Open Health Data Hub



All posts
|3 min read

Why College Graduates Are Now the Heaviest Drinkers in America

heavy drinking by education levelbinge drinking college graduatesalcohol consumption trends BRFSShigh income alcohol usedrinking disparities education

Among college-educated men, binge drinking has held above 18% for over a decade. That's higher than any other education group in the dataset, and it hasn't meaningfully come down.

The conventional picture of problem drinking in America centers on economic stress, limited opportunity, and lower educational attainment. The data tells a different story.

College Graduates Lead Every Education Group on Binge Drinking

Start with the baseline comparisons. In 2015, high school graduates had a binge drinking prevalence of 16.1%. Those who didn't finish high school came in at 12.9%. College graduates, by contrast, sat well above both groups, with male graduates reaching 21.2% in 2017, the highest recorded rate in the dataset for any education-sex combination.

By 2024, high school graduates had drifted down to 15.3%. Non-graduates ticked up slightly, from 12.9% in 2015 to 14.1% in 2024, but remained the lowest group. College graduate males, meanwhile, stood at 18.7%. The educational gradient on binge drinking doesn't run in the direction most people assume. More education correlates with more binge drinking, not less.

This matters because public health messaging and clinical screening tools are often calibrated around the assumption that higher-risk drinkers are concentrated in lower-income, lower-education populations. If the actual distribution is inverted, those tools are aimed at the wrong people.

The Gender Gap Among Graduates Is Narrowing, But Slowly

Within the college graduate group, the male-female gap has compressed only modestly over the past decade. In 2014, male graduates binged at 19.7% versus 13.5% for female graduates, a gap of 6.2 percentage points. By 2024, that gap had closed to 5.9 percentage points (18.7% vs. 12.8%).

That's a shift of just 0.3 points over ten years. The convergence narrative around female drinking, which has received substantial media attention, doesn't show up strongly in this population. Female college graduates actually recorded their lowest binge drinking rate in the dataset in 2024, at 12.8%, while male rates remained elevated. The gap persists not because women are catching up, but because men haven't come down.

For providers, this means the elevated-risk college graduate population is still predominantly male, and the clinical picture hasn't changed as dramatically as cultural commentary might suggest.

Income Adds Another Layer: Higher Earners Drink More Heavily

The income pattern reinforces the education finding. Among females earning $50,000 or more, heavy drinking prevalence was 7.8%, higher than females in every lower income bracket. Women earning under $15,000 had a heavy drinking rate of just 4.0%. That's nearly double the rate at the top of the income distribution compared to the bottom.

For men, the income gradient is flatter. Males earning $50K or more had a heavy drinking prevalence of 7.2%, the same as males earning $35,000 to $50,000. The income effect on heavy drinking is more pronounced for women than for men, which means rising female incomes and educational attainment could push female heavy drinking rates higher over time even as binge drinking rates hold steady or fall.

Put the education and income findings together: the Americans most likely to design health surveys, set clinical guidelines, and write insurance policy are also the Americans with the highest measured rates of binge and heavy drinking. The people assessing the problem may belong to the highest-risk group.

Female college graduates' binge drinking dropped to 12.8% in 2024, the lowest value recorded for that group in the dataset, even as male college graduate rates held at 18.7%. What drove that divergence, and whether it holds, is the question the next decade of data will have to answer.

Explore the data yourself

Run your own queries against 240M+ rows of federal health data using natural language — powered by AI.

Start analyzing