Explaining Right Skew

Welcome back to Skewed Distributions! This is lesson three of four, so we are well past the halfway mark. In the first two lessons we built a toolkit for seeing and naming skewed shapes: we learned to spot asymmetry with one long tail and to label the skew by the direction that tail points. Now comes the deeper question — what makes a distribution skewed in the first place?

In this lesson we focus on right skew and the surprisingly simple mechanism behind it. We will see why variables like household income, home prices, and customer wait times almost always lean the same way. Once the mechanism clicks, we will be able to predict right skew before we even look at a graph.

Recognizing a shape and explaining its cause are two very different skills. If someone shows us a histogram of home sale prices and asks, "Is this right-skewed?" we can answer by finding the long tail. But if a colleague mentions a variable we have never plotted — say, the time customers spend on hold with a support line — and asks, "What shape would you expect?" we need a reason, not just a visual memory.

That reason comes from understanding the process that generates the data. When we know what kind of process tends to create right skew, we can apply that logic to any new variable and make a well-founded prediction. The rest of this lesson builds that reasoning, one ingredient at a time.

Many everyday quantities share a quiet but important feature: they cannot drop below some low value. This lower boundary is often called a floor. Here are a few familiar examples:

• Household income is often treated as having a floor near $0 for this kind of example.
• Home sale prices cannot be negative — a house sells for at least some positive amount.
• Wait time on a support line starts at zero; no one waits a negative number of minutes.
• Reaction time has a biological minimum around 150–200 milliseconds; our nervous system simply cannot fire faster.

A floor squeezes the left side of the distribution. Values cannot spread out very far in that direction because the boundary stops them. Think of it like a wall on the left: data can pile up against it, but it cannot pass through.