December 14, 2018
Presenting results on the effectiveness of pre-hire assessments and selection tools is not the most enthralling. Unless you’re a scientist or statistician, there is nothing alluring, stimulating, or exciting about a validity coefficient. Generally, it’s not worth trying to explain the pure excitement of finding a moderate correlation (such as, r = .25) between your pre-hire assessment and job performance at a significance level of p < .001! Just imagine the reaction of business leaders as you present these results: What does “.25” mean, anyway? Is that like 25%? That doesn’t seem impressive at all! And who can possibly be excited about something described as “moderate?” I had a moderate pizza last night and read a moderately good book, neither of which were all that great in the end.
In the hiring and selection context, validity coefficients are often used to represent how well an assessment predicts performance on the job. However, these coefficients (e.g., “r = .25”) don’t speak to everyone. Their meaning can be confusing and inaccessible to an audience that doesn’t have a working knowledge of how to interpret them. Even if you do explain the meaning behind this number, it can still be challenging for non-technical audiences to fully appreciate its importance. Rather than providing a research stats lecture each time you share validity results, why not present those results in a memorable fashion that requires less explanation?
Hack #1: Put it into context
One approach is to present correlation coefficients in comparison to more relatable examples. In a recent validation study of a client’s pre-hire assessment, we supplied clear frames of reference for the audience, including the relationship between NSAIDs (e.g., ibuprofen) on pain reduction (r = .14) and the relationship between weight and height in US adults (r = .44). These examples provided the context we needed to showcase our validity coefficient (r = .41) as pretty awesome! It was easier for the business to understand the strength of the relationship, and certainly more memorable.
Hack #2: Embrace the odds
A second approach is to utilize odds ratios instead of validity coefficients, as these can be easier to conceptualize. An odds ratio represents the odds that an outcome will occur, compared to the odds that it will not occur given a certain condition. For example, you may find that there is a strong correlation between call center applicants’ scores on a pre-hire employee assessment and the number of first-call resolutions they complete. The odds ratio might be expressed like this: “Top assessment scorers were 3 times more likely to obtain 95% first-call resolutions than bottom assessment scorers.” The benefit to using odds ratios as compared to validity coefficients is that you don’t need statistical software to calculate the odds, and they can be easier for non-technical audiences to understand.
Hack #3: Add it up
Cost and time savings are perhaps two of the most compelling reasons for action. You can take validity coefficients and odds ratios one step further by assigning them tangible values. For example, a reduction in turnover translates into time (and cost) savings related to processing terms and replacement costs (i.e., engaging in a new cycle of recruiting, onboarding, and training). Increases in sales performance or customer retention translate into increased revenue and cost savings for the organization. Your findings are more meaningful to key stakeholders when you cash out results in terms of the hours and dollars saved.
With the rise in popularity of infographics, memes, GIFs, Instagram, and the like, let’s face it – life is just more visual than it used to be. In addition, we know from cognitive psychology that images or pictures are more memorable than words on a page. Simply, we pay more attention to them (i.e., spend more time processing them). Thus, it makes sense to use eye-catching visuals to the greatest extent possible when showing results. Sure, bullet points can convey the same information and are easier to jot down, but they may not be recalled later. In contrast, using graphs, images, and pictures to tell a data story increases the likelihood that the audience will later recall your study’s results.
A lot of brain power (and coffee!) is expended in collecting, cleaning, and analyzing data. It would be a shame for a significant finding between results of employee assessments and success on the job to be undervalued or unrecognized. Maximize your data scientists’ efforts by investing the time to make those numbers more palatable, approachable, relatable, visual, and just plain fun.