How to Measure the Diversity of a Place

“Diverse” is one of the most loosely used words in the way we talk about places. People apply it to any city with a large minority population, any big metro, any place that isn’t monochrome. But diversity is a measurable property, and measuring it precisely produces some genuinely surprising results — including the fact that several of America’s most heavily minority cities rank among the least diverse in the country. The tool that makes sense of this is a diversity index, and it’s worth understanding what it actually calculates.

The question the index answers

A diversity index boils a place’s entire racial and ethnic makeup down to a single number by answering one precise question:

If you picked two residents of this place at random, what’s the probability they’d belong to different groups?

That’s it. If nearly everyone shares one background, two random people will almost always match, and the probability of a difference is low — a low diversity score. If the population is spread evenly across many groups, two random people will usually differ, and the score is high. The index runs from near 0 (total homogeneity) to a theoretical maximum approached only when many groups are present in equal shares.

Mechanically, you take each group’s share of the population, square it, add those squares up (that’s the chance two random people match), and subtract from one. Squaring is what makes the measure sensitive to evenness: a group at 90% contributes 0.81 to the matching probability all by itself, so any place dominated by one group scores low no matter how many small groups it also contains. Nationally, the index comes out around 0.61 — better than even odds that two random Americans differ by group.

Why “minority-heavy” and “diverse” aren’t the same thing

This is where the index overturns a common assumption. Because it measures evenness, not the mere presence of minorities, a place can be overwhelmingly non-white and still score as barely diverse — if that non-white population is concentrated in a single group.

The data shows exactly this. Among larger cities, some of the lowest diversity scores belong to places like Laredo, Texas and Hialeah, Florida, where around 90% or more of residents are Hispanic. These cities are almost entirely minority — and precisely because they’re so dominated by one group, two random residents almost always share it, so the index reads them as homogeneous. Diversity, measured honestly, is about the mix, not the label.

The most diverse cities tell the opposite story — no single group anywhere near a majority:

City	Diversity index
Vallejo, CA	0.78
Federal Way, WA	0.78
Jersey City, NJ	0.77
Oakland, CA	0.77

Places like Vallejo and Oakland score at the very top not because they have the most of any one group, but because White, Black, Hispanic, and Asian residents each hold a substantial share — the population is genuinely spread out. That’s what maximum diversity actually looks like in the arithmetic.

What the index can and can’t see

The diversity index is powerful because it’s a single, comparable number — you can rank all 31,555 places by it, which raw percentages don’t let you do. But it has real limits worth keeping in mind.

It’s only as good as its categories. The index uses the standard Census race and ethnicity groups, and — as we cover in how the Census measures race and ethnicity — those come with quirks. Hispanic origin is a separate dimension from race, and how the two are combined affects the score. Our index treats the major groups consistently so places are comparable, but no single-number summary can capture every dimension of who lives somewhere.

It says nothing about integration. A city can score highly diverse in aggregate while its neighborhoods remain sharply segregated block by block. The index measures the mix within a boundary, not whether the groups actually live side by side. A high citywide score and deep neighborhood segregation can coexist — the index sees the first and is blind to the second.

And it’s scale-dependent. A place measured as one unit will almost always look more diverse than any of its individual neighborhoods, because aggregating hides internal sorting. Comparing a whole city to a single small town isn’t quite comparing like with like.

Reading a diversity score

With those caveats, the index is one of the most useful summary numbers in a profile — as long as you read it for what it is: a measure of how evenly mixed a place is, not how many minorities live there.

A high score means no single group dominates. A low score means one group does — and that group might be a national minority. Don’t read “low diversity” as “mostly white”; read it as “mostly one group, whichever group that is.” Set the index beside the actual composition and you get the full picture: the number tells you how mixed, and the breakdown tells you mixed of what.

We compute a diversity index for every place on the site and rank states and cities by it — most diverse states is topped by Hawaii, the most evenly mixed state in the country, and most diverse cities surfaces the urban leaders. Read them knowing what the number is really counting, and “diverse” stops being a vague compliment and becomes something you can actually measure.