A degree ranking describes adults living in a state
Our bachelor’s degree ranking reports the share of adults age 25 and older with a bachelor’s degree or higher. It describes residents’ qualifications. It does not follow a cohort of children through the state’s schools or assess what those schools contributed to learning.
Graduates can move between states. A state can attract adults educated elsewhere, while adults educated in the state may leave. Age composition and labor markets also shape who lives there. The adult degree share therefore cannot isolate school effectiveness.
Keep the denominator consistent
The high school attainment measure includes adults age 25 and older who have a diploma or equivalency, including those with higher qualifications. The advanced-degree measure uses the same adult age floor but a higher qualification threshold.
These are shares of adults, not shares of current students. High school completion among adults is different from the adjusted cohort graduation rate for students. A graduation rate follows a defined student cohort and needs its own treatment of transfers and other changes.
Achievement, spending, and attainment are separate measures
Student assessments describe performance in a subject at a grade level. School spending describes resources under an accounting definition. Degree attainment describes completed qualifications among a population. A single “education ranking” that combines them is a composite with method choices, not a direct observation of one property.
The National Assessment of Educational Progress provides a common assessment framework. Scores and uncertainty need to be read by grade, subject, and assessment year. State-defined proficiency rates from different tests should not be treated as a uniform yardstick merely because each is expressed as a percentage.
Per-pupil spending also needs context about the expenditure definition and denominator. It is not a direct measure of learning or a causal estimate of what extra spending would produce. Comparing spending and outcomes across states cannot, by itself, establish the effect of a policy.
A close rank may not establish a difference
The ACS estimates in this edition have published 90% margins of error. A difference in the point estimates can be smaller than the uncertainty in their difference. Use the state comparison tool to test selected pairs instead of treating neighboring ranks as distinct tiers of performance.
The tests are pairwise and do not form a definitive league table. They also do not explain why a difference exists. A statistical comparison is evidence about two estimates under a particular method, not a judgment about students, teachers, or residents.
Use a title that names what was measured
“States by bachelor’s degree attainment” identifies both the concept and an observable measure. “Best schools” implies a much broader judgment that this dataset does not support. This edition publishes three adult-attainment measures; it does not publish a school-quality composite, NAEP ranking, or school-spending ranking.
See Census educational attainment definitions, the linked producer files on each ranking, and the site methodology for units, ranking rules, and uncertainty treatment.
Worked example: the qualification threshold
In Massachusetts, 48.3% of adults age 25 and older have a bachelor’s degree or higher; 22.6% have an advanced degree. Both use the adult age-25-plus denominator. The second group is contained within the first, so adding these percentages would count advanced-degree holders twice.
The difference, 25.7 percentage points, describes the point-estimate share with a bachelor’s degree as their highest qualification. Our independent-state test is not appropriate for a difference between these overlapping groups within one state. This example illustrates the denominator and overlap; it does not estimate the performance of Massachusetts schools. Compare the advanced-degree table.