🔬 Methodology

The analysis behind the Collegiate Survivability Index (CSI) and related visualizations on UniversityDeathpool.com is rooted in data science, financial modeling, and institutional behavior trends observed over time.

This methodology outlines our approach to collecting, analyzing, forecasting, and interpreting data for over 1,300 private, not-for-profit universities across the United States.

🔢 Data Collection and Sources

Data was collected for all private, not-for-profit universities offering undergraduate degrees, provided that their information was publicly available via:

IRS Form 990 filings (financial data)
IPEDS (Integrated Postsecondary Education Data System) (academic and enrollment metrics)
U.S. Census and state-level projections (demographic trends)

We analyzed historical data from 2017 through 2023, using this timeline to establish trends and evaluate institutional health. Universities with missing, inaccessible, or incomplete data were excluded from the survivability analysis.

⏲ Forecasting to 2028

To project future conditions, we used the Holt-Winters exponential smoothing model, which allowed us to forecast individual data points for each institution through 2028, accounting for both seasonality and trend. These projections fed into the broader calculation of each institution’s risk.

🧬 Weighting Model: Learning from the Past

Our weighting system for the three CSI components—Adjusted CFI (aCFI), Academic Efficiency, and Market Saturation—was developed by analyzing characteristics commonly found among universities that have closed since 2017.

Each factor was assigned a weight based on its predictive value. This balance was informed by both regression analysis and machine learning optimization techniques.

We also applied a Sobol Sequence algorithm to generate quasi-random, low-discrepancy sampling across our multidimensional input space, improving the robustness of scenario simulations.

In parallel, a Bayesian model was used to estimate the posterior likelihood of institutional closure based on historical inputs, further validating the importance of each survivability factor.

📊 The Purpose and Construction of aCFI

The Adjusted Composite Financial Index (aCFI) was created to address the shortcomings of the traditional CFI used in higher education finance.

Specifically, the traditional CFI:

Penalizes institutions with large depreciation expenses, even if those are non-cash
Excludes restricted net assets, which are often available for use in financial crisis

Our aCFI model:

Removes depreciation from net income calculations
Includes restricted assets, under the assumption that most institutions would seek release of restrictions prior to insolvency

This approach provides a more realistic representation of a school’s short-to-midterm financial flexibility.

🕵️ Assumptions and Limitations

Like all models, our methodology includes a number of assumptions and known limitations:

Data availability: Universities with incomplete or missing 990/IPEDS data were excluded from analysis
Restricted funds: We assume that in cases of financial stress, most institutions would attempt to release donor restrictions
Depreciation: Treated as a non-cash expense and removed from consideration in the aCFI model
Program-level nuance: The model does not account for internal academic or athletic program dependencies
Market assumptions: Local competition and demographic conditions were weighted by state-level data, not regional micro-data
Projections: Forecasts are inherently uncertain; the Holt-Winters model assumes past trends will continue

Despite these limitations, the methodology offers a consistent, transparent, and predictive approach to identifying institutions at greatest risk of closure.

As new data becomes available, all models and visualizations will be updated accordingly.