Institutional Spending Composition and Student Outcomes: Analysis Panel and Replication Code (2015–2023)
Description
Institution-year panel dataset and SAS replication code for a study examining whether the composition of institutional spending (instruction, academic support, student services, and institutional support) predicts graduation and first-year retention rates at U.S. four-year colleges, independent of total spending. The panel covers 2,225 public and private nonprofit four-year institutions across eight collection years (2015-2019, 2021-2023; 2020 excluded due to pandemic-related reporting disruption), drawn from the National Center for Education Statistics' Integrated Postsecondary Education Data System (IPEDS). The dataset includes 18,760 raw institution-year observations; the complete-case analysis sample used in the primary models comprises 13,868 institution-years across 1,971 institutions. Variable definitions and construction are documented in the accompanying SAS script.
Files
Steps to reproduce
1. Download both files: institutional_spending_panel.csv (the analysis dataset) and institutional_spending_analysis_FINAL.sas (the replication script). 2. Place both files in the same folder. 3. Open institutional_spending_analysis_FINAL.sas in SAS 9.4 or later (requires base SAS plus SAS/STAT and SAS/ETS for PROC SURVEYREG and PROC PANEL). 4. In Step 0 of the script, edit the %let projpath macro variable to point to the folder containing the CSV file. 5. Run the script from top to bottom. It will: import and filter the panel; construct log-transformed spending variables; build the Mundlak (1978) correlated random-effects device; estimate the primary graduation-rate and retention-rate models; run robustness checks (alternative spending control, public/private subsamples, winsorized subsample); and run the two-way fixed-effects and Hausman diagnostics reported in the paper's Methods section. 6. A summary at the end of the script maps each output to its corresponding table or section in the paper.