When summarizing test results, we often use the median or a percentile, instead of the average, to represent the data set. The median, where 50% of the samples are below and 50% are above, is used because its not affected by a few outlier measurements.

A great example of averages getting skewed happened in 2013, where the average pay for a tech worker in San Mateo county soared by 38% in that year.  The cause of this increase? Mark Zuckerberg cashed in $3.3 Billion of his Facebook stock. With Zuckerberg’s pay, the average was $290K, without his pay it was $210K. His pay alone increased the average by $80K, or 38%.

Another reason to be wary of averages.

“Mark Zuckerberg at the 37th G8 Summit in Deauville 018 v1” by Guillaume Paumier – Own work. Licensed under CC BY 3.0 via Wikimedia Commons – https://commons.wikimedia.org/wiki/File:Mark_Zuckerberg_at_the_37th_G8_Summit_in_Deauville_018_v1.jpg#/media/File:Mark_Zuckerberg_at_the_37th_G8_Summit_in_Deauville_018_v1.jpg