Multidimensional Scale of Perceived Social Support (MSPSS)

Abstract

The Multidimensional Scale of Perceived Social Support (MSPSS) is a brief, psychometrically sound, self-report instrument designed to evaluate an individual’s cognitive appraisal of the adequacy and availability of their social support resources. Developed by Gregory D. Zimet and colleagues in 1988, the instrument addresses a critical distinction within psychological science: the differentiation between the structural properties of social networks (such as network size or contact frequency) and the functional, subjective perception of support quality. The MSPSS comprises 12 items distributed evenly across three distinct contextual sources: Family, Friends, and a Significant Other. Each item is scored on a 7-point Likert scale ranging from 1 (Very Strongly Disagree) to 7 (Very Strongly Agree), facilitating both global support indices and granular, source-specific evaluations. Psychometric investigations across diverse clinical, adolescent, geriatric, and cross-cultural populations have repeatedly established the scale’s robust internal consistency (Cronbach’s alpha typically ranging from .85 to .93 across subscales), stable test-retest reliability, and clear three-factor latent structure. This article provides an exhaustive examination of the MSPSS, elucidating its theoretical foundations, structural and construct validity, reliability parameters, factor-analytic evidence, scoring methodologies, and broad applicability in clinical, epidemiological, and health psychology paradigms.

Keywords

Multidimensional Scale of Perceived Social Support, MSPSS, perceived social support, psychometrics, family support, peer support, significant other, stress-buffering hypothesis, factor structure, construct validity

Authors

The Multidimensional Scale of Perceived Social Support was conceptualized, developed, and initially validated by an interdisciplinary team of clinical psychologists and medical researchers:

• Gregory D. Zimet, Ph.D. — Professor of Pediatrics and Clinical Psychology, Indiana University School of Medicine, Indianapolis, Indiana, USA. Dr. Zimet has authored numerous studies on adolescent health behaviors, psychosocial oncology, and psychometric instrument design.
• Nancy W. Dahlem, Ph.D. — Department of Psychiatry, University of Colorado Health Sciences Center, Denver, Colorado, USA.
• Sara G. Zimet, Ed.D. — Department of Psychiatry, University of Colorado Health Sciences Center, Denver, Colorado, USA.
• Gordon K. Farley, M.D. — Professor of Child Psychiatry, Day Treatment Center, University of Colorado Health Sciences Center, Denver, Colorado, USA.

Inquiries regarding the historical development of the scale may be directed to Dr. Gregory D. Zimet at the Department of Pediatrics, Section of Adolescent Medicine, Indiana University School of Medicine.

Purpose

The fundamental purpose of the Multidimensional Scale of Perceived Social Support is to quantify an individual’s subjective appraisal of social support sufficiency derived from three essential relational domains: family members, peer networks, and a central significant other. Prior to the introduction of the MSPSS in 1988, the prevailing psychometric landscape was bifurcated. Researchers typically relied on either extensive, cumbersome batteries that assessed multiple operational categories of support (e.g., informational, instrumental, appraisal, and emotional support) across complex matrices, or simplistic sociological counts of an individual’s social network ties. Both paradigms exhibited significant limitations in applied clinical and empirical contexts.

Extensive measurement batteries placed an undue cognitive and temporal burden on vulnerable patient populations, such as medically ill individuals, psychiatric inpatients, and adolescents. Conversely, structural network indices frequently demonstrated tenuous correlations with mental health outcomes, failing to capture whether an individual actually felt cared for, valued, or psychologically cushioned against life stressors. The MSPSS was purposefully engineered to bridge this methodological divide by providing a brief, economically administered 12-item instrument that preserves conceptual nuance by parsing support into its primary relational sources rather than functional micro-modalities.

In clinical practice, the MSPSS serves as an indispensable diagnostic and monitoring tool. Clinicians utilize source-specific subscale profiles to identify ecological vulnerabilities. For instance, an adolescent experiencing severe familial conflict may exhibit profoundly suppressed Family subscale scores while maintaining elevated Friend or Significant Other scores, indicating localized compensatory pathways that therapeutic interventions can leverage. In medical psychology and behavioral medicine, the scale is routinely deployed to examine how perceived support mitigates the psychological morbidity associated with chronic physical illnesses, including cardiovascular disease, cancer, diabetes, and human immunodeficiency virus (HIV). In research domains, the MSPSS functions as an established operationalization of perceived support within structural equation modeling, epidemiological surveys, and intervention trials evaluating stress resilience, treatment adherence, and post-traumatic adaptation.

Psychological Construct

The core psychological construct measured by the MSPSS is perceived social support. Psychometric and theoretical consensus defines perceived social support not as an objective accounting of received interpersonal transactions, but as a stable cognitive appraisal or schema regarding the availability and adequacy of supportive connections when necessity arises. It reflects an internalized sense of relational security, emotional validation, and logistical responsiveness.

Rather than conceptualizing social support as a monolithic, unidimensional phenomenon, the MSPSS operationalizes the construct as a tri-partite, multidimensional framework consisting of three distinct relational sources:

1. Significant Other Support

The Significant Other subscale (comprising items 1, 2, 5, and 10) captures the perception of having a designated, central figure who functions as a dedicated source of comfort, emotional intimacy, and responsive caregiving. Zimet et al. (1988) deliberately utilized the generalized descriptor “special person” to accommodate diverse developmental and social realities without forcing individuals into restrictive traditional categories. Depending on the respondent’s life stage, cultural background, and relational circumstances, this “special person” may represent a romantic partner, spouse, mentor, coach, counselor, teacher, healthcare worker, or non-familial confidant. This subscale measures deep dyadic attachment, emotional containment, and the certainty of individualized attention during episodes of distress.

2. Family Support

The Family subscale (comprising items 3, 4, 8, and 11) assesses the perceived availability, emotional warmth, and active assistance originating from familial kinship systems. Items in this dimension evaluate the degree to which family members actively attempt to help, offer emotional support, share in decision-making processes, and foster an environment where problems can be discussed openly without fear of rejection or condemnation. In developmental psychology, familial support serves as the primary developmental bedrock, influencing early attachment security, distress tolerance, and cognitive appraisals of self-worth.

3. Friend Support

The Friend subscale (comprising items 6, 7, 9, and 12) measures the perceived dependability, empathy, and emotional reciprocal sharing provided by peer relationships. Items evaluate the respondent’s confidence that friends will attempt to assist in moments of adversity, remain dependable when crises emerge, provide reciprocal sharing of affective states (“joys and sorrows”), and serve as accessible sounding boards for personal conflicts. Peer support represents an essential developmental domain, particularly during adolescence and emerging adulthood, facilitating individuation from the family of origin and fostering horizontal social belonging.

Importantly, the MSPSS assesses the cognitive evaluation of these domains rather than behavioral enactments of received support. Empirical literature consistently demonstrates that perceived support correlates more robustly with positive psychological indices, neuroendocrine regulation, and affective well-being than actual received support, which can occasionally inadvertently induce feelings of incompetence, indebtedness, or situational distress.

Theoretical Framework

The theoretical architecture of the MSPSS is situated primarily within the intersection of cognitive appraisal models of stress and psychosocial resource theory, most notably the stress-buffering hypothesis articulated by Sheldon Cohen and Thomas Ashby Wills (1985), alongside Richard Lazarus and Susan Folkman’s (1984) Transactional Model of Stress and Coping.

The Stress-Buffering Hypothesis and Direct Effects Models

Cohen and Wills (1985) posited two distinct mechanisms through which social relationships influence psychological and physical health: the direct (or main) effect model and the buffering model. The direct effect model posits that social integration provides regular positive experiences, stable social roles, and a predictable environment that confers psychological well-being irrespective of stress exposure. Conversely, the stress-buffering model asserts that social support acts as a protective shield specifically under conditions of elevated stress, attenuating the detrimental physiological and psychological consequences of stressors.

Within this framework, the MSPSS specifically captures the psychological mechanisms underpinning buffering. When confronted with an environmental demand or threat, an individual undergoes primary appraisal (evaluating the threat’s severity) and secondary appraisal (evaluating available coping resources). High perceived social support, as measured by the MSPSS, profoundly alters secondary appraisal: individuals perceive that they possess the interpersonal collateral necessary to cope with demands, thereby preventing benign stress from escalating into severe psychological distress or pathological neuroendocrine activation (e.g., prolonged hypothalamic-pituitary-adrenal [HPA] axis dysregulation).

Attachment Theory and Working Models

The MSPSS is also closely aligned with John Bowlby’s Attachment Theory. Bowlby hypothesized that early relational experiences with primary caregivers consolidate into “internal working models” of self and others. Individuals with secure working models perceive others as accessible, reliable, and benign, and the self as worthy of affection and care. The 12 items of the MSPSS operationalize these cognitive schemas across developmental trajectories, mapping directly onto the expectation that relational partners will remain proximate and emotionally responsive during times of vulnerability.

The Convoy Model of Social Relations

Toni Antonucci and Robert Kahn’s Convoy Model of Social Relations (1980) further elucidates the multi-source design of the MSPSS. The Convoy Model conceptualizes individuals as moving through the life course surrounded by concentric circles of social relationships varying in stability, closeness, and function. The innermost circle consists of stable, irreplaceable attachments (family and primary significant others), while outer circles encompass peers and broader associates. By explicitly disaggregating the measurement of support into Family, Friends, and Significant Other, the MSPSS honors the structural reality that human beings maintain distinct functional convoys that vary dynamically across life transitions.

Validity

The construct, convergent, discriminant, and predictive validity of the MSPSS have been subjected to rigorous empirical evaluation across hundreds of peer-reviewed investigations globally.

Construct and Factorial Validity

Construct validity was initially established by Zimet et al. (1988) in an undergraduate sample (N = 275) and subsequently reaffirmed in a clinical adolescent psychiatric inpatient cohort (Dahlem et al., 1991) and pediatric clinic populations (Canty-Mitchell & Zimet, 2000). Confirmatory factor analytic investigations routinely demonstrate that the hypothesized three-factor model provides a significantly superior fit to empirical data compared to alternative one-factor (unidimensional support) or two-factor (familial vs. non-familial) models. Goodness-of-fit indices across diverse cultural groups typically meet or exceed standard psychometric thresholds (e.g., Comparative Fit Index [CFI] > .95, Tucker-Lewis Index [TLI] > .95, and Root Mean Square Error of Approximation [RMSEA] < .06).

Convergent Validity

Convergent validity is evidenced by robust, theoretically congruent correlations between the MSPSS and validated measures of emotional distress, psychological well-being, and social functioning:

• Depressive Symptoms: MSPSS scores consistently correlate negatively with validated depression instruments, including the Beck Depression Inventory (BDI; correlations typically ranging from r = −.25 to −.45, p < .001) and the Center for Epidemiologic Studies Depression Scale (CES-D). Higher perceived support reliably associates with lower affective, somatic, and cognitive symptoms of depression.
• Anxiety and Somatization: Significant inverse associations are observed between the MSPSS and the Hopkins Symptom Checklist (HSCL), the State-Trait Anxiety Inventory (STAI), and the Generalized Anxiety Disorder-7 (GAD-7) scale (r values typically ranging from −.20 to −.38).
• Loneliness: The MSPSS exhibits strong inverse correlations with the UCLA Loneliness Scale (r values between −.55 and −.70), demonstrating that higher perceived support directly counters the painful subjective isolation characteristic of chronic loneliness.
• Self-Esteem and Optimism: Positive associations are repeatedly documented between MSPSS subscales and the Rosenberg Self-Esteem Scale (RSES; r = .35 to .50) as well as the Life Orientation Test-Revised (LOT-R), verifying that perceived interpersonal support fosters positive global self-evaluations.

Discriminant Validity

Discriminant validity is supported by the scale’s ability to delineate distinct functional domains. For example, in studies of college students living away from home, the Friend subscale often exhibits higher absolute levels and distinct predictive pathways toward campus social adjustment compared to the Family subscale. Conversely, in populations coping with severe chronic physical illness (e.g., end-stage renal disease, oncology), Family and Significant Other support frequently emerge as the sole significant predictors of functional adaptation, while Friend support operates independently. Furthermore, the MSPSS correlates weakly or non-significantly with measures of social desirability (e.g., the Marlowe-Crowne Social Desirability Scale, r < .15), establishing that high scores reflect genuine perceived social availability rather than impression management tendencies.

Predictive and Ecological Validity

Prospective longitudinal designs have validated the predictive utility of the MSPSS. Baseline MSPSS scores predict long-term medical treatment adherence, lower rates of psychiatric readmission, attenuated cortisol reactivity during laboratory stress inductions (e.g., the Trier Social Stress Test), and lower allostatic load markers in aging cohorts over multi-year follow-ups.

Reliability

The psychometric reliability of the MSPSS has been established across multiple analytical parameters, including internal consistency, test-retest stability, and cross-sample reliability generalization.

Internal Consistency

Internal consistency metrics for the MSPSS total scale and its three subscales consistently satisfy Nunnally’s rigorous criteria (≥ .80) for research and clinical assessment. In the seminal psychometric investigation conducted by Zimet et al. (1988), Cronbach’s alpha coefficients were determined as follows:

• Significant Other Subscale: α = .91
• Family Subscale: α = .87
• Friends Subscale: α = .85
• Total MSPSS Scale: α = .88

Subsequent psychometric replications have repeatedly substantiated these findings. Dahlem, Zimet, and Walker (1991) reported alphas of .91 (Significant Other), .87 (Family), .89 (Friends), and .88 (Total Scale). In an adolescent urban sample, Canty-Mitchell and Zimet (2000) recorded alpha coefficients of .92 for Significant Other, .88 for Family, .89 for Friends, and .93 for the Total Scale. Reliability generalization meta-analyses across international adaptations (e.g., Spanish, Chinese, Turkish, Arabic, French, and Urdu) consistently yield mean internal consistency estimates between .85 and .94 for the total instrument.

Test-Retest Reliability

Temporal stability has been verified across varying test-retest intervals. In the initial validation study (Zimet et al., 1988), a subsample of 69 university students completed the instrument across a 2- to 3-month latency period. The resulting test-retest correlation coefficients were:

• Significant Other: r = .72 (p < .001)
• Family: r = .85 (p < .001)
• Friends: r = .75 (p < .001)
• Total Scale: r = .85 (p < .001)

These findings substantiate that the MSPSS captures stable, enduring cognitive schemas rather than transient affective states or immediate contextual fluctuations, while still maintaining sufficient sensitivity to capture genuine shifts in life circumstances (such as divorce, bereavement, or relocation).

Factor Analysis

The structural dimensionality of the MSPSS has been extensively scrutinized through Exploratory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA).

Exploratory Factor Analysis (EFA)

In the foundational scale construction study by Zimet et al. (1988), the 12 items were subjected to principal components factor analysis followed by orthogonal (Varimax) and oblique rotations. The mathematical extraction revealed three distinct factors with eigenvalues substantially exceeding Kaiser’s criterion of 1.0:

• Factor 1 (Family): Eigenvalue = 5.71, accounting for 47.6% of the total variance.
• Factor 2 (Friends): Eigenvalue = 1.69, accounting for 14.1% of the total variance.
• Factor 3 (Significant Other): Eigenvalue = 1.05, accounting for 8.8% of the total variance.

Collectively, the three-factor solution explained 70.5% of the total item variance. Factor loadings after orthogonal rotation were exceptionally clean, with primary item loadings ranging from .73 to .91, and cross-loadings across alternative factors uniformly remaining beneath .35. Specifically:

• Items 1, 2, 5, and 10 loaded definitively onto the Significant Other factor (.79 to .89).
• Items 3, 4, 8, and 11 loaded definitively onto the Family factor (.73 to .87).
• Items 6, 7, 9, and 12 loaded definitively onto the Friends factor (.75 to .84).

Confirmatory Factor Analysis (CFA)

Modern psychometric literature has deployed structural equation modeling (SEM) to contrast alternative factor configurations. Competing models typically include:

1. Unidimensional Model: All 12 items load onto a single general perceived support latent factor.
2. Two-Factor Model: Familial support versus non-familial support (collapsing Friends and Significant Other into a singular peer factor).
3. Three-Factor Oblique Model: Three distinct but correlated latent factors representing Family, Friends, and Significant Other.
4. Bifactor Model / Second-Order Model: A global social support factor underlying the three primary source dimensions.

Across diverse clinical and non-clinical cohorts, CFA consistently demonstrates the statistical superiority of the three-factor correlated model and the hierarchical second-order model. Typical fit indices reported in literature include χ²/df ratios between 1.5 and 2.5, CFI > .96, TLI > .95, SRMR < .04, and RMSEA < .055 (with 90% confidence intervals bounded within .04 and .07). Measurement invariance testing across sex, age, and cultural groups has repeatedly verified full configural, metric, and scalar invariance, demonstrating that the underlying construct functions identically across demographic categories.

Instrument / Measurement Tool

The Multidimensional Scale of Perceived Social Support is structured as follows:

• Instrument Type: Standardized self-report psychometric questionnaire.
• Number of Items: 12 items.
• Administration Format: Paper-and-pencil, computer-administered (web/survey platforms), or structured clinical interview.
• Completion Time: Approximately 3 to 5 minutes.
• Target Population: Adolescents (ages 12+) and adults across clinical and community settings.
• Response Scale: 7-point Likert scale:
  • 1 = Very Strongly Disagree
  • 2 = Strongly Disagree
  • 3 = Mildly Disagree
  • 4 = Neutral
  • 5 = Mildly Agree
  • 6 = Strongly Agree
  • 7 = Very Strongly Agree
• Subscale Item Breakdown:
  • Significant Other Subscale: Items 1, 2, 5, and 10
  • Family Subscale: Items 3, 4, 8, and 11
  • Friends Subscale: Items 6, 7, 9, and 12
• Scoring Methodology:
  • There are no reverse-scored items; all items are positively phrased.
  • Subscale Mean Calculation: Sum the item responses for a subscale and divide by 4.
  • Total Scale Mean Calculation: Sum all 12 item responses and divide by 12.
  • Alternatively, sum scores can be utilized: subscale sums range from 4 to 28, and total scale sums range from 12 to 84. However, mean scores (ranging from 1 to 7) are standard in literature to retain direct interpretability relative to the Likert scale anchors.
• Score Interpretation Guidelines (Mean Scores):
  • Low Perceived Support: Mean score of 1.0 to 2.9
  • Moderate Perceived Support: Mean score of 3.0 to 5.0
  • High Perceived Support: Mean score of 5.1 to 7.0

Permissions & Fee and Test Year

The Multidimensional Scale of Perceived Social Support was originally published in 1988 in the Journal of Personality Assessment. In keeping with the scientific ethos of the scale’s creators, the MSPSS is placed in the public domain for non-commercial research, academic, and clinical healthcare evaluation purposes.

No licensing fees or formal written permissions are required from Dr. Gregory D. Zimet or his co-authors to administer, reproduce, or translate the instrument for non-commercial research or clinical practice, provided that appropriate scholarly attribution and reference to the original 1988 validation publication are maintained. Any planned commercial use, incorporation into for-profit proprietary software systems, or fee-for-service testing platforms necessitates direct communication with the primary author.

References

• Antonucci, T. C., & Kahn, R. L. (1980). A convoy model of social relations: Theoretical and methodological considerations. In Aging and Social change (pp. 53–71). Academic Press.
• Canty-Mitchell, J., & Zimet, G. D. (2000). Psychometric properties of the Multidimensional Scale of Perceived Social Support in urban adolescents. American Journal of Community Psychology, 28(3), 391–400. https://doi.org/10.1023/A:1005109522457
• Cohen, S., & Wills, T. A. (1985). Stress, social support, and the buffering hypothesis. Psychological Bulletin, 98(2), 310–357. https://doi.org/10.1037/0033-2909.98.2.310
• Dahlem, N. W., Zimet, G. D., & Walker, R. R. (1991). The Multidimensional Scale of Perceived Social Support: A study of adolescent psychiatric patients. Journal of Personality Assessment, 56(3), 438–445. https://doi.org/10.1207/s15327752jpa5603_7
• Kazarian, S. S., & McCabe, S. B. (1991). Dimensions of social support in the MSPSS: Factorial validity, reliability, and relations to depression in adolescents. Journal of Community Psychology, 19(2), 170–179. https://doi.org/10.1002/1520-6629(199104)19:2<170::AID-JCOP2290190208>3.0.CO;2-E
• Lazarus, R. S., & Folkman, S. (1984). Stress, appraisal, and coping. Springer Publishing Company.
• Tonsing, K. N., Zimet, G. D., & Tse, S. (2012). Assessing psychometric properties of the Multidimensional Scale of Perceived Social Support in South Asian ethnic minorities in Hong Kong. International Journal of Social Psychiatry, 58(4), 395–403. https://doi.org/10.1177/0020764011400262
• Zimet, G. D., Dahlem, N. W., Zimet, S. G., & Farley, G. K. (1988). The Multidimensional Scale of Perceived Social Support. Journal of Personality Assessment, 52(1), 30–41. https://doi.org/10.1207/s15327752jpa5201_2
• Zimet, G. D., Powell, S. S., Farley, G. K., Werkman, S., & Berkoff, K. A. (1990). Psychometric characteristics of the Multidimensional Scale of Perceived Social Support. Journal of Personality Assessment, 55(3-4), 610–617. https://doi.org/10.1207/s15327752jpa5503&4_17

13. Items of the Scale (Questionnaire)