Bimodal Distribution

A bimodal distribution has two distinct peaks, often showing that one dataset actually combines two different groups.

A bimodal distribution is a distribution with two clear peaks, which usually means the data are being generated by two different underlying groups rather than one single average pattern.

Why economists care

Economists often summarize data with a mean, median, and variance. That can hide important structure. If wages, firm productivity, or test scores are bimodal, the issue is often not “more spread” but “two populations in one sample.”

Common examples include:

  • wages split between low-skill and high-skill workers,
  • firm productivity split between informal and formal producers,
  • inflation expectations split between well-anchored and poorly anchored households.

How to read it

Two humps in a histogram suggest bimodality, but economists still have to ask whether the pattern is real:

  • Is the sample large enough?
  • Are the peaks stable across years or subsamples?
  • Is the apparent bimodality just noise from bin choice?

If the pattern is real, the next step is usually to model the separate groups directly instead of forcing one average response on everyone.

Economic interpretation

Bimodality often points to segmentation, threshold effects, or regime differences. In labor economics, that may mean two skill tiers. In development economics, it may reflect a poverty trap where some households stay near a low-income equilibrium while others move to a higher one.

That is why bimodality matters for policy. A single policy aimed at the average observation may fit neither group well.

Knowledge Check

### What does a bimodal distribution usually suggest? - [x] The sample may contain two distinct underlying groups - [ ] The data have zero variance - [ ] The mean and median must be identical - [ ] The distribution is always normal > **Explanation:** Two peaks often indicate that one sample is mixing observations generated by different economic processes or populations. ### Why can a simple average be misleading in bimodal data? - [x] Because the average may fall between two groups and describe neither one well - [ ] Because averages are never used in economics - [ ] Because bimodal data always have no median - [ ] Because histograms cannot display two peaks > **Explanation:** When two clusters exist, the overall mean can sit in a low-density region that is not representative of either cluster. ### In applied economics, what is a common next step after detecting real bimodality? - [x] Model the subgroups or regimes separately - [ ] Drop the data immediately - [ ] Replace the sample with a normal distribution - [ ] Ignore the pattern and report only the mean > **Explanation:** Economists usually try to identify the mechanism behind the two peaks, such as segmentation, selection, or multiple equilibria.