Open almost any "best counties for schools" list and you are really looking at a map of household income. Rank counties by raw proficiency rates and the winners are the places where high earners already cluster: leafy suburbs, exclusive school districts, towns with median home prices north of a million dollars. The schools post big numbers. The kids walked in the door with most of those numbers already.
That tells you where to buy a house if you can afford the house. It tells you almost nothing about whether the schools are any good at the job of teaching.
So we ranked counties a different way. Instead of asking which schools score highest, we asked which counties have the most schools that beat what their poverty level would predict. The answer puts a stretch of the Texas border at the top, and it bumps a lot of famous wealthy counties down to the middle of the pack. That is not a glitch. That is the whole point.
How We Measured "Beating the Odds"
Here is the method in plain English. Within each state, we run a regression of school proficiency against the share of free and reduced-price lunch (FRL) students at each school. FRL share is the standard proxy for student poverty. That regression draws a line: given how poor a school's students are, here is roughly the proficiency you would expect.
Every school then sits somewhere relative to that line. Schools well above the line are doing more with what they have. Schools well below it are doing less. We take the schools in the top tier of that gap, the strong over-performers, and call them outperforming. Then we rank counties by the share of their scored schools that land in that top tier.
A few guardrails. We only included counties with at least 40 scored schools, so a single great elementary in a tiny county can't fluke its way to the top. The scoring runs on NCES enrollment and demographic data plus state assessment results. And because the comparison is within each state, a Texas school is judged against Texas, a California school against California. We are measuring effort against circumstances, not absolute scores. Keep that distinction in mind, because it changes everything below.
The Leaderboard
Share of each county's scored schools that outperform their predicted level:
- Webb County, TX (Laredo) · 50.0%
- Cameron County, TX (Brownsville) · 30.9%
- Kent County, MI (Grand Rapids) · 28.6%
- Grafton County, NH · 28.6%
- Hudson County, NJ (Jersey City area) · 28.5%
- Hidalgo County, TX (McAllen / Rio Grande Valley) · 28.4%
- Boone County, MO (Columbia) · 26.1%
- Collier County, FL (Naples) · 25.4%
- Boulder County, CO · 25.0%
- Miami-Dade County, FL · 23.6%
- Durham County, NC · 23.5%
- Orange County, CA · 22.9%
- Queens County, NY · 22.7%
- San Francisco County, CA · 22.3%
Half the schools in Webb County beat their predicted level. Half. To put that in context, the cutoff for outperforming is the top tier in the state, so in a perfectly average county you would expect a small slice of schools to clear it. Webb clears it with one school in two.
The Rio Grande Valley Is the Story
Three of the top six counties sit on the Texas border with Mexico: Webb (Laredo), Cameron (Brownsville), and Hidalgo (McAllen). These are not the counties that show up on glossy "best schools" maps. They are high-poverty. They are heavily Hispanic. A large share of students are English learners, many of them learning to read and test in a language they did not speak at home.
By the raw-score logic that drives most rankings, that profile is supposed to be a death sentence. Poverty plus English learners plus a border ZIP code is the exact combination people point to when they explain away low scores. And the schools here do not post the highest absolute proficiency in Texas. They are not trying to. What they do is clear the bar their own demographics set, over and over, at a rate that almost nowhere else in the country matches.
There is a real pattern under this. South Texas border districts have spent years building bilingual instruction, dual-language programs, and a college-going culture in places the rest of the state wrote off. The data here is not a feel-good anecdote. It is 50% of Webb County's schools doing it and a third of Cameron's. When the same kind of community shows up three times in your top six, that is a finding, not noise.
It Isn't Only the Underdogs
The list is not a clean morality tale where poor counties win and rich ones lose. Some affluent counties earn their spot honestly. Collier County, Florida, which is Naples and some of the most expensive real estate in the Southeast, lands at 25.4%. It would post high raw scores no matter what. It also beats its own raised bar, which is harder than it sounds, because the regression sets the bar higher for low-poverty schools too.
Boulder County, Colorado (25.0%) and San Francisco (22.3%) tell the same story. Wealthy, educated, expensive, and still outperforming the line drawn for places like them. That is the version of "good schools in a rich county" that actually means something. The schools are adding value on top of the advantages, not just banking the advantages.
Then there is the middle of the list, where the big metros sit. Miami-Dade (23.6%), Durham, NC (23.5%), Orange County, CA (22.9%), and Queens (22.7%). These are enormous, diverse, immigrant-heavy counties with a lot of moving parts, and they are quietly punching above their weight across hundreds of schools each. Queens alone is one of the most linguistically diverse places on earth, and more than one in five of its schools beats expectations.
Two more worth naming because they break the coastal-metro frame. Kent County, Michigan (Grand Rapids, 28.6%) and Hudson County, New Jersey (Jersey City, 28.5%) both crack the top five. Grand Rapids is a working Midwestern city, not a hedge-fund suburb. Hudson County is dense, immigrant-heavy, and sits in the shadow of Manhattan rents. And Grafton County, New Hampshire (28.6%) is the rural outlier, a reminder that beating expectations is not only an urban trick.
The Caveat
This measure rewards beating expectations, not absolute achievement and you should understand exactly what it does and does not say.
A wealthy county can post some of the highest raw proficiency scores in its state and still land in the middle of this list, because it was supposed to score high and merely did. That is not a knock on those schools. It is the math working as designed. If you are a family who can move anywhere and you only care about peak test scores, this ranking is not your shopping list, and the affluent districts you would expect are still excellent in absolute terms.
But if you want to know which schools are actually good at teaching, as opposed to which schools enrolled the easiest students, beating expectations is the cleaner signal. A school that lifts low-income kids past the line drawn for them is doing the harder job. Webb County does that at scale. Plenty of richer places don't.
What To Do With The Data
If you are house-hunting, run both numbers. Pull the raw proficiency for a county and pull how its schools do against expectations. A place that is high on both is genuinely strong. A place that is high on raw scores but mediocre on beating expectations is mostly selling you its neighbors' incomes, which is fine if that is what you want, but know that is what you are buying.
If you are a parent already in one of these counties, especially the border and Midwest entries that never make the magazine lists, the takeaway is simpler. Your schools are outrunning their circumstances, and the people running them figured out something most districts haven't. That is worth more than a high score in a town where every kid arrives reading at grade level.
And if you are a reporter or a policymaker looking for what works, the addresses are right there in the top six. The answers to closing achievement gaps are not theoretical. They are operating, at scale, in Laredo and Brownsville and McAllen, and almost nobody is writing about them.
Data note: rankings are based on allk12's BeatsExpectations scoring, a per-state regression of school proficiency on student-poverty share built from NCES SY 2024-25 data and state assessment results. Counties shown have at least 40 scored schools.
Sources
NCES: Common Core of Data, School and Agency Reports (enrollment, demographics, free and reduced-price lunch)
NCES Condition of Education: Concentration of Public School Students Eligible for Free or Reduced-Price Lunch
Brookings: Standardized Test Scores Track Family Wealth, Not Just Ability
Stanford Educational Opportunity Project (SEDA): How Socioeconomic Inequality Predicts Academic Achievement Gaps
