UCLA Loneliness Scale – Version 3 (UCLA-LS3)

1. Abstract

The UCLA Loneliness Scale – Version 3 (UCLA-LS3) represents the gold standard self-report psychometric instrument designed to assess subjective loneliness and perceived social isolation. Developed by Daniel W. Russell in 1996, the scale builds upon its historical predecessors—the original 1978 UCLA Loneliness Scale and the 1980 Revised UCLA Loneliness Scale (Version 2)—by systematically simplifying linguistic syntax, reversing ambiguous phrasing, and eliminating double negatives. These refinements were enacted to enhance structural accessibility and maintain psychometric fidelity across diverse cohorts, particularly older adults, clinical populations, and individuals with lower literacy levels. Comprising 20 items, the instrument utilizes a balanced design consisting of 11 negatively worded (lonely) statements and 9 positively worded (non-lonely) statements, scored along a 4-point Likert scale ranging from 1 (“Never”) to 4 (“Always”). Total scores span from 20 to 80, with higher aggregates reflecting heightened subjective isolation and relational dissatisfaction.

Extensive psychometric investigations substantiate the instrument’s exemplary reliability and validity across heterogeneous developmental and demographic strata. Internal consistency coefficients consistently range between Cronbach’s α = .89 and .94 across college undergraduate, public school teacher, nurse, and community-dwelling elderly samples. Longitudinal test-retest evaluations over a 12-month interval substantiate temporal stability (r = .73). Structurally, confirmatory factor analytic studies support a bifactor model featuring a robust, overarching global loneliness dimension accompanied by two orthogonal methodological factors corresponding to item directionality, thereby attenuating acquiescence bias. Demonstrating robust convergent and discriminant validity, UCLA-LS3 scores correlate meaningfully with self-esteem, depressive symptoms, social support, and neurobiological stress markers while preserving discriminant divergence from general negative affectivity. Consequently, the UCLA-LS3 remains the most widely cited and psychometrically vetted instrument in behavioral medicine, affective epidemiology, social neuroscience, and clinical gerontology.

2. Keywords

UCLA Loneliness Scale, UCLA-LS3, Daniel W. Russell, loneliness assessment, social isolation, subjective isolation, psychometrics, relational connectedness, perceived social support, structural equation modeling, construct validity, bifactor model, scale validation.

3. Authors

The UCLA Loneliness Scale – Version 3 was developed and psychometrically standardized by Daniel W. Russell, Ph.D., an esteemed psychometrician and professor in the Department of Human Development and Family Studies at Iowa State University. Prior to his tenure at Iowa State, Dr. Russell served on the faculties of the University of California, Los Angeles (UCLA) and the University of Iowa, where he directed numerous large-scale epidemiological investigations examining the intersection of stress, social support, coping, and geriatric mental health.

The historical lineage of the UCLA Loneliness Scale series reflects a legacy of foundational scholarship conducted in collaboration with prominent social psychologists at UCLA, notably Letitia Anne Peplau, Ph.D. (Professor Emerita of Psychology at UCLA, a pioneer in relationship science and social psychology) and Mary Anne Ferguson, Ph.D. (formerly Cutrona), along with Carolyn E. Cutrona, Ph.D. (Professor of Psychology at Iowa State University). While the original 1978 scale (Russell, Peplau, & Ferguson) and the 1980 revised version (Russell, Peplau, & Cutrona) were collaborative endeavors, the Version 3 revision was authored exclusively by Daniel W. Russell (1996) to systematically address structural, cognitive, and linguistic limitations encountered when administering earlier iterations to non-collegiate adult and geriatric populations.

Primary Author Affiliation:
Daniel W. Russell, Ph.D.
Department of Human Development and Family Studies
College of Human Sciences, Iowa State University
Ames, Iowa, United States
Correspondence concerning the psychometric validation of Version 3 was originally archived through the Center for Health Services Research, College of Medicine, University of Iowa.

4. Purpose

The primary purpose of the UCLA Loneliness Scale – Version 3 is the quantitative operationalization and standardized assessment of subjective loneliness—defined conceptually as the distressing affective state that emerges when a discrepancy exists between an individual’s desired and perceived interpersonal relationships. Loneliness is fundamentally subjective; it does not represent an objective deficit in social network size, nor does it denote physical aloneness, solitude, or social contact frequency. Rather, the UCLA-LS3 captures the qualitative emotional distress stemming from perceived relational deficits, social unfulfillment, and perceived social alienation.

The methodological rationale necessitating Version 3 emerged directly from empirical difficulties encountered with the 1980 Revised UCLA Loneliness Scale (Version 2). Although Version 2 was psychometrically sound among college undergraduate cohorts, researchers attempting to deploy the instrument in community, clinical, and geriatric populations encountered severe measurement error driven by cognitive complexity. Specifically, Version 2 incorporated double-negative statements, convoluted conditional syntax, and ambiguous phrasing (e.g., “I do not feel that I am not part of a group”). Older adults and respondents with limited educational attainment exhibited elevated rates of missing data, misinterpretations, and response artifacting. Russell (1996) developed Version 3 to dramatically simplify the cognitive processing demands of the items, re-engineering statements to ensure universal readability while preserving the theoretical breadth of the construct.

In clinical practice, the UCLA-LS3 functions as an indispensable diagnostic and monitoring instrument. Severe loneliness has been empirically recognized as an independent risk factor for mortality, equivalent in hazard ratio to tobacco smoking, severe obesity, and physical inactivity. Clinicians deploy the instrument across psychiatric, primary care, and geriatric environments to screen for social detachment, evaluate the psychosocial sequelae of major life transitions (e.g., bereavement, chronic illness, retirement), and assess the efficacy of psychotherapeutic interventions such as Cognitive Behavioral Therapy (CBT) adapted for social anxiety, interpersonal psychotherapy, and community-based social prescribing initiatives.

Within empirical research, the UCLA-LS3 serves as the foundational metric across behavioral medicine, neurobiology, and epidemiology. Researchers utilize the scale to elucidate the pathophysiological pathways connecting perceived isolation to autonomic dysfunction, hypothalamic-pituitary-adrenal (HPA) axis hyperactivation, systemic inflammation (e.g., elevated C-reactive protein, interleukin-6), nocturnal sleep fragmentation, and accelerated cognitive decline in neurodegenerative diseases such as Alzheimer’s disease. By offering a standardized 20-item continuum, the scale facilitates cross-study meta-analyses, longitudinal health tracking, and the structural modeling of social determinants of health across diverse human lifespans.

5. Psychological Construct

The psychological construct assessed by the UCLA-LS3 is multidimensional in phenomenological experience yet captured as a pervasive psychological continuum. Russell’s psychometric architecture operationalizes loneliness according to the cognitive-discrepancy tradition established by Peplau and Perlman. Under this formulation, loneliness constitutes a negative subjective state resulting from perceived inadequacies across intimate, interpersonal, and collective domains of an individual’s social network.

Although Russell scored the instrument unidimensionally to reflect overall global loneliness, structural research identifies three central phenomenological facets embedded within the 20 items:

• Relational Connectedness (Intimate Others): This dimension encompasses feelings of emotional closeness, deep reciprocal understanding, and the presence of primary attachment figures. It reflects whether an individual feels fundamentally known, validated, and psychologically bonded to specific significant others. For example, Item 10 (“How often do you feel close to people?”), Item 13 (“How often do you feel that no one really knows you well?”), and Item 16 (“How often do you feel that there are people who really understand you?”) target the presence or absence of profound emotional intimacy. Deficits in this domain yield what Robert S. Weiss historically classified as “emotional loneliness”—the agonizing desolation stemming from the loss or lack of an attachment figure.
• Collective Connectedness (Social Others / Social Integration): This dimension measures the extent to which an individual feels integrated into a broader peer group, community, or social network that shares common values, worldviews, and recreational pursuits. Items capturing this construct include Item 5 (“How often do you feel part of a group of friends?”), Item 6 (“How often do you feel that you have a lot in common with the people around you?”), and Item 8 (“How often do you feel that your interests and ideas are not shared by those around you?”). These items address social affiliation and collective belonging; deficits here precipitate “social loneliness,” characterized by feelings of boredom, marginalization, and aimlessness.
• Perceived Social Isolation and Companionship Availability: This dimension taps into the acute perception of being physically and psychologically excluded, neglected, or devoid of social recourse during times of distress. Items such as Item 2 (“How often do you feel that you lack companionship?”), Item 3 (“How often do you feel that there is no one you can turn to?”), Item 4 (“How often do you feel alone?”), Item 11 (“How often do you feel left out?”), and Item 14 (“How often do you feel isolated from others?”) evaluate the profound awareness of isolation, abandonment, and the absence of a reliable social buffer.

Crucially, the construct evaluated by the UCLA-LS3 distinguishes itself fundamentally from objective solitude. Solitude is a voluntary, non-pathological state often associated with emotional restoration, creative contemplation, and autonomous self-regulation. In sharp contrast, loneliness measured by the UCLA-LS3 represents an unchosen, painful, and dysphorically experienced deficit accompanied by feelings of rejection, vulnerability, and hypervigilance toward social threat.

6. Theoretical Framework

The UCLA-LS3 is grounded in an interdisciplinary theoretical synthesis combining cognitive discrepancy models, interpersonal attachment theory, and evolutionary neurobiology.

Cognitive Discrepancy Perspective

The primary theoretical foundation of the instrument originates from the Cognitive Discrepancy Model of Loneliness formulated by Letitia Anne Peplau and Daniel Perlman (1982). According to this cognitive model, loneliness is neither an automatic consequence of living alone nor a direct function of social network size. Instead, it is mediated entirely through cognitive appraisal. The individual continuously compares their perceived level of social contact against their internalized standards or desired level of social contact. Loneliness is experienced precisely when a perceived mismatch occurs—either quantitatively (insufficient number of friends or interaction frequency) or qualitatively (relationships lack depth, reciprocal authenticity, or emotional responsiveness). The UCLA-LS3 measures this cognitive evaluation through relative self-appraisals (e.g., “How often do you feel that people are around you but not with you?”), where the respondent’s internal expectations govern their response rating.

Attachment Theory and Provisions of Social Relationships

The theoretical architecture also draws heavily upon John Bowlby’s Attachment Theory and the relational provisions framework delineated by sociologist Robert S. Weiss (1973). Weiss posited that healthy human functioning necessitates distinct relational provisions: attachment (provided by romantic partners or primary attachment figures), social integration (provided by networks of friends and colleagues), reassurance of worth, reliable alliance, and guidance. The UCLA-LS3 operationalizes both emotional loneliness (the severance of secure attachment bonds, reflected in items measuring closeness and being understood) and social loneliness (the failure to achieve social integration, reflected in items assessing companionship and shared group membership).

Evolutionary Theory of Loneliness (ETL)

Contemporary interpretations of the UCLA-LS3 construct are deeply aligned with the Evolutionary Theory of Loneliness articulated by John T. Cacioppo and Louise C. Hawkley. ETL conceptualizes loneliness as an evolved, adaptive biological drive analogous to physical pain, hunger, or thirst. Just as hunger signals a metabolic deficit requiring caloric intake to preserve organismic survival, loneliness operates as a social alarm signaling that the individual’s protective social perimeter has eroded, placing them at increased ancestral risk of predation and starvation. Consequently, high scores on the UCLA-LS3 reflect an organism trapped in an evolutionary state of social threat, characterized by implicit cognitive hypervigilance, autonomic arousal, elevated vascular resistance, and an altered transcriptional profile in immune cells (upregulation of pro-inflammatory genes and downregulation of antiviral genes).

7. Validity

The psychometric validity of the UCLA Loneliness Scale – Version 3 has been empirically verified across numerous developmental, occupational, and cross-cultural populations through rigorous assessment of construct, convergent, discriminant, predictive, and factorial validity parameters.

Construct and Convergent Validity

In his seminal validation study, Russell (1996) evaluated the scale across four diverse samples: college undergraduates (N = 487), public school teachers (N = 316), hospital nurses (N = 286), and community-dwelling older adults (N = 2,370). Convergent validity was robustly demonstrated via significant, substantial correlations with other validated measures of subjective well-being, psychological distress, and social relationships:

• Depressive Symptomatology: UCLA-LS3 scores correlated positively with depression as measured by the Center for Epidemiologic Studies Depression Scale (CES-D) and the Beck Depression Inventory (BDI), with coefficients typically ranging from r = .52 to .65 across cohorts.
• Self-Esteem: Consistent with theoretical models linking self-concept and interpersonal security, the scale exhibited strong negative correlations with the Rosenberg Self-Esteem Scale (r = -.49 to -.59).
• Social Support: Strong inverse correlations were recorded between the UCLA-LS3 and the Social Provisions Scale (ranging from r = -.51 to -.69), demonstrating that perceived inadequacy of social provisions directly tracks elevated loneliness.
• Anxiety and Neuroticism: UCLA-LS3 scores correlated moderately with trait anxiety (r = .38 to .46) and the Neuroticism domain of the NEO Personality Inventory (r = .40 to .48), confirming that individuals experiencing chronic loneliness also experience generalized affective distress.

Discriminant Validity

A critical psychometric hurdle in affective measurement involves demonstrating that loneliness does not represent a redundant proxy for depression, negative affectivity, or introversion. Russell (1996) deployed confirmatory structural equation modeling (SEM) to test the discriminant validity of the UCLA-LS3 against the CES-D and the Rosenberg Self-Esteem Scale. Modeling loneliness, depression, and self-esteem as distinct latent constructs yielded a significantly superior fit to the observed covariance matrix compared to single-factor or two-factor collapsed models across all sample populations (Δχ² tests, p < .001). Furthermore, partial correlation analyses revealed that when depressive symptomatology and self-esteem were statistically partialled out, UCLA-LS3 scores remained significantly predictive of social network deficits, objective social interaction frequency, and perceived social dissatisfaction.

Predictive and Criterion Validity

Substantial empirical literature documents the criterion and predictive validity of the UCLA-LS3 regarding psychological, somatic, and physiological outcomes:

• Morbidity and Mortality: In longitudinal epidemiological prospective cohorts (e.g., the Health and Retirement Study), elevated UCLA-LS3 scores independently predict cardiovascular morbidity, hypertension incidence, metabolic dysregulation, and all-cause mortality over 5- to 10-year follow-up intervals, even after controlling for socioeconomic status, baseline physical health, and clinical depression.
• Neuroendocrine and Immune Biomarkers: Higher scores on the scale predict elevated circulating levels of interleukin-6 (IL-6), elevated tumor necrosis factor-alpha (TNF-α), impaired glucocorticoid receptor sensitivity, and blunted immune response to viral inoculation.
• Cognitive Function: Among geriatric populations, elevated baseline UCLA-LS3 scores significantly predict accelerated rates of cognitive decline and doubled hazard ratios for clinical diagnosis of mild cognitive impairment (MCI) and Alzheimer’s disease.

8. Reliability

The UCLA-LS3 displays exceptionally high internal consistency and longitudinal stability across developmental stages, sociocultural environments, and clinical conditions.

Internal Consistency

In the primary psychometric validation conducted by Russell (1996), internal consistency estimates (Cronbach’s α) demonstrated high precision across all evaluated groups:

• College Undergraduates: α = .92
• Hospital Nurses: α = .94
• Public School Teachers: α = .89
• Community-Dwelling Elderly: α = .89

Item-total correlations across these cohorts ranged from .36 to .74, demonstrating that every individual item contributed meaningfully to the total score variance without redundancy. Subsequent independent psychometric investigations have replicated these high internal consistency metrics, with composite reliability (ρ_c) and McDonald’s omega (ω) values routinely exceeding .90, demonstrating minimal measurement error across the latent continuum.

Temporal Stability (Test-Retest Reliability)

Because loneliness encompasses both a baseline trait-like vulnerability and responsive state-like fluctuations driven by life circumstances, test-retest evaluations must account for observation windows. Russell (1996) evaluated temporal stability among community-dwelling older adults across a 12-month test-retest interval, reporting an impressive stability coefficient of r = .73. Shorter test-retest windows, such as 2-week to 4-week intervals utilized in clinical trial validation studies, yield reliability coefficients between r = .82 and .88, confirming high test-retest reliability during periods of contextual life stability while remaining sensitive to therapeutic or structural interventions.

Standard Error of Measurement

The Standard Error of Measurement (SEM) across standard populations ranges between 2.10 and 2.65 points on the 20-to-80 scale, yielding narrow confidence intervals for clinical decision-making and individual score interpretations.

9. Factor Analysis

The latent factor architecture of the UCLA Loneliness Scale has served as a central subject of psychometric modeling in personality assessment. While the scale was designed to provide a single composite score of global loneliness, linear factor analyses historically yielded conflicting structural conclusions due to methodological artifacts.

EFA and CFA Findings in Russell (1996)

Early exploratory factor analyses (EFA) on earlier versions often yielded two or three factors that split strictly along item wording: positively phrased items loaded onto one factor, while negatively phrased items loaded onto another. In his 1996 validation study, Russell subjected the UCLA-LS3 to rigorous confirmatory factor analysis (CFA) using structural equation modeling to adjudicate between competing theoretical specifications:

1. A strict Unidimensional Model (single bipolar loneliness construct).
2. A Two-Factor Correlated Model (reflecting loneliness vs. social connectedness).
3. A Three-Factor Correlated Model (based on Austin’s 1983 categorization of isolation, social others, and intimate others).
4. A Bifactor / Methodological Model comprising one substantive general loneliness factor and two orthogonal method factors (one representing positively worded items and the other representing negatively worded items).

Across all four samples analyzed by Russell (students, teachers, nurses, and older adults), the Bifactor / Methodological Model demonstrated superior, robust fit indices compared to all alternative configurations. Representative fit statistics for this bifactor model included:

• Goodness of Fit Index (GFI) ≥ .92
• Comparative Fit Index (CFI) ≥ .90 (ranging up to .95 in nurse and student samples)
• Root Mean Square Error of Approximation (RMSEA) ≤ .055 to .068

Russell concluded that the multidimensionality observed in standard factor extractions was an artifact of method variance associated with item directionality rather than substantively distinct theoretical constructs. When the method variance driven by positive and negative wording is explicitly isolated and controlled, the remaining variance is dominated by a single substantive dimension of general subjective loneliness.

Subsequent Multidimensional Formulations

Despite Russell’s defense of unidimensional scoring with method factors, other psychometricians (e.g., McWhirter, 1997; Hawkley et al., 2005) have established that three correlated first-order factors provide meaningful conceptual utility in specialized research designs:

• Factor 1: Isolation (e.g., feeling alone, isolated, left out, no one to turn to).
• Factor 2: Relational Connectedness (e.g., feeling close to people, having people who understand you).
• Factor 3: Collective Connectedness (e.g., feeling part of a group, having common interests with peers).

Nevertheless, because the general loneliness dimension accounts for the preponderance of common item variance (explained common variance [ECV] typically exceeding 70%), total score summation remains the standard, empirically validated scoring practice worldwide.

10. Instrument / Measurement Tool

The UCLA Loneliness Scale – Version 3 is a standardized, self-administered psychometric questionnaire. Below is the operational measurement profile of the instrument:

• Instrument Type: Self-report psychological scale / psychometric screening inventory.
• Target Population: Adolescents (ages 12+), college students, working adults, clinical outpatients, and older adults. Suitable for general community and clinical populations.
• Administration Time: Approximately 5 to 7 minutes.
• Administration Modes: Paper-and-pencil questionnaire, computerized survey (web/desktop), or face-to-face/telephone structured interview.
• Item Count: 20 items.
• Item Balance: 11 negatively worded items (indicating loneliness) and 9 positively worded items (indicating social connection).
• Response Scale: 4-point Likert scale: 1 = Never, 2 = Rarely, 3 = Sometimes, 4 = Always.
• Scoring Rules:
  • Reverse Scoring: Items 1, 5, 6, 9, 10, 15, 16, 19, and 20 are positively phrased and must be reversed prior to computing the overall scale score (i.e., 1 = 4, 2 = 3, 3 = 2, 4 = 1).
  • Direct Scoring: Items 2, 3, 4, 7, 8, 11, 12, 13, 14, 17, and 18 are scored directly as marked (1 = 1, 2 = 2, 3 = 3, 4 = 4).
  • Total Score Calculation: Sum all 20 items after reverse-scoring. Total scores range from 20 to 80. Higher scores indicate greater degrees of subjective loneliness.
• Interpretive Stratification & Normative Cutoff Benchmarks:
  • Score 20 – 34: Low / minimal loneliness (normal social connectedness).
  • Score 35 – 49: Moderate loneliness (mild to intermittent relational dissatisfaction).
  • Score 50 – 64: Moderately high loneliness (significant social isolation; common threshold for clinical concern).
  • Score 65 – 80: Severe loneliness (profound social alienation and intense relational distress).

11. Permissions & Fee and Test Year

The UCLA Loneliness Scale – Version 3 was officially published in 1996 by Daniel W. Russell in the peer-reviewed psychometrics journal Journal of Personality Assessment.

Permissions and Accessibility:
In accordance with academic fair-use conventions and Dr. Russell’s published guidance, the UCLA Loneliness Scale – Version 3 is generally accessible free of charge for non-commercial academic research, university teaching, and clinical evaluation. The full instrument and scoring key were published openly within the 1996 journal article to encourage wide scientific dissemination. Researchers using the scale must provide full academic attribution and citation to Russell (1996).

For commercial applications, proprietary healthcare software integrations, pharmaceutical clinical trials, or sponsored corporate assessments, interested entities should consult the copyright holder, Taylor & Francis (the current publisher of the Journal of Personality Assessment), and the author’s representative estate/institution (Iowa State University) to secure formal commercial licensing agreements.

12. References

Below are primary foundational references documenting the design, theoretical foundations, psychometric validation, and factor structure of the UCLA Loneliness Scale series:

• Austin, B. A. (1983). Factorial structure of the UCLA Loneliness Scale. Journal of Social Psychology, 119(2), 207–210. https://doi.org/10.1080/00224545.1983.9924463
• Bowlby, J. (1982). Attachment and loss: Vol. 1. Attachment (2nd ed.). Basic Books.
• Cacioppo, J. T., & Hawkley, L. C. (2009). Perceived social isolation and cognition. Trends in Cognitive Sciences, 13(10), 447–454. https://doi.org/10.1016/j.tics.2009.06.005
• Hawkley, L. C., Browne, M. W., & Cacioppo, J. T. (2005). How can I connect with thee? Let me count the ways. Psychological Science, 16(10), 798–804. https://doi.org/10.1111/j.1467-9280.2005.01617.x
• Hughes, M. E., Waite, L. J., Hawkley, L. C., & Cacioppo, J. T. (2004). A short scale for measuring loneliness in large surveys: Results from two population-based studies. Research on Aging, 26(6), 655–672. https://doi.org/10.1177/0164027504268574
• McWhirter, B. T. (1997). Factor analysis of the Revised UCLA Loneliness Scale with high school students. Personality and Individual Differences, 23(1), 47–55. https://doi.org/10.1016/S0191-8869(97)00021-9
• Peplau, L. A., & Perlman, D. (Eds.). (1982). Loneliness: A sourcebook of current theory, research, and therapy. John Wiley & Sons.
• Russell, D. W. (1996). UCLA Loneliness Scale (Version 3): Reliability, validity, and factor structure. Journal of Personality Assessment, 66(1), 20–40. https://doi.org/10.1207/s15327752jpa6601_2
• Russell, D., Peplau, L. A., & Cutrona, C. E. (1980). The revised UCLA Loneliness Scale: Concurrent and discriminant validity evidence. Journal of Personality and Social Psychology, 39(3), 472–480. https://doi.org/10.1037/0022-3514.39.3.472
• Russell, D., Peplau, L. A., & Ferguson, M. L. (1978). Developing a measure of loneliness. Journal of Personality Assessment, 42(3), 290–294. https://doi.org/10.1207/s15327752jpa4203_11
• Weiss, R. S. (1973). Loneliness: The experience of emotional and social isolation. The MIT Press.

13. Items of the Scale (Questionnaire)