Level of Expressed Emotion (LEE)

Abstract

The Level of Expressed Emotion (LEE) scale is a specialized psychometric self-report instrument developed by John D. Cole and Shahe S. Kazarian in 1988 to evaluate perceived expressed emotion (EE) within an individual's immediate interpersonal environment. Derived as an economical, psychometrically robust alternative to the resource-intensive Camberwell Family Interview (CFI), the LEE assesses the degree to which an individual experiences their significant other or family members as critical, hostile, overinvolved, or unsupportive. Comprising 60 dichotomous (True/False) items, the instrument measures four correlated yet distinct clinical dimensions: Intrusiveness, Emotional Response, Attitude toward Illness, and Tolerance/Expectations. Psychometric investigations consistently demonstrate strong internal consistency across clinical and non-clinical samples, with full-scale Cronbach's alpha coefficients typically exceeding .88 and subscale alphas ranging between .72 and .87. Test-retest reliability estimates across multi-week intervals demonstrate stability coefficients between .78 and .85. The instrument exhibits exceptional predictive validity, prospectively forecasting psychiatric relapse and rehospitalization rates in patients diagnosed with schizophrenia, mood disorders, and eating disorders across 1-, 2-, and 5-year longitudinal follow-up cohorts. Cross-cultural adaptations, including Dutch, Spanish, and adolescent versions, support the factorial stability of its core components while offering crucial empirical insights into the systemic dynamics that moderate course, recovery, and vulnerability in chronic psychological and somatic conditions.

Keywords

Level of Expressed Emotion, LEE, Expressed Emotion, Schizophrenia Relapse, Interpersonal Environment, Intrusiveness, Emotional Response, Psychometrics, Camberwell Family Interview, Family Functioning

Authors

The Level of Expressed Emotion (LEE) scale was conceived, constructed, and psychometrically standardized by:

• John D. Cole, Ph.D.: Department of Psychology, University of Western Ontario, and associated psychiatric hospital systems, London, Ontario, Canada. Specializing in clinical psychometrics, family environmental factors, and longitudinal predictors of serious mental illness.
• Shahe S. Kazarian, Ph.D.: Department of Psychology, University of Western Ontario, London, Ontario, Canada; later Professor of Psychology at the American University of Beirut. Renowned authority on clinical psychology, cross-cultural psychopathology, family emotional climates, and the operationalization of systemic stress factors in psychiatric populations.

Purpose

The primary clinical and empirical purpose of the Level of Expressed Emotion (LEE) scale is to quantify an individual's subjective perception of the emotional environment established by a key significant other (such as a parent, spouse, or primary caregiver). Historically, expressed emotion was evaluated via the Camberwell Family Interview (CFI), a semi-structured clinician-administered audiotaped interview that required approximately one to two hours to conduct and three to four hours to code by certified, trained raters. While the CFI remains an established diagnostic reference in psychiatric research, its profound clinical inefficiency, training costs, and reliance on observer inference spurred the development of self-report analogues. The LEE was systematically designed to resolve these methodological bottlenecks, allowing investigators and clinicians to capture the family affective climate rapidly and directly from the patient’s experiential perspective.

From an applied clinical perspective, high levels of expressed emotion represent one of the most reliable environmental predictors of symptomatic relapse in chronic psychiatric illnesses, particularly schizophrenia, bipolar disorder, and major depressive disorder. When individuals transition from inpatient psychiatric stabilization back into home environments characterized by pervasive hostility, relentless criticism, or suffocating emotional overinvolvement, their neurobiological stress systems undergo chronic activation, markedly elevating the risk of decompensation. The LEE provides mental health professionals with a granular, 60-item diagnostic tool that pinpoints the exact operational dimensions of familial stress. Rather than treating “family toxicity” as a monolithic, ambiguous construct, the LEE disentangles whether the primary stressor stems from structural boundary violations (Intrusiveness), catastrophic emotional reactivity (Emotional Response), dismissive or stigmatizing cognitive appraisals (Attitude toward Illness), or perfectionistic, inflexible behavioral standards (Tolerance/Expectations).

In addition to psychiatric applications, the LEE has garnered substantial utility in behavioral medicine and eating disorder clinics. In anorexia nervosa and bulimia nervosa, the scale informs family-based therapy protocols by illustrating whether caregivers inadvertently maintain symptoms through hyper-surveillance or punitive blame. Furthermore, the scale functions as an empirical outcome metric for family psychoeducation and systemic family therapy, tracking whether targeted psychosocial interventions successfully down-regulate emotional tension, cultivate genuine empathy, and restore healthy interpersonal boundaries across longitudinal phases of care.

Psychological Construct

Expressed Emotion (EE) is a multifaceted psychosocial construct conceptualizing the overarching affective and behavioral climate within an interpersonal dyad, typically involving an individual diagnosed with a clinical disorder and their influential caregiving figures. The Level of Expressed Emotion (LEE) scale models this construct across four distinct, clinically validated subscales:

1. Intrusiveness

The Intrusiveness dimension measures boundary dissolution, overprotection, behavioral micro-management, and uninvited meddling in personal autonomy. Operationally, it captures the extent to which the significant other fails to acknowledge psychological, physical, and communicative individuation. Manifestations include reading private mail, entering private spaces without permission, monitoring whereabouts obsessively, interrupting private conversations, and persistently demanding that activities be conducted collectively. In high-EE environments, intrusiveness generates pervasive feelings of infantilization, surveillance, and loss of agency in the patient, depriving them of the mastery experiences required for psychosocial recovery.

2. Emotional Response

This subscale assesses the caregiver’s affective reactivity, volatility, panic, and capacity for self-regulation when confronted with stress, crisis, or symptomatic fluctuations. Items capture whether the relative loses their temper, panics during minor setbacks, reacts with hysteria or dramatic anxiety, and communicates personal fragility. Rather than providing a steadying, emotionally regulating presence, high-scoring relatives amplify systemic turmoil, shifting the emotional burden back onto the vulnerable individual. This dynamic forces the patient into emotional accommodation or leads to chronic hyperarousal, directly destabilizing the patient’s autonomic nervous system.

3. Attitude toward Illness

Attitude toward Illness assesses the significant other's cognitive attributions, empathy, and acceptance regarding the patient’s symptoms, limitations, and help-seeking behaviors. In low-EE contexts, caregivers understand limitations, validate distress, and support clinical intervention. Conversely, high-EE relatives exhibit hostile, invalidating, and stigmatizing attributions: they perceive symptoms as deliberate manipulation (“says I just want attention”), accuse the individual of feigning incapacity, view treatment-seeking as weakness, or hold the patient morally culpable for their distress. This critical, accusatory attributional style directly fosters demoralization, internalized stigma, and treatment non-compliance.

4. Tolerance/Expectations

The Tolerance/Expectations dimension operationalizes the rigidity of the behavioral standards imposed upon the patient, evaluating whether the relative demonstrates flexibility and compassion or enforces rigid, unrealistic developmental and functional milestones. It probes whether the relative expresses irritation when productivity wavers, puts the patient down for making mistakes, expects identical exertion during acute episodes, and withholds tolerance. When caregivers fail to adjust their functional expectations to accommodate clinical disability, the patient experiences chronic failure, intense guilt, and perceived rejection, triggering secondary depressogenic and anxious cascades.

Theoretical Framework

The Level of Expressed Emotion scale is grounded in the confluence of the Vulnerability-Stress-Coping Model of psychopathology, pioneered by Joseph Zubin and Bonnie Spring (1977), and the seminal interpersonal empirical paradigms established by George Brown, Michael Rutter, and Christine Vaughn at the Medical Research Council Social Psychiatry Unit in London during the 1960s and 1970s.

Within the Vulnerability-Stress-Coping framework, severe mental disorders (such as schizophrenia and recurrent affective psychoses) possess an underlying biological or neurodevelopmental diathesis. However, the phenotypic expression of symptoms, the trajectory of the disease, and the occurrence of acute psychotic episodes are heavily mediated by environmental stressors and individual coping mechanisms. The emotional climate of the home functions as an ongoing, pervasive socio-environmental stressor. Brown and colleagues observed that patients discharged from psychiatric facilities into households where relatives displayed high levels of criticism, hostility, or emotional over-involvement demonstrated relapse rates two to three times higher than those returning to calm, low-EE settings, even after controlling for baseline clinical severity and medication adherence.

From an attribution theory perspective (Weiner, 1985), high expressed emotion represents a breakdown in systemic cognitive processing. High-EE relatives tend to make internal, controllable, and stable attributions regarding the patient's symptomatic deficits (e.g., viewing negative symptoms such as avolition or social withdrawal as intentional laziness or spite), leading to emotional frustration, anger, and punitive interactions. Conversely, low-EE relatives generate external or uncontrollable attributions (e.g., viewing symptoms as manifestations of an authentic neurobiological disorder), fostering sympathy, patience, and adaptive scaffolding.

Furthermore, the LEE incorporates concepts from family systems theory and structural family models (Minuchin, 1974). Salvador Minuchin emphasized that healthy family functioning necessitates clear, flexible boundaries between subsystems. The Intrusiveness and Emotional Response components of the LEE reflect classic systemic pathologies: “enmeshment,” in which personal autonomy is suppressed in favor of intrusive togetherness, and systemic affective dysregulation, in which boundaries fail to protect vulnerable members from parental or partner distress.

Validity

The psychometric validity of the Level of Expressed Emotion scale has been extensively documented across psychiatric, community, and developmental samples over several decades.

Construct and Factorial Validity

During the initial development of the scale, Cole and Kazarian (1988) administered an extensive pool of potential items to both psychiatric inpatients and university undergraduates to determine if self-reported perception could reliably map onto the four theoretical pillars of expressed emotion. Confirmatory and exploratory analyses substantiated that the 60 items cleanly segregated into the four hypothesized dimensions (Intrusiveness, Emotional Response, Attitude toward Illness, and Tolerance/Expectations), with significant intercorrelations between subscales confirming that they tap a unified overarching construct while preserving meaningful discriminant validity between dimensions.

Predictive Validity

The most compelling psychometric evidence for the LEE lies in its robust predictive validity regarding psychiatric outcomes. In a seminal longitudinal investigation, Cole and Kazarian (1993) followed psychiatric cohorts across 1-year, 2-year, and 5-year post-discharge intervals. Patients who rated their significant others as demonstrating high EE on the LEE at index discharge exhibited significantly faster and more frequent rehospitalization rates than those rating their environment as low in EE. This longitudinal predictive capability matched the historical benchmarks set by the clinician-rated Camberwell Family Interview, verifying that the patient’s own phenomenological experience of their familial environment is as clinically predictive of biological decompensation as expert observer ratings.

Convergent and Discriminant Validity

The LEE displays strong convergent validity when benchmarked against established measures of interpersonal conflict, family dysfunction, and relational stress. Significant positive correlations have been observed with the Family Assessment Device (FAD), the Family Environment Scale (FES) Conflict subscale, and the Perceived Criticism Measure (PCM). Conversely, the scale demonstrates robust negative correlations with measures of perceived social support (e.g., the Multidimensional Scale of Perceived Social Support; MSPSS) and family cohesion. Demonstrating discriminant validity, the LEE correlates weakly with broad trait neuroticism and generalized psychiatric distress, demonstrating that it measures the specific interactive properties of the dyadic relationship rather than generalized negative affectivity or acute distress.

Reliability

The Level of Expressed Emotion scale demonstrates sound reliability metrics across diverse clinical populations, linguistic translations, and research designs.

Internal Consistency

In the foundational psychometric evaluations conducted by Cole and Kazarian (1988), the full 60-item instrument demonstrated high internal consistency, yielding a full-scale Cronbach's alpha of .91 among psychiatric samples and .89 among non-clinical populations. The internal consistencies of the individual subscales, which each contain 15 items, have been reported as follows:

• Intrusiveness: Cronbach's alpha ranging between .74 and .84.
• Emotional Response: Cronbach's alpha ranging between .78 and .86.
• Attitude toward Illness: Cronbach's alpha ranging between .71 and .82.
• Tolerance/Expectations: Cronbach's alpha ranging between .73 and .85.

Subsequent psychometric adaptations, such as the Spanish validation by Sepúlveda et al. (2012) among relatives of individuals with eating disorders and the Dutch adaptation by Gerlsma et al. (1992), have reported comparable internal reliability coefficients, with full-scale alphas consistently exceeding .85.

Test-Retest Stability

Test-retest reliability assessments conducted across stable psychiatric outpatients and non-clinical participants over intervals spanning two to six weeks have yielded stability coefficients ranging from r = .78 to r = .87 for the total score, with subscale test-retest coefficients maintaining levels between .72 and .82. These statistics demonstrate that the LEE measures stable relational environments while remaining sensitive to systemic shifts following therapeutic family interventions.

Factor Analysis

Extensive exploratory and confirmatory factor analyses have examined whether the theoretical four-factor structure of the LEE holds across varied clinical cohorts, age demographics, and cultural translations.

In the initial exploratory factor analysis (EFA) conducted by Cole and Kazarian (1988) using principal axis factoring with varimax and oblimin rotations, a distinct four-factor solution accounted for the major portion of common variance. The four extracted factors matched the theoretical model: Factor I represented Emotional Response (affective reactivity, panic, irritation), Factor II captured Intrusiveness (boundary incursions, lack of privacy), Factor III loaded heavily on Attitude toward Illness (understanding limitations, seeking help, accusations of faking), and Factor IV corresponded to Tolerance/Expectations (perfectionism, impatience, tolerance of mistakes). Item loadings on primary factors were predominantly moderate to high (ranging between .40 and .76), with minimal problematic cross-loadings across secondary dimensions.

Hale, Raaijmakers, Gerlsma, and Meeus (2007) investigated the structural invariance of the LEE across developmental cohorts, specifically testing whether the four-factor structure established in adults applied equally to adolescents. Employing multi-group confirmatory factor analysis (CFA), the researchers demonstrated acceptable goodness-of-fit indices (RMSEA < .06; CFI > .90; TLI > .90) for the four-factor model across both groups, establishing metric and scalar invariance. Similarly, Sepúlveda et al. (2012) conducted CFA on the Spanish version, confirming that the four core latent variables adequately model the interpersonal dynamic in eating disorders, though highlighting that shortened versions (e.g., 38-item or 24-item derivations) may provide streamlined fit indices in specific acute medical settings.

Instrument / Measurement Tool

The Level of Expressed Emotion scale is structured as follows:

• Instrument Type: Self-report psychological assessment / dyadic interpersonal environment scale.
• Respondent Perspective: Completed by the patient/client regarding a specific, identified significant other (e.g., mother, father, spouse, primary caregiver), or completed by relatives evaluating their own behavioral stance toward the patient (in caregiver-adapted versions).
• Administration Format: Available in paper-and-pencil or secure computer-based digital formats; self-administered or read aloud by an interviewer in clinical contexts where reading comprehension is compromised.
• Administration Time: Approximately 10 to 15 minutes.
• Total Number of Items: 60 items (divided equally into four 15-item subscales).
• Response Scale: Dichotomous forced-choice format: True or False.
• Subscale Composition:
  • Intrusiveness: Evaluates boundary violations and privacy interference.
  • Emotional Response: Evaluates affective volatility, panic, and distress tolerance.
  • Attitude toward Illness: Evaluates cognitive attributions, empathy, and recognition of clinical limitations.
  • Tolerance/Expectations: Evaluates standards of performance, patience, and criticism.
• Scoring Procedures:
  • Items are scored in the direction of high expressed emotion. An item indicating a high-EE behavior or attitude receives a score of 1 when marked “True”, and 0 when marked “False”.
  • Positively phrased, supportive, or tolerant items (e.g., “Calms me down when I'm upset”, “Can see my point of view”, “Understands my limitations”) are reverse-scored: answering “False” receives a score of 1, and answering “True” receives a score of 0.
  • Subscale scores are calculated by summing the keyed responses for each 15-item cluster (subscale range: 0 to 15).
  • A total LEE score is generated by summing all four subscale scores, yielding a continuous global EE score ranging from 0 to 60.
• Clinical Interpretation: Higher scores correspond to higher levels of perceived expressed emotion and greater interpersonal stress. Researchers and clinicians utilize either continuous scores or sample-based cutoffs (e.g., scores in the upper tercile or median splits) to stratify participants into “High-EE” versus “Low-EE” environments for clinical intervention and prognostic risk modeling.

Permissions & Fee and Test Year

The Level of Expressed Emotion (LEE) scale was formally developed and published in 1988 by John D. Cole and Shahe S. Kazarian in the Journal of Clinical Psychology. The instrument was subsequently reprinted and documented in major reference compilations, notably Joel Fischer and Kevin J. Corcoran's Measures for Clinical Practice and Research: A Sourcebook (Oxford University Press, 2007).

The scale was developed under academic research funding and is widely considered to reside in the academic public domain for non-commercial, clinical, educational, and empirical research applications. Researchers and practitioners may administer the instrument without paying royalty fees, provided that appropriate scholarly attribution and standard academic citations (Cole & Kazarian, 1988; Fischer & Corcoran, 2007) are maintained. Any commercial reproduction, inclusion in commercial software packages, or proprietary distribution must seek formal copyright authorization from the original authors and the corresponding journal copyright holders.

References

• Brown, G. W., Birley, J. L., & Wing, J. K. (1972). Influence of family life on the course of schizophrenic disorders: A replication. The British Journal of Psychiatry, 121(562), 241–258. https://doi.org/10.1192/bjp.121.3.241
• Cole, J. D., & Kazarian, S. S. (1988). The Level of Expressed Emotion scale: A new measure of expressed emotion. Journal of Clinical Psychology, 44(3), 392–397. 3.0.CO;2-J” target=”_blank” rel=”noopener noreferrer”>https://doi.org/10.1002/1097-4679(199303)49:2<216::AID-JCLP2270490212>3.0.CO;2-J
• Fischer, J., & Corcoran, K. J. (2007). Measures for clinical practice and research: A sourcebook (4th ed., Vol. 1, pp. 359–361). Oxford University Press.
• Gerlsma, C., Hale, W. W., Leeman, P., & Mersch, P. P. (1992). The Level of Expressed Emotion scale: A useful tool for clinical practice and research? Acta Psychiatrica Scandinavica, 85(4), 312–317. https://doi.org/10.1111/j.1600-0447.1992.tb01476.x
• Hale, W. W., Raaijmakers, Q. A. W., Gerlsma, C., & Meeus, W. (2007). Does the Level of Expressed Emotion (LEE) questionnaire have the same factor structure for adolescents as it has for adults? Social Psychiatry and Psychiatric Epidemiology, 42(3), 215–220. https://doi.org/10.1007/s00127-006-0150-1
• Minuchin, S. (1974). Families and family therapy. Harvard University Press.
• Sepúlveda, A. R., Anastasiadou, D., del Río, A. M., & Carrobles, J. A. (2012). The Spanish validation of Level of Expressed Emotion scale for relatives of people with eating disorders. The Spanish Journal of Psychology, 15(2), 825–839. https://doi.org/10.5209/rev_SJOP.2012.v15.n2.38892
• Vaughn, C. E., & Leff, J. P. (1976). The measurement of expressed emotion in the families of psychiatric patients. British Journal of Social and Clinical Psychology, 15(2), 157–165. https://doi.org/10.1111/j.2044-8260.1976.tb00021.x
• Weiner, B. (1985). An attributional theory of achievement motivation and emotion. Psychological Review, 92(4), 548–573. https://doi.org/10.1037/0033-295X.92.4.548
• Zubin, J., & Spring, B. (1977). Vulnerability: A new view of schizophrenia. Journal of Abnormal Psychology, 86(2), 103–126. https://doi.org/10.1037/0021-843X.86.2.103

Items of the Scale