UCSF Academic Freedom Report Concept Note

Background

National survey data (collected between December 2022 and December 2023) from AAC&U/AAUP/NORC Academic Freedom and Civil Discourse in Higher Education report, published in 2025, documented declines in perceived academic freedom, rising self-censorship, and concerns about external interference. Nearly 60% of faculty reported self-censorship, and over one-third felt constrained by institutional or political pressures. The current political climate in the United States has likely intensified the chilling effect on legitimate expression and research since the data has been collected. Recent events, such as the University of California, Berkeley investigations under federal scrutiny for campus speech (Guardian, Sept 2025), exemplify how administrative and political actions threaten autonomy and well-being. While burnout and stress among faculty are well-documented, no causal study has yet linked academic freedom constraints to mental health outcomes using quasi-experimental methods.

Within the University of California (UC), the University Committee on Academic Freedom (UCAF) is mandated to safeguard academic freedom principles. While California is classified as a “no restrictive legislation” state, UC faculty likely experience indirect pressures from federal funding agencies that are not captured by state-level categorizations. Understanding these dynamics may inform UC Academic Senate decision-making, guide institutional supports, and strengthen advocacy for faculty autonomy.

University of California, San Francisco’s (UCSF) health sciences focus—and the absence of any undergraduate program—creates a distinct academic environment where faculty face unique pressures around research topics, including public health policy, pharmaceutical industry relationships, and clinical guideline development. These features may shape academic freedom dynamics differently than in traditional liberal arts or university settings. The lack of an undergraduate population concentrates UCSF’s teaching and scholarship within professional and graduate education, where research agendas are closely tied to clinical and policy implications. These dynamics underscore the need to generate UCSF-specific data to inform targeted interventions and contribute to broader national understanding. Documenting the extent of self-censorship or concern about expressing scholarly within UCSF’s distinct institutional structure would provide insight into the intersection of academic freedom, research climate, and faculty mental health.

In addition to leveraging core items from the AAC&U/AAUP/NORC survey, this study proposes integrating data collection instruments from organizational psychology—including Edmondson’s Psychological Safety Scale and the Organizational Trust Inventory—to examine perceptions of psychological safety, collegial trust, and institutional support. These measures will enable exploration of how faculty experience the balance between expression and professional risk. Future extensions may also incorporate research integrity–related measures, including perceptions of institutional protection for ethical inquiry and publication independence, to map the broader ecology connecting academic freedom, psychological climate, and mental health.

In alignment with the Committee for Academic Freedom bylaw requiring a campus-based academic freedom report every five years, we propose conducting the report using a robust research methodology outlined in this concept note. We will apply for UCSF Institutional Review Board approval for the study proposed. The collected data will be analyzed and used to prepare an internal report on the state of academic freedom at UCSF, as well as a draft manuscript for peer review. 

Methods

Study Design

  • Cross-sectional survey of UC faculty and comparable institutions (target N = 1,000).
  • Primary exposure: Self-reported academic freedom constraints (binary/ordinal composite index).
  • Outcomes: Continuous wellbeing and mental health scores (WHO-5, PHQ-9, etc.).

Causal Inference Strategy

  1. Propensity Score Matching (PSM): Match constrained and unconstrained respondents based on demographic and institutional covariates.
  2. Coarsened Exact Matching (CEM): Ensure exact balance across key dimensions.
  3. Future work: implement DiD or micro-level synthetic control (Xu 2017; Athey 2021) for panel data.

Objective and Aims

Objective: To quantify and understand the landscape of academic freedom at an academic medical university with no undergraduate program and the impact of academic freedom constraints on faculty mental health and well-being.

  • Aim 1: Use cross-sectional matching methods to estimate associations between exposure to academic freedom constraints and faculty mental health outcomes, distinguishing between intramural constraints (e.g., university sanctions) and extramural constraints (e.g., federal grant cancellation based on research topic).
  • Aim 2: Prepare for longitudinal extension using difference-in-differences (DiD) and synthetic control designs once repeated survey waves or policy shocks data are available.
  • Aim 3: Translate empirical findings into policy guidance for UCAF and peer institutions to strengthen academic freedom protections.

Survey Instrument

The study adapts core domains from the AAC&U 2025 Academic Freedom Survey, supplemented by validated mental health and demographic instruments.

  • Academic Freedom Domains (AAC&U, 2025)
    (See Appendix A for the full reconstructed question list and trimming recommendations)
    1. Institutional climate for open discourse
    2. Self-censorship and topic avoidance
    3. External pressures (political, donor, media, governmental)
    4. Administrative oversight and investigations
    5. Perceived protections and due process confidence
    6. Trust in leadership and peer support for academic freedom
  • Mental Health Measures (Proposed)
    (See Appendix B for proposed items and justification.)
    • WHO-5 Wellbeing Index (positive affect; 5 items)
    • PSS-10 Perceived Stress Scale (10 items)
  • Sociodemographic & Faculty Characteristics
    (See Appendix C for the proposed faculty demographic section.)
    • School, rank, tenure status, series (i.e., Adjunct, Clinical X, HS Clinica, In residence, Ladder rank)
    • Gender, race/ethnicity, age, political affiliation (optional)
    • Faculty governance participation

Sample Size & Power Simulation Summary:

Power analyses were conducted to ensure adequate sensitivity to detect policy-relevant differences in well-being across faculty exposed to varying levels of academic freedom constraints. Assuming a two-tailed α = 0.05 and 80% power under a matched or weighted two-group design (exposed vs. unexposed), a total sample of approximately 1,000 respondents (≈500 per group) will detect standardized mean differences of d = 0.22–0.25 on continuous outcomes such as the WHO-5, PHQ-9, and GAD-7. This estimate incorporates a conservative design effect of 1.2–1.3 to account for covariate adjustment and weighting. With an effective analytic N of 850–900 (allowing for 10–15% attrition), detectable effects remain in the small-to-moderate range (d ≈ 0.24–0.27). Subgroup analyses (e.g., by rank or disciplinary field) will retain ≥0.80 power for d ≥ 0.33. The resulting target sample of N ≈ 1,000 provides adequate precision to detect meaningful, policy-relevant associations between perceived constraints on academic freedom and faculty mental health, while maintaining analytic efficiency for subgroup exploration and matching-based causal inference.