Weighted Employee Pay Averages vs. Medians
Question: In my role as a Compensation Manager, I compare my company’s internal pay rates against the marketplace. But I’m always curious as to whether I should be using the Weighted Average or the Median as the “central measure of tendency”? Is one better than the other?
Answer: Perhaps not “better” but there are good reasons to use one or the other. It depends upon how you value each of these measures, and to some extent upon what you are hoping to accomplish. Let’s look at the differences.
Of these two measures, the vast majority of Compensation practitioners use the Weighted Average as the central measure of tendency. It is “weighted” by the number of employees whose pay rates make up the Average. Put simply, the Weighted Average moves up if there are more employees paid at higher rates. The average is driven down if more employees are paid at lower rates. Thus it is “weighted.”
This measure is considered more “realistic” than the Median, whose value is unchanged if the highest or lowest rates increase or decrease in any way. Thus the Median is seen as failing to consider the effects of very high or very low rates of pay.
You may notice that in large on-line databases of pay data that the sponsor of the data uses the Median commonly in lieu of the Weighted Average. This is because the job matching is not as accurate, producing a high standard deviation (i.e.. the range of reported pay rates is wider). There is so much dissimilarity in the data that the survey site’s sponsor compensates for this through the use of the Median as a measure of central tendency.
It could be said that it’s good to use some measure, whether the Weighted Average or the Median, rather than not make a market comparison of your internal data. But the Weighted Average is dominant in its use because it considers the effect of all of the pay rates in the survey upon the average, whether they are high or low.