits. Choose traits that have definitions that are as different from one another as possible and you won’t risk muddying your measurement.
There are many ways to improve the accuracy of measurement when trying to evaluate your teammates. However, accurate measurements are only possible when you select distinct variables and attribute behaviors to the right traits. Starting with these three steps will ensure that your evaluations give you the accurate information you need to make good decisions.
