Performative Progressivism Study
Performative Progressivism and Integrity Fatigue: Reported Toll of Ideological Conformity in Higher Education
Kevin M. Waldman & Forest Romm
psychFORM Research Lab
Author Note
psychFORM Research Lab is an independent research entity unaffiliated with university departments or funding sources. All data were collected by research staff using standardized, non-evaluative protocols. Correspondence concerning this article should be addressed to Kevin M. Waldman or Forest Romm, psychFORM Research Lab, 3450 N Lake Shore Drive, Apt. 2304, Chicago, IL 60657. Email: kevin@psychform.com.
Abstract
In contemporary campus culture, belonging is increasingly negotiated through affective alignment, creating conditions in which students learn that how one appears to feel about sanctioned issues may matter more than what one actually believes. This multi-year investigation (2023–2025) assessed N = 1,452 undergraduates at Northwestern University (n = 720) and the University of Michigan (n = 732) via structured, in-person surveys administered in non-evaluative settings. A core design feature was instrumental neutrality: each directional item was mirrored by its inverse (counter-question), item order was randomized, and prompts avoided emotive or moralized language. This paired, counterbalanced architecture minimizes leading-question bias and permits clear inference about the direction of respondents’ own convictions rather than the survey’s framing.
Across both institutions, students reported high rates of ideological self-performance (88.2%), coursework misrepresentation (82.4%), and self-censorship (gender identity: 78%; political belief: 72%; family values: 68%). A gender-discourse module showed elevated numbness (72%), confusion (68%), and anxiety (64%), alongside a substantial desire for authenticity (61%). Grounded in attachment-based models of belonging and safety, minority stress/concealment research, and discrepancy/dissonance frameworks, we introduce Integrity Fatigue - a cumulative affective-cognitive depletion associated with sustained incongruence between private belief and public performance in reputationally evaluative contexts. Findings indicate that persistent impression management in academically evaluative environments may condition students away from autonomous identity formation and toward externally regulated alignment.
Keywords: ideological conformity, impression management, self-censorship, authenticity, identity development, campus climate
Introduction
Institutions of higher education are not only sites of instruction; they are relational ecologies in which belonging and safety are continuously negotiated. Attachment-derived perspectives hold that individuals calibrate self-disclosure based on anticipated responsiveness from significant others and communities; when inclusion appears conditional, people preferentially down-regulate divergence and up-regulate conformity to preserve proximity and approval (Ainsworth, 1989; Bowlby, 1988). For socially marked groups, the minority stress framework further predicts that concealment - the strategic inhibition of stigmatizable thoughts or identities - functions as a chronic regulator of threat, but at psychological cost (Meyer, 2003; Pachankis, 2007). Complementarily, discrepancy theories propose that sustained misalignment among the actual, ought, and ideal selves yields characteristic affective profiles (e.g., anxiety, guilt, dejection), while classic cognitive dissonance accounts describe the load imposed by maintaining incompatible cognitions or behaviors under social pressure (Higgins, 1987; Festinger, 1957).
Across these literatures, a convergent prediction emerges: when belonging and status appear contingent on affective alignment, people will perform agreement even when private belief diverges - especially in evaluative settings where reputational stakes are salient (e.g., classrooms, graded assignments, peer-visible discussions). We contend that contemporary campus climates can instantiate precisely these contingencies. To specify the phenomenon, we advance Integrity Fatigue - a construct denoting the cumulative depletion that accrues when individuals repeatedly inhibit authentic belief expression and substitute ideological impression management to maintain social and academic viability in reputational economies.
The present study tests this developmental-sociocognitive account using a neutral, counterbalanced survey instrument administered in person at two elite universities. We examined (a) prevalence of ideological self-performance; (b) domain-specific self-censorship; (c) misrepresentation in coursework; (d) private beliefs versus public conformity in gender-discourse contexts; and (e) affective correlates (numbness, confusion, anxiety, desire for authenticity). We hypothesized that behaviors consistent with impression management would be common and associated with an affective profile predicted by discrepancy and concealment frameworks, even when measurement was explicitly balanced to avoid directional priming.
Method participants
Undergraduates were recruited at Northwestern University (n = 720) and the University of Michigan (n = 732) between Fall 2023 and Spring 2025 using standardized campus-intercept procedures at pre-randomized locations and times. At Michigan, 244 participants identified as male and 488 as female; at Northwestern, 261 identified as male and 459 as female. Across institutions, self-described ideology was 67% liberal, 21% moderate, and 12% conservative.
Nonresponse for sensitive counter-items was minimal (average 7 %).
Design and Procedure
We employed a cross-sectional survey design with in-person administration to reduce satisficing and context confounds associated with remote panels. We approached potential participants in public campus spaces (e.g., libraries, quads) and invited them to complete a brief, anonymous instrument printed surveys. No identifying information (name, email, IP, device ID) was collected. Participants were spaced to prevent visual spillover. Enumerators were instructed to maintain neutral affect, avoid paraphrasing items, and refrain from clarifying content beyond reading the item verbatim.
To minimize order effects and moral priming, all ideological items were paired with their inverse (counter-question), and item order was randomized at the pair level and within pair. Response options used symmetric Likert anchors (e.g., Strongly disagree↔Strongly agree; Never↔Very often) or dichotomous Yes/No where conceptually appropriate. Sensitive modules (e.g., gender-discourse beliefs) were separated by neutral filler items to reduce carryover.
measures Methodological Rationale and Design Safeguards
- Dual-item Mirroring - Every value-pressure item was paired with its inverse (e.g., “Do you feel pressured to align?” vs. “Do you feel free not to align?”). This ensured measurement of internal permission and detected latent fear through response asymmetry.
- Contextual Tagging - Students identified where they felt pressure (classroom, peer discussion, online submission). Classroom settings accounted for 62 % of reported self-censorship.
- Cross-Campus Verification - Parallel survey instruments administered at UM and NU with identical phrasing. Site coefficients within 3 % of each other → strong cross-validation.
- Theoretical Anchoring - Analyses interpreted via Erikson (1968) on identity vs role confusion, Marcia (1966) on foreclosure, and Self-Determination Theory (Deci & Ryan, 2000) on introjected regulation.
Theoretical Frame
Performative Progressivism is defined as strategic exaggeration of ideological alignment to preserve academic and social safety. Integrity Fatigue emerges as its affective correlate — a chronic depletion from sustained suppression of authentic belief expression.
This framework integrates:
- Erikson’s ego development model (autonomy vs role confusion),
- Marcia’s identity status paradigm (foreclosure under external authority),
- Self-Determination Theory (loss of autonomy through introjected motivation), and
- Festinger’s cognitive dissonance as institutionalized behavioral adaptation.
Bias-Control Features
- Paired, inverse wording for all directional constructs
- Randomization within and between pairs
- Symmetric anchors and neutral lexical choices (no moralized terminology)
- In-person anonymity to reduce reputational demand characteristics
- Filler blocks to disrupt carryover in identity-salient modules
Ethics and Protection of Human Subjects
Procedures conformed to ethical standards for minimal-risk survey research. Participation was voluntary, uncompensated, and anonymous; respondents could skip any item. No protected health information or identifiers were collected.
Analytic Plan
We report descriptive prevalence estimates and cross-tabulations by institution and gender for interpretive clarity. For instrument validity, we inspected directional symmetry by comparing endorsement of primary items versus counter-items, anticipating non-equivalence if true directional effects exist. (Inferential models available upon request; the present report emphasizes prevalence and construct definition.)
Performative Progressivism and Integrity Fatigue: Phase I
Purpose of counter-balancing questions
The goal of this expansion is to integrate counter-balancing questions into each domain of inquiry to avoid directional bias and capture the full developmental spectrum of ideological adaptation - including authenticity, resilience, and moral agency.
Where the original questions identified conformity and suppression, the new parallel items measure self-trust, differentiation, and critical independence.
This dual design yields a complete psychological map:
Conformity Pressure ↔ Authentic Resistance
Moral Fatigue ↔ Moral Reflection
Self-Censorship ↔ Self-Expression
Section I - Identity Architecture A. Ideological Self-Performance
Item (A1). “Have you ever pretended to hold more progressive views than you believed to be accepted socially or academically?”
Results. Yes 88.2 % (n = 1,280); No 11.8 % (n = 172)
Campus × Gender. UM-Female 91 % (444/488); UM-Male 83 % (203/244); NU-Female 88 % (404/459); NU-Male 88 % (229/261)
Counter (A1-c). “In your classes, reasoned disagreement with progressive views is treated with respect.”
Agree 21 % | Disagree 72 % | NR 7 %
1) Construct and Rationale
This dimension operationalizes performative alignment—the strategic exaggeration of ideological agreement to safeguard social/academic standing. The primary item was intentionally framed in first-person, past-behavior terms to reduce social desirability bias (“Have you ever pretended…” rather than “Do students pretend…”). The counter-item measures perceived permission structure for dissent, an essential validity check: if performative alignment is high and perceived respect for disagreement is low, the behavioral report is plausibly adaptive rather than confessional.
2) Distribution and Patterning
The overall endorsement (88.2 %) positions ideological self-performance as a normative rather than exceptional behavior at both universities. Female students at UM show the highest rate (91 %), suggesting stronger social contingency or tighter reputational coupling in that context. NU males and females converge (both 88 %), indicating that at Northwestern, the pressure is less gender-differentiated and more structurally ambient.
3) Cross-Campus and Gender Contrast
- UM vs NU. UM’s slightly higher female rate (91 % vs NU 88 %) and lower male rate (83 % vs NU 88 %) produce a wider gender gap at UM. This pattern is consistent with a climate where women perceive stronger interpersonal evaluation in classroom and peer settings, while men at NU report pressures comparable to women - a more uniformly moralized classroom ecology.
- Implication. Where gaps narrow (NU), the climate functions less through gendered peer micro-norms and more through generalized institutional signaling.
4) Counter-Item Interpretation (A1-c)
Only 21 % perceive respectful treatment of reasoned disagreement; 72 % disagree. The 51–67 point gap between self-performance (88 %) and perceived respectful disagreement (21 %) yields a coherent behavioral logic: students perform because environments code dissent as reputationally costly. The 7 % NR suggests students were willing to render a judgment on classroom norms, bolstering interpretive confidence.
5) Validity Checks and Potential Confounds
- Demand effects: Mitigated by past-tense, behaviorally specific wording.
- Pluralistic ignorance: High endorsement may reflect perceived “everyone does it”; however, the counter-item’s low permission independently substantiates the adaptive value of performance.
- Single-item limitation: Because A1 is a single indicator, we rely on triangulation with C1 (coursework misrepresentation), B1 (self-censorship), and D1 (harm-coding) for convergent validity - each aligns in direction and magnitude.
6) Developmental Meaning
Within an Eriksonian frame, students default to role protection over exploration. In SDT terms, motivation shifts toward introjected regulation (avoid guilt/shame/punishment) rather than integrated values. The repeated act of self-performance initiates the Integrity Fatigue loop: short-term safety → long-term depletion.
B. Self-Censorship Domains
Primary (B1). “Intentionally self-censored to avoid ideological conflict.”
Counter (B1-c). “Faculty or peers have explicitly encouraged honest disagreement on sensitive issues.”
Yes 17 % | No 76 % | NR 7 %
1) Construct and Rationale
Self-censorship is differentiated by domain to reveal where evaluative threat concentrates. We expected gender to dominate due to identity essentialism - topics framed as core to persons rather than propositions. The counter-item evaluates whether explicit encouragement for dissent exists; absence of such signaling is known to depress authenticity even in “tolerant” climates.
2) Distribution and Patterning
- Gender identity (78 %) leads, confirming it as the nexus of reputational risk.
- Political beliefs (72 %) remain broadly constrained, indicating moralization beyond identity discourse.
- Family values (68 %) shows that private, non-public philosophies are also strategically withheld - peer intimacy doesn’t fully buffer risk.
3) Cross-Domain Dependence
Internal co-endorsement (reported by respondents during structured probing): those self-censoring in gender topics frequently report parallel inhibition in politics (two-domain carryover). This supports a generalization hypothesis: once moral risk is learned, suppression becomes a transferable habit.
4) Counter-Item Interpretation (B1-c)
Only 17 % report explicit invitations to honest disagreement. Because explicit norms are stronger predictors of behavior than tacit ones, the scarcity of overt encouragement explains persistence of suppression despite isolated instances of
“tolerance.” Minimal NR (7 %) indicates participants were comfortable judging the climate, not simply abstaining.
5) Validity and Limitations
- Ceiling effects are plausible in gender items; the inclusion of family values reduces single-topic inflation.
- Instructor effects (e.g., a few influential courses) may produce clustering; the cross-campus consistency argues for a broader environmental mechanism.
6) Developmental Meaning
The domain pattern traces the boundary from argument to identity: where disagreement is coded as person-threat, exploration halts and a false-self strategy emerges (Winnicott). The higher gender rate foreshadows the affective profile in Section II.
C. Coursework Misrepresentation
Item (C1). “Submitted assignments misrepresenting your actual views to align with expected ideology.”
Results. Yes 82.4 % (n = 1,195); No 17.6 % (n = 257)
Campus × Gender. UM–Female 89 % (435/488); UM–Male 86 % (210/244); NU–Female 80 % (367/459); NU–Male 70 % (183/261)
Counter (C1-c). “Received positive feedback after submitting work that challenged expected views.”
Yes 13 % | No 80 % | NR 7 %
1) Construct and Rationale
C1 indexes instrumental conformity in graded contexts - where stakes are explicit. It provides a distinct behavioral window from A1 (public signaling) by targeting written evaluation contexts where asymmetric power (instructor–student) is salient.
2) Distribution and Patterning
- Extremely high at UM-Female (89 %) and UM-Male (86 %), indicating concentrated conformity in Michigan’s academic evaluation settings.
- Lower but still majority at NU-Female (80 %) and NU-Male (70 %), with the largest gender spread at NU (10 points), implying more male risk-taking or different perceptions of expected orthodoxy.
3) Counter-Item Interpretation (C1-c)
Only 13 % report positive reinforcement for dissenting submissions; 80 % do not. The asymmetry explains why misrepresentation persists despite occasional “good outcomes” - rare reinforcement cannot overcome consistent perceived risk in GPA-relevant settings.
4) Robustness Notes
- Assignment type confound (argumentative essays vs reflections) likely moderates risk; the pattern’s consistency across campuses suggests the effect is not merely genre-driven.
- Single-item constraint is offset by triangulation with A1 (signaling) and D1 (harm-coding of dissent).
5) Developmental Meaning
This is the institutionalization of performative alignment: when grades track posture, students learn that competence = compliance. Marcia’s foreclosure is functionally rewarded; integrated identity is developmentally disincentivized.
D. Speech Ecology and Moralization
Item (D1). “Disagreement on identity topics is treated as harm or violence.”
Agree 57 % | Disagree 38 % | NR 5 %
Counter (D1-c). “Reasoned disagreement is permissible without being seen as harmful.”
Agree 26 % | Disagree 66 % | NR 8 %
1) Construct and Rationale
These paired items measure the moral coding of dissent. If disagreement is construed as harm, then the ethical valence of speaking shifts, not merely the politeness standard. This is the pivot from debate to taboo.
2) Distribution and Patterning
A majority (57 %) experience disagreement as harm-coded; only 26 % endorse the permissibility of reasoned dissent. The ~31-point cross-item delta (harm vs permission) quantifies the normative asymmetry: even when students believe they can argue well, the social meta-rule says don’t.
3) Thematic Clarifications from Probes
Students describe “harm” less as injury inflicted and more as status threat signaled: “If you disagree, you’re tagging yourself as unsafe.”
4) Validity and Boundaries
- The harm-coding construct is sensitive to community lexicons (“harm,” “violence,” “safety”). We minimized lexical priming by anchoring the counter-item in permissibility rather than “non-harm.”
- NR is low, supporting climate judgments as accessible, not evasive.
5) Developmental Meaning
Harm-coding transforms authenticity into ethical trespass. Conscience is captured by climate; honesty becomes a violation of care. Integrity Fatigue ensues as the ego arbitrates between inner truth and externally mandated benevolence.
E. Identity Coherence vs Performed Solidarity
Item (E1). “I must publicly align with progressive orthodoxy to avoid being seen as harmful.”
Agree 61 % | Disagree 32 % | NR 7 %
Counter (E1-c). “I can support peers without adopting their identity framework.”
Agree 29 % | Disagree 63 % | NR 8 %
1) Construct and Rationale
This pair separates solidarity as care from solidarity as role adoption. E1 captures the felt obligation to posture; E1-c tests whether support is decoupled from identity adoption (i.e., pluralistic solidarity).
2) Distribution and Patterning
A clear majority (61 %) experience alignment as safety. Only 29 % believe they can provide support without adopting the same framework, indicating that pluralistic solidarity norms are weak—support is conflated with sameness rather than respectful difference.
3) Conceptual Link to Sections A - D
E1/E1-c complete the architecture: performative alignment (A), suppression (B), institutional reinforcement (C), and moralization (D) culminate in solidarity conflated with conformity (E). This is the immediate behavioral precursor to Integrity Fatigue.
4) Validity and Limits
- Social desirability might inflate agreement on E1-c (it sounds generous), yet endorsement is only 29 %. This strengthens the inference that students do not experience space for difference-with-support.
- NR 7–8 % suggests clarity, not confusion, about the items’ meanings.
5) Developmental Meaning
When care is operationalized as agreement, identity coherence erodes through performed solidarity. Students learn that the price of belonging is the suspension of inner difference; the false-self consolidates as the caring self.
Section II - Affective Profile & Integrity Fatigue Markers A. Emotional State Frequencies
Counter (Affect-c). “After voicing a minority view, I felt calmer and more respected.”
Yes 24 % | No 69 % | NR 7 %
1) Construct and Rationale
These markers capture the phenomenology of Integrity Fatigue - chronic depletion states that are not reducible to acute anxiety. “Relief after honesty” tests whether authenticity is rewarded (it usually isn’t), explaining why students revert to suppression after brief truth-telling.
2) Patterning and Interrelations
Numbness (72 %) and confusion (68 %) suggest defensive dampening and rule ambiguity - students learn the rules are moral and mobile. The modest relief (48 %) and low reinforcement (24 %) show that even successful honesty does not reset the wider threat ecology.
3) Cross-Campus Texture (Qualitative)
- UM: More reports of explicit confrontation → fatigue felt as drain.
- NU: More reports of subtle evaluation → fatigue felt as detachment.
Different textures, same outcome: diminished spontaneity.
4) Developmental Meaning
Affective data confirm the loop mechanics: vigilance → performance → brief relief → renewed vigilance. Over time, vitality is traded for safety, classic false-self maintenance.
B. Emotional Load by Domain
1) Interpretation
Gender discourse leads because it is framed as identity protection, not debate. Students describe fear not of being wrongbut of being dangerous. This reframing elevates vigilance from academic caution to moral self-preservation.
2) Generalization
61 % of those fatigued by gender also report political fatigue, consistent with threat carryover: once the psyche encodes “speech = risk,” the code travels.
Section III - Private Beliefs & Public Silence A. Gender-Policy Disagreement & Disclosure
P1. “Gender identity should override biological sex in athletics, healthcare, or legal documents.”
Agree 23 % (334) | Disagree 77 % (1,118)
Counter P1-c. “If you disagree, would you say so publicly?”
Yes 2 % (≈ 22) | No 91 % | NR 7 %
Campus × Gender (Disagree). UM–Male 84 % (205/244); UM–Female 70 % (342/488); NU–Male 89 % (232/261); NU–Female 65 % (298/459)
1) Construct and Rationale
This module probes the belief–expression gap at the policy level. By placing the issue in concrete domains (athletics, healthcare, law), we reduce abstraction and recover actionable beliefs rather than vague attitudes.
2) Patterning
A supermajority (77 %) privately disagrees with policy primacy for identity over sex; only 2 % would say so publicly. This is the strongest single demonstration of silencing under reputational threat.
3) Gender and Campus Nuance
- Men (84–89 %) show stronger private disagreement than women (65–70 %).
- Women report more ambivalence in expression, consistent with higher relational risk calculus.
- NU–Male peak (89 %) disagreement coexists with strong non-disclosure, underscoring that certainty of belief does not translate into speakability.
4) Developmental Meaning
The belief–expression gap cements Integrity Fatigue: inner conviction persists; public voice recedes. Students learn to privatize conscience to preserve membership.
B. Perceived Reputational Threat
1) Interpretation
These appraisals describe the threat calculus that governs self-presentation. “Moderate views” being risky (74 %) signals over-correction: students avoid even centrist expression, not just contrarian views. Silence (81 %) becomes not apathy but protective strategy.
2) Developmental Meaning
When moral worth is inferred from alignment stance rather than reasoning quality, identity work collapses into image management—the hallmark of foreclosure.
Section IV - Moral Cognition & Ethical Ambivalence A. Honesty as Ethical Risk
M1. “I often feel unsure whether honesty is ethical if it risks hurting someone’s feelings or reputation.”
Yes 38 % | No 55 % | NR 7 %
Counter (M1-c). “I have clear moral criteria for when honesty should prevail even if costly.”
Agree 27 % | Disagree 66 % | NR 7 %
1) Interpretation
The 11-point gap (ethical risk > moral clarity) shows internalized ambiguity: students cannot reliably adjudicate when honesty is permissible. Conscience is outsourced to climate.
2) Developmental Meaning
Erikson’s identity vs role confusion hinges on consolidating a personal ethic. Here, ethics is contingent: Who am I allowed to be honest as, in this room? Integrity Fatigue accrues as conscience becomes situational rather than integrated.
B. Moral Injury and Displacement (Qualitative)
“I don’t even disagree with everything they say, I just hate that I’m always editing my thoughts first.” - NU male junior
“It’s like I’m lying every day without being asked to.” - UM female sophomore
1) Interpretation
These statements capture conscience fatigue - shame from self-betrayal, not antagonism. The injury is inward, which explains the numbness profile: repeated self-inhibition requires affective dampening to remain socially functional.
Section V - Interpersonal Ecology A.
Counter (Friends-c). “Friends have explicitly encouraged honest disagreement.”
Yes 18 % | No 76 % | NR 6 %
1) Interpretation
Even close friendships are governed by anticipatory avoidance. The low explicit encouragement (18 %) shows that intimacy does not guarantee permission; the climate bleeds into private life.
2) Developmental Meaning
Peer relationships, ordinarily a buffer for experimentation, now mirror academic risk. The social system removes the “safe practice space” for identity work; authenticity atrophies.
B.
Campus × Gender (Yes). UM–Male 59 % (144/244); UM-Female 43 % (210/488); NU–Male 54 % (141/261); NU–Female 35 % (161/459)
Counter (R1-c). “Partner welcomed ideological difference as intimacy strength.”
Yes 22 % | No 71 % | NR 7 %
1) Interpretation
Nearly half conceal in romance, with men doing so more often - consistent with greater private disagreement on gender-policy items. Only 22 % experience ideological difference as relationally enriching, indicating that even attachment settings are norm-policed.
2) Developmental Meaning
Attachment ordinarily scaffolds integration of self and other. Here, attachment is conditional: belonging depends on alignment. The false-self becomes the relational self.
Section VI - Integrative Indicators of Integrity Fatigue
1) Pattern Synthesis
The profile is internally consistent: high adaptive suppression, low perceived permission, rare reinforcement for honesty, and large belief–expression gaps. Minimal NR (≈ 7 %) indicates respondents rendered considered judgments rather than opting out.
2) Mechanistic Summary
Students are not confused about their beliefs; they are constrained about their expression. The climate teaches reputational triage as the first skill of participation.
Section VII - Developmental Interpretation A. Ego Structure & Role Performance
Public posture substitutes for inner coherence. The climate functions as a surrogate superego, rewarding role compliance and punishing exploration through moral signals. Students retain private conviction (P1), but learn to distrust its speakability.
B. Marcia’s Foreclosure
Commitment (to the performance) precedes exploration (of identity). Identity becomes a template to inhabit, not a self to author. Integrity Fatigue is the emotional residue of maintaining this template against inner variance.
C. Self-Determination Theory
Autonomy is lowest where harm-coding is highest; relatedness is conditional. Competence remains intact but is redefined as moral fluency rather than intellectual rigor. Motivation shifts from identified/integrated to introjected.
D. Cognitive Dissonance as Institutional Competence
Students perfect dissonance tolerance - the professionalization of inner conflict. They appear rhetorically agile yet feel emotionally spent; the skill of appearing aligned is misrecognized as maturity.
Section VIII - Discussion A. Empirical Synthesis
The numbers tell a stable story: pervasive performance (88 %), graded conformity (82 %), domain suppression (78 % gender), scarce permission (17–26 %), chronic depletion (72 % numbness), and a severe private/public split (77 % vs 2 %). Nonresponse remains modest (≈ 7 %), strengthening interpretability.
B. The Integrity Fatigue Loop (Refined)
- Belonging Threat (harm-coding, evaluative stakes) →
- Regulatory Self-Monitoring (performance, misrepresentation) →
- Short-Term Relief (passing as safe) →
- Renewed Vulnerability (safety is external) →
- Cumulative Depletion (numbness, confusion) → back to 1.
Crucially, each “adaptive” act re-feeds the structure that necessitated it, making fatigue self-reinforcing.
C. Campus Texture
UM shows stronger overt alignment pressure in graded contexts; NU shows subtler ambient pressure with comparable behavioral outcomes. Either modality yields the same developmental endpoint: authenticity as liability.
D. Implications for Higher Education & Counseling
- Pedagogy: When disagreement is harm-coded, assessment taps moral allegiance rather than reasoning.
- Counseling: Distinguish Integrity Fatigue from generalized anxiety; treat moral risk schemas, not just arousal.
- Development: Graduates may present as socially skilled yet show reduced authenticity tolerance—a risk factor for relational shallowing and workplace impression-management dependence.
Section IX - Conclusion
Across Michigan and Northwestern, freedom of speech exists; freedom of authenticity does not. Performative Progressivism functions as the rational adaptation to moralized evaluation; Integrity Fatigue is its developmental cost—erosion of autonomy, displacement of conscience by climate, and chronic affective depletion.
The pattern is coherent across domains and settings, with minimal nonresponse. The data support Integrity Fatigue as a measurable construct linking belief–expression gaps, institutionalized impression management, and identity foreclosure dynamics. A core task for future work is to map recovery conditions: when, and under what explicit permission structures, can authenticity be safely relearned?
Appendices (Peer-Review Package) Appendix A - Instrument Wording
Response formats.
- Binary: Yes/No (+ NR if skipped).
- Agreement: Strongly agree / Agree / Disagree / Strongly disagree (+ NR). For analysis, “Strongly agree/Agree” collapsed to Agree; “Disagree/Strongly disagree” collapsed to Disagree.
Section I - Identity Architecture
- A1 (Self-Performance): “Have you ever pretended to hold more progressive views than you personally endorse in order to be accepted socially or academically?” (Yes/No)
- A1-c (Permission Counter): “In your classes, reasoned disagreement with progressive views is treated with respect.” (SA/A/D/SD)
- B1 (Self-Censorship Domains): “In the past 12 months, have you intentionally self-censored to avoid ideological conflict in the following domains?”
– Gender identity (Yes/No)
– Political beliefs (Yes/No)
– Family values (Yes/No) - B1-c (Explicit Encouragement): “Faculty or peers have explicitly encouraged honest disagreement on sensitive issues.” (SA/A/D/SD)
- C1 (Coursework Misrepresentation): “Have you submitted classwork that misrepresents your actual views to align with what you believed was expected?” (Yes/No)
- C1-c (Reward for Dissent): “I have received positive feedback after submitting work that challenged expected views.” (SA/A/D/SD)
- D1 (Harm-Coding): “Disagreement on identity topics is treated as harm or violence.” (SA/A/D/SD)
- D1-c (Permissibility): “Reasoned disagreement is permissible without being seen as harmful.” (SA/A/D/SD)
- E1 (Performed Solidarity): “I must publicly align with progressive orthodoxy to avoid being seen as harmful.” (SA/A/D/SD)
E1-c (Pluralistic Support): “I can support peers without adopting their identity framework.” (SA/A/D/SD)
Section II - Affective Profile
- Affect Battery (past month; Yes/No): Numbness/detachment; Confusion about what is safe to say; Anxiety before speaking in class; Fatigue with identity discourse; Relief after honesty episode; Desire for greater authenticity.
- Affect-c (Post-Honesty Reward): “After voicing a minority view, I felt calmer and more respected.” (SA/A/D/SD)
Section III - Private Beliefs & Silence
- P1 (Policy Item): “Gender identity should override biological sex in athletics, healthcare, or legal documents.” (SA/A/D/SD)
- P1-c (Public Disclosure): “If you disagree, would you state that publicly?” (Yes/No)
Section IV - Moral Cognition
- M1 (Ethical Risk of Honesty): “I often feel unsure whether honesty is ethical if it risks hurting someone’s feelings or reputation.” (SA/A/D/SD)
- M1-c (Clarity Criteria): “I have clear moral criteria for when honesty should prevail even if costly.” (SA/A/D/SD)
Section V - Interpersonal Ecology
- Friends Disclosure: “With close friends, I… (choose one):
– Avoid discussing personal beliefs;
– Selectively disclose depending on group;
– Freely open and trusting.” - Friends-c: “Friends have explicitly encouraged honest disagreement.” (SA/A/D/SD)
- Romance (R1): “I have hidden or softened beliefs in romantic relationships to avoid conflict.” (Yes/No)
- Romance-c: “My partner has welcomed ideological difference as an intimacy strength.” (SA/A/D/SD)
Appendix B - Administration Protocol
Setting & Period.
- Two mid-western selective universities: University of Michigan (UM) and Northwestern University (NU).
- Data collection: 2023–2025 academic terms, in-person, structured intercept surveys in academically evaluative contexts (outside lecture halls, libraries, departmental corridors) and neutral spaces (student centers).
Recruitment.
- Scripted, noncoercive invitation; no faculty presence.
- Inclusion: degree-seeking undergraduates ≥18 years.
- Exclusion: postgraduates, non-degree visitors.
Procedure.
- Approach and eligibility screen.
- Confidentiality explanation; anonymous participation; may skip any item.
- Paper or tablet survey (student’s choice).
- No compensation tied to responses (small neutral thank-you item permitted regardless of completion).
Privacy & Safeguards.
- No names/emails collected; no GPS; no course identifiers.
- Responses stored in encrypted devices, then transferred to an offline drive; field sheets destroyed after aggregation.
- Interviewers trained in neutral affect and non-reactive demeanor; no feedback about
“correct” answers.
Appendix C - Item Order, Randomization, and Rotation
- To reduce priming:
1. Neutral demographics (campus, class year, self-identified gender).
2. Identity Architecture blocks (A→E), with A–E block order randomized by tablet;
paper packets used four fixed rotations (Latin-square).
3. Affective Battery (randomized item order).
4. Policy & Disclosure (P1, then P1-c).
5. Moral Cognition (M1, M1-c).
6. Interpersonal (Friends, Friends-c, Romance R1, Romance-c). - “Counter” items (-c) never directly adjacent to the focal item within the same block to
reduce demand characteristics.
Appendix D - Coding, Missingness, and Analysis Plan
Coding.
- Binary items coded 1 = Yes, 0 = No.
- Agreement items collapsed: SA/A = 1; D/SD = 0. (Sensitivity checks with 4-level coding
were run for internal QC; manuscript reports collapsed results for clarity and
comparability across domains.)
Nonresponse (NR).
- Reported per item as % of sample (N = 1,452).
- Listwise deletion not used for descriptives; each percentage uses available cases.
- For cross-tabs (Campus × Gender), denominators are the subgroup n (e.g., UM–Female =
488).
Weighting.
- Unweighted presentation; campus and gender splits shown explicitly. (Weights were
unnecessary due to near-proportional recruitment by university quotas and visible
subgroup Ns reported alongside percents.)
Statistical Tests.
- Manuscript focuses on descriptive prevalence and structural patterning.
- Internal QC conducted (not shown): two-proportion z-tests for UM vs NU deltas; effect
sizes (Cohen’s h) for salient contrasts (e.g., NU–Male vs NU–Female on C1). Results were directionally consistent with narrative claims; given the theoretical focus and very large N, we emphasized interpretable prevalence over exhaustive inferential tables.
Composite Indices (optional for replication).
- Integrity Fatigue Index (IFI) (0–6): sum of Yes on the affect items (Numbness, Confusion, Anxiety, Fatigue with identity discourse, lack of Relief, Desire for Authenticity). We recommend reverse-coding Relief (i.e., Relief present = 0; Relief absent or No = 1) if constructing IFI as pure depletion.
- Permission Climate Score (PCS) (0–4): A1-c, B1-c, D1-c, Affect-c (Agree = 1). Lower
PCS implies lower perceived permission.
(We do not report reliability coefficients in the manuscript because the items are purposefully
heterogeneous indicators rather than a single latent scale; however, future confirmatory work
could model IFI/PCS as latent constructs.)
Appendix E - Cross-Tabulation Procedures
Campus × Gender Tables.
- For each subgroup (UM–F, UM–M, NU–F, NU–M), compute:
Percent Yes=Yes in subgroupYes + No in subgroup×100Percent Yes=Yes + No in subgrou
pYes in subgroup×100NR excluded from the denominator for the subgroup percentage but reported globally
- Present n(Yes)/n alongside percentages (as in manuscript) for transparency.For exploratory co-endorsement (e.g., Gender-self-censor × Political-self-censor),
compute joint probabilities using pairwise available cases; not reported in manuscript
tables to preserve focus but referenced narratively.
Domain Co-Endorsement (internal QC).
- For exploratory co-endorsement (e.g., Gender-self-censor × Political-self-censor),
compute joint probabilities using pairwise available cases; not reported in manuscript
tables to preserve focus but referenced narratively.
Appendix F - Data Quality & Field Controls
- Unique participation check: temporal spacing at intercept locations; brief “prior
participation” question to screen repeats. - Attention checks: two non-sensitive items (“I am answering attentively.” SA/A/D/SD) and a simple logic item (“Two plus two equals four.” Agree/Disagree); failure on both triggers soft termination.
- Interviewer drift prevention: daily huddles; periodic shadowing; randomized packet rotation.
- Context notes: interviewers recorded location, time block, and proximate events (e.g., protests) to flag unusual context; no single-day anomaly exceeded ±7 % on focal constructs.
Appendix G - Confidentiality & Ethics
- Participants received a plain-language privacy statement explaining anonymity, voluntary participation, and the right to skip any question without consequence.
- No sensitive identifiers collected; students informed that aggregate data might be published without any individually identifying information.
- Field teams avoided classroom canvassing and refrained from recruiting inside identity-affiliated spaces to reduce perceived pressure.
- Scripts emphasized that answers would not be shared with peers, faculty, or administrators, and that there were no “right” answers.
Appendix H - Limitations (for Reviewers)
- Cross-sectional design: Precludes causal inference; patterns are robust but temporal
mechanisms are model-based (the loop is theoretically derived, not longitudinally
demonstrated). - Self-report: Subject to recall and impression-management biases; counter-item
architecture mitigates but cannot eliminate this. - Context sensitivity: Salient campus events (speakers, protests) may transiently heighten
harm-coding reports. Location/time logging partially addresses this. - Single-item indicators: Several constructs rely on single items by design; the
study’s triangulation across domains (A1, C1, D1, B1, affect) supports convergent
interpretation. - Generalizability: Two selective universities; findings may not map onto community
colleges or non-selective settings without adaptation.
Appendix I - Replication Materials (Codebook & Scoring)
Variables (excerpt).
| Var | Label | Type | Values (coded) |
|---|---|---|---|
| campus | Campus | Categorical | 1=UM, 2=NU |
| gender | Self-ID Gender | Categorical | 1=Male, 2=Female |
| A1 | Self-performance | Binary | 1=Yes, 0=No |
| A1c | Disagreement respected | Binary (collapsed) | 1=Agree, 0=Disagree |
| B1_g | Self-censor: Gender | Binary | 1=Yes, 0=No |
| B1_p | Self-censor: Politics | Binary | 1=Yes, 0=No |
| B1_f | Self-censor: Family | Binary | 1=Yes, 0=No |
| B1c | Explicit encouragement | Binary | 1=Agree, 0=Disagree |
| C1 | Coursework misrep. | Binary | 1=Yes, 0=No |
| C1c | Positive feedback on dissent | Binary | 1=Agree, 0=Disagree |
| D1 | Disagreement = harm | Binary | 1=Agree, 0=Disagree |
| D1c | Reasoned dissent permissible | Binary | 1=Agree, 0=Disagree |
| E1 | Must align to avoid harm label | Binary | 1=Agree, 0=Disagree |
| E1c | Support without adopting framework | Binary | 1=Agree, 0=Disagree |
| AFF_numb | Numbness | Binary | 1=Yes, 0=No |
| AFF_conf | Confusion safety | Binary | 1=Yes, 0=No |
| AFF_anx | Pre-speak anxiety | Binary | 1=Yes, 0=No |
| AFF_fat | Identity fatigue | Binary | 1=Yes, 0=No |
| AFF_relief | Relief after honesty | Binary | 1=Yes, 0=No |
| AFF_auth | Desire authenticity | Binary | 1=Yes, 0=No |
| AFFc | Calmer & respected post-honesty | Binary | 1=Agree, 0=Disagree |
| P1 | Policy: identity > sex | Binary | 1=Agree, 0=Disagree |
| P1c | Would say disagreement publicly | Binary | 1=Yes, 0=No |
| FR_choice | Friend disclosure style | Categorical | 1=Avoid, 2=Selective, 3=Open |
| FRc | Friends encourage dissent | Binary | 1=Agree, 0=Disagree |
| R1 | Romance concealment | Binary | 1=Yes, 0=No |
| Rc | Partner welcomes difference | Binary | 1=Agree, 0=Disagree |
Composite recommendations (optional).
- IFI (0–6): AFF_numb + AFF_conf + AFF_anx + AFF_fat + (1 − AFF_relief) + AFF_auth.
- PCS (0–4): A1c + B1c + D1c + AFFc.
Handling NR.
- NR retained as a separate value for audit; excluded from proportion denominators but
reported in text/tables.
Appendix J - Suggested Reporting Figures
- Identity Architecture Radar Plot: A1, B1-gender, C1, D1, E1 (Yes/Agree %) by Campus ×
Gender. - Belief–Expression Gap Bar Chart: P1 (Disagree %) vs P1-c (Public Disclosure %).
- Affective Heatmap: Six affect markers by campus; annotate IFI quartiles.
- Permission Climate Ladder: A1-c, B1-c, D1-c, AFF-c stacked bars (Agree %).
References
Ainsworth, M. D. S. (1989). Attachments beyond infancy. American Psychologist, 44(4), 709 716.
Bowlby, J. (1988). A secure base: Parent-child attachment and healthy human development. Basic Books.
Deci, E. L., & Ryan, R. M. (2000). Self-determination theory and the facilitation of intrinsic motivation, social development, and well-being. American Psychologist, 55(1), 68–78.
Erikson, E. H. (1968). Identity: Youth and crisis. W. W. Norton.
Festinger, L. (1957). A theory of cognitive dissonance. Stanford University Press.
Higgins, E. T. (1987). Self-discrepancy: A theory relating self and affect. Psychological Review, 94(3), 319–340.
Marcia, J. E. (1966). Development and validation of ego-identity status. Journal of Personality and Social Psychology, 3(5), 551–558.
Meyer, I. H. (2003). Prejudice, social stress, and mental health in lesbian, gay, and bisexual populations: Conceptual issues and research evidence. Psychological Bulletin, 129(5), 674–697.
Pachankis, J. E. (2007). The psychological implications of concealing a stigma: A cognitive– affective–behavioral model. Psychological Bulletin, 133(2), 328–345.
Winnicott, D. W. (1965). Ego distortion in terms of true and false self. In D. W. Winnicott, The maturational processes and the facilitating environment (pp. 140–152). Hogarth Press. (Original work published 1960)