The image below shows a distribution with positive skew or right skew.

The image below shows a distribution with negative skew or left skew.

Some people find these concepts counterintuitive, assuming that skew translates roughly to "where most of the values are." In fact, skew might be better described as "where the extreme values are."

In continuous probability density functions, interpreting skew can be complicated. In empirical social science, we usually talk about distributions of a finite number of observations; these are sometimes easier to describe because the "tails" do not extend outward to infinity.

Here is a histogram of a sample variable with a positive skew.

Here is a histogram of a sample variable with a negative skew.

With this relatively common kind of data, where there is one clear peak in the distribution and most of the extreme values fall on one side of the peak, it is easy to visualize the skew at a glance.

To help remember what positive and negative
(or right and left) skew look like, students can
**look for the extreme values** or
**imagine an arrow** pointing in the direction of the skew.

To some people, the long tail of the histogram looks a bit like an arrow pointing in the direction of the skew.