Understanding R-square value in compensation analysis

Understanding R-square value in compensation analysis

For this episode of Sunday Skills, I wanted to explain some statistical concepts used by many compensation consultants when they produce compensation analyses for their clients.Very often, the results are presented “raw”, with a nice graph and a R2 or R-square value (or a series of values) mentioned on the side.

But many HR Generalists, and even Compensation specialists, have not studied statistics, and as a result they don’t always understand what that really means for their business. So here is a plain English, simple explanation of what the data means. Please bear in mind that I am not a statistician, so if you spot a mistake, please let me know through the comments or the Contact Me page and I will do my best to correct…

Now, let’s imagine that you are interested in understanding how tenure, education and job grade influence pay.  So you perform a regression analysis, which allows you to “predict” the strength of the relationship between the variables : tenure and pay, then job grade and pay, and education and pay. When there is only one variable (tenure or job grade), the regression is called simple linear regression and is usually represented by a line in the middle of the data points. That’s the typical graph that you get from your compensation consultants.

The R-square value explains the strength of this relationship. The closer it is to 1 (or 100%), the more it explains the result. But in reality you never get 1. So for example, the R-square value between job grade and pay is 52%. It means that job grade “explains” 52% of an employee’s pay.

But life is never that simple, is it ? There is not only one factor which influences pay. As mentioned above, for example you suspect that tenure also influences pay. So you perform another regression analysis, and the result is that R-square is, say 29%. So tenure alone “explains” 29% of pay. Less than job grade.

You then perform a multiple linear regression to understand how the two independent variables (job grade and tenure) influence pay. Combined together, the R-square value is now for example 63%.

63% – 52% = 11%. It means that tenure explains 11% of pay, on top of job grade.

If you do this analysis multiple times, for each criteria that you think influences pay, you will be able to build a model “predicting” pay based on these criteria. For instance you can combine, job grade, tenure and education to explain pay.

Once you have exhausted all the independent, quantitative criteria, you may get to a total R2 value of, let’s say, 76% or 0.76. In our example with 3 variables, job grade explains 52% of pay, with tenure explaining an extra 11% of pay and education predicting another 13% of pay. The final 24% (100%-73%) are linked to qualitative criteria that drive pay.

I hope you find this explanation of R-square interpretation useful and that it will help you understand the results of your compensation analyses better so that you can make informed decisions in the future.

 

Liked this post ? Why don’t you subscribe by email to receive all my future articles ?

 

Related posts :

 

Print Friendly, PDF & Email

Speak Your Mind

*