Our Methodology

Earnings Data

Salary figures are sourced from the US Census Bureau's Post-Secondary Employment Outcomes (PSEO) dataset. We use median (50th percentile) earnings at 1, 5, and 10 years after graduation. Cells marked as suppressed by the Census Bureau (small cohort sizes) are excluded.

Purchasing Power Adjustment

We adjust nominal salaries using a state-level cost-of-living (COL) index informed by BEA regional price parity benchmarks and refreshed with newer price-trend estimates.

Formula:

Adjusted Salary = Nominal Salary × (Target COL Index / School State COL Index)

For our "National Purchasing Power" score, the target COL index is 1.0000, which represents the national average. Some published COL tables use 100 as the national baseline; in this app we normalize those values to 1.0000 internally. A state with an index of 0.88 means goods and services cost about 12% less than the national average, while a state with an index of 1.84 means prices are about 84% higher than the national average.

Break-Even / Payback Timer

The payback timer answers: "How many years until cumulative earnings advantage exceeds total college investment?"

Total Investment = (Annual Cost of Attendance × 4) + (HS Baseline Salary × 4)

The HS baseline is $38,792/year (BLS 2023 median for high school diploma holders). Opportunity cost represents earnings foregone during 4 years of college. We do not apply a discount rate in this simplified model.

Earnings between the three PSEO data points (year 1, 5, 10) are linearly interpolated. If break-even is not reached by year 10, we extrapolate using the year-10 earnings advantage.

Career Verdicts

Career Verdict text is generated using the OpenAI API and constrained to use only data passed in from the official Census PSEO dataset. We keep the prompts tightly grounded in the imported statistics, review the data mappings and prompt behavior, and revise cached outputs when issues are identified.

All AI-generated verdicts are grounded in official Census datasets and undergo a human-in-the-loop review process to ensure data integrity.

Limitations

• PSEO data is released with a 2-3 year lag.
• Small institutions or programs may have suppressed (missing) data.
• Earnings reflect all graduates regardless of employment type (full-time, part-time).
• COL indices are state-level averages and may not reflect specific cities.