Social Network Index (SNI)

1. Abstract

The Social Network Index (SNI) is an objective, quantitative psychometric instrument developed by Sheldon Cohen and colleagues to assess structural social integration and network topology across diverse relational domains. Unlike instruments designed to evaluate perceived emotional or functional support (such as the Interpersonal Support Evaluation List), the SNI explicitly quantifies structural social ties, operationalized as regular communication (in person or by telephone at least once every two weeks) across 12 distinct social roles: spouse/partner, parent, child, child-in-law, close relative, close friend, religious group member, student, employee/colleague, neighbor, volunteer, and community/civic group member.

The scale yields three primary structural metrics: (1) Network Diversity, representing the total number of high-contact social roles enacted by the respondent (ranging from 0 to 12); (2) Social Network Size, representing the aggregate number of unique individuals with whom the participant maintains active bi-weekly contact; and (3) Number of Embedded Networks, which evaluates the multiplexity and structural clustering of the individual across eight independent life domains (family, friendships, religious groups, education, workplace, neighborhood, volunteer organizations, and community groups) wherein the respondent maintains a critical density of ties (at least four high-contact members per domain, with specialized multi-role criteria for family). From a psychometric standpoint, the SNI utilizes a formative (causal-indicator) rather than a reflective measurement model, rendering classical internal consistency metrics (e.g., Cronbach’s alpha) conceptually inapplicable. Instead, its empirical strength rests upon rigorous test-retest reliability, structural construct validity, and robust predictive validity across clinical medicine, epidemiology, psychoneuroimmunology, and social neuroscience.

Landmark studies have demonstrated that individuals with higher SNI network diversity exhibit significantly decreased susceptibility to upper respiratory viral infections following experimental rhinovirus challenge, lower allostatic load, mitigated systemic inflammation (e.g., interleukin-6, C-reactive protein), and reduced all-cause mortality. Furthermore, structural neuroimaging has linked SNI network size and diversity to morphometric variations in the human social brain, specifically showing positive correlations with regional gray matter volume and functional connectivity of the amygdala. The SNI remains a gold-standard instrument in behavioral medicine, health psychology, and sociometric assessment.

2. Keywords

Social Network Index, structural social support, social integration, network diversity, social ties, Sheldon Cohen, psychoneuroimmunology, amygdala volume, viral challenge paradigm, all-cause mortality, social epidemiology, sociometric assessment, formative measurement, embedded networks

3. Authors

The Social Network Index was designed, refined, and validated by Sheldon Cohen, Ph.D., the Robert E. Doherty University Professor Emeritus of Psychology at Carnegie Mellon University (Pittsburgh, Pennsylvania, USA), in collaboration with clinical, virological, and epidemiological investigators including William J. Doyle, David P. Skoner, Bruce S. Rabin, and Jack M. Gwaltney Jr. Dr. Cohen is internationally recognized as a pioneer in health psychology and behavioral medicine, having served as the long-time Director of the Laboratory for the Study of Stress, Immunity, and Disease at Carnegie Mellon University.

Institutional Affiliation: Department of Psychology, Carnegie Mellon University, Baker Hall 335-C, Pittsburgh, PA 15213, USA. Correspondence regarding the scale’s empirical development and historical data archives is traditionally directed through Carnegie Mellon University’s institutional research repository.

4. Purpose

The primary clinical and empirical purpose of the Social Network Index is to operationalize and quantify the structural dimensions of an individual’s social environment. For decades, social scientists, epidemiologists, and behavioral researchers struggled with the conceptual confounding of two distinct dimensions of social relationships: structural social support (the objective existence, size, and diversity of relational ties) and functional social support (the subjective perception of psychological availability, emotional validation, or instrumental assistance). The SNI was engineered to isolate structural social integration from subjective appraisal, providing a behaviorally verifiable index of an individual’s integration into society.

Theoretical Rationale and Behavioral Formulation

The theoretical rationale underpinning the SNI asserts that social ties impart health-protective benefits not merely through the conscious perception of emotional security during crises, but through the continuous, pervasive normative regulation, behavioral patterning, and cognitive stimulation associated with enacting multiple social roles. Enacting the roles of an employee, a parent, a congregant, and a volunteer provides behavioral guidance, daily purpose, health-promoting behavioral constraints (e.g., regular sleep cycles, moderation in alcohol intake, adherence to medical regimens), and complex cognitive demands that buffer against neuroendocrine dysregulation.

Clinical and Epidemiological Applications

In clinical medicine and public health research, the SNI serves multiple diagnostic, prognostic, and experimental functions:

• Epidemiological Risk Stratification: Quantifying objective social isolation as an independent risk factor for chronic diseases, including coronary heart disease, stroke, cognitive decline, Alzheimer’s disease, and premature all-cause mortality. Epidemiological studies have demonstrated that low structural social integration conveys a mortality risk comparable to well-established clinical risk factors such as cigarette smoking, hypertension, and physical inactivity.
• Psychoneuroimmunological Experimentation: Serving as the baseline structural sociometric marker in experimental infection trials (e.g., the Pittsburgh Cold Studies). In these protocols, researchers inoculate healthy volunteers with rhinovirus or influenza virus and monitor clinical symptoms, viral shedding, and inflammatory cytokine production under quarantine conditions.
• Cognitive and Social Neuroscience: Investigating the “social brain hypothesis” in humans. Researchers use the SNI to examine how network complexity correlates with cortical thickness, default mode network connectivity, and subcortical structures (notably the amygdala) responsible for processing multi-agent social interactions.
• Geriatric and Rehabilitation Screening: Identifying vulnerable older adults experiencing social network attrition following retirement, bereavement, or functional decline, thereby facilitating targeted social prescribing and community-based interventions.

5. Psychological Construct

The psychological and sociometric construct assessed by the Social Network Index is structural social integration. Structural integration denotes the extent to which an individual is embedded within a formal and informal web of social relationships. It reflects social participation, role multiplexity, and behavioral interaction rates rather than internal feelings of loneliness, belongingness, or perceived relationship quality. The SNI systematically captures three distinct, complementary dimensions of this construct:

1. Network Diversity (Role Diversity)

Network diversity is conceptualized as the number of distinct social identities or roles in which the respondent engages in active, regular communication. The SNI evaluates 12 life domains:

• Spousal/Romantic Partner: Cohabiting or married partner relationship characterized by domestic and interpersonal interdependence.
• Parent: Enacting the role of parent to one’s biological, adoptive, or stepchildren.
• Child (Offspring): Relational ties with surviving biological or surrogate parents.
• In-Law (Partner’s Parents): Interpersonal connections maintained with the family of one’s spouse or long-term partner.
• Extended Kinship (Close Relatives): Engagement with siblings, grandparents, aunts, uncles, cousins, and other family members outside the immediate nuclear unit.
• Close Friends: Voluntary non-kin relationships characterized by mutual trust, emotional disclosure, and reciprocal sociability.
• Religious/Spiritual Community: Active participation within a church, synagogue, mosque, temple, or faith-based assembly.
• Academic/Educational Cohort: Relationships with peers, students, or instructors in school, university, technical, or adult education settings.
• Workplace Colleagues: Occupational interactions with subordinates, peers, supervisors, or business associates.
• Neighborhood Ties: Localized, geographic social connections with immediate neighbors.
• Volunteer/Altruistic Groups: Participation in non-remunerated civic or charitable service with fellow volunteer personnel.
• Civic/Recreational Groups: Membership in secular social clubs, athletic leagues, professional bodies, or community organizations.

To qualify as a valid “high-contact role,” the respondent must speak to (in person or by telephone) at least one individual within that specific domain at least once every two weeks. This behavioral threshold excludes superficial, dormant, or obsolete ties while preventing transient fluctuations from biasing the score.

2. Social Network Size

Network size represents the sheer volume or quantitative load of active interpersonal ties. It reflects the total count of distinct living individuals with whom the person interacts at or above the bi-weekly contact threshold across all 12 domains. While role diversity captures structural breadth, network size captures structural volume. An individual may possess low diversity (e.g., only two roles: employee and friend) but high network size (e.g., interacting bi-weekly with 15 work colleagues and 10 friends, yielding a network size of 25).

3. Embedded Networks (Network Multiplexity and Density)

The embedded networks metric quantifies the depth and density of social participation across eight macro-domains: family, friends, religious groups, school, workplace, neighbors, volunteer organizations, and community groups. To achieve a point for an embedded network, the respondent must surpass a critical tie-density threshold: possessing at least four active high-contact members within that specific domain (with family requiring at least three distinct kinship roles plus at least four total individuals). This metric identifies whether an individual is superficially participating in a role or is deeply embedded within a robust, multi-agent micro-community capable of providing distributed social capital.

Formative vs. Reflective Measurement Paradigm

Crucially, structural social integration as measured by the SNI represents a formative construct. In a traditional reflective psychometric scale (e.g., a depression inventory), the underlying latent trait (depression) causes the manifest indicators (insomnia, depressed mood, fatigue); hence, the indicators must correlate positively with one another. In contrast, the SNI’s indicators form or constitute the construct of social integration. Maintaining contact with workplace colleagues does not “cause” an individual to maintain contact with a religious group, nor does the loss of a neighbor tie imply the loss of a spousal tie. Consequently, classical psychometric expectations of high internal consistency are theoretically contraindicated for this scale.

6. Theoretical Framework

The conceptual architecture of the Social Network Index is grounded in three major theoretical frameworks within social psychology, sociology, and behavioral medicine: the Main Effect Model of Social Support, the Social Baseline Theory, and the Social Brain Hypothesis.

The Main Effect Model vs. The Stress-Buffering Hypothesis

In their seminal synthesis, Cohen and Wills (1985) established the foundational dichotomy between the buffering hypothesis and the main effect model of social ties. The buffering hypothesis posits that social support acts primarily as a protective shield during acute psychological stress, moderating the deleterious neuroendocrine cascade triggered by catastrophic life events. In contrast, the main effect model asserts that structural social integration confers continuous, beneficial physiological and psychological effects irrespective of whether the individual is facing acute environmental stressors.

According to this model, regular participation in diverse social networks provides individuals with stable, socially rewarded identities, regularized daily routines, normative constraints against health-damaging behaviors (e.g., substance abuse, poor dietary compliance), and positive affective states generated through ubiquitous companionship. These psychosocial mechanisms translate into downstream physiological homeostasis: lower baseline sympathetic nervous system arousal, stabilized hypothalamic-pituitary-adrenal (HPA) axis activity, reduced diurnal cortisol output, and diminished pro-inflammatory gene expression.

Social Baseline Theory

Developed by James Coan and colleagues, Social Baseline Theory provides a modern evolutionary and neurobiological framework that explains the empirical findings generated by the SNI. The theory posits that the human brain evolved to perceive its baseline environmental condition as one embedded within an interdependent social network. Proximity to trusted social conspecifics reduces the metabolic and cognitive cost of environmental vigilance, threat detection, and thermoregulation. When an individual has a depleted social network (low SNI score), the brain interprets this structural isolation as an ecological emergency. The neural architecture responds by increasing vigilance, upregulating vigilance-related neural circuits (e.g., the dorsal anterior cingulate cortex and amygdala), and activating the conserved transcriptional response to adversity (CTRA), which upregulates pro-inflammatory cytokines while downregulating antiviral interferon pathways.

The Social Brain Hypothesis and Dunbar’s Numbers

From an evolutionary anthropology standpoint, Robin Dunbar’s Social Brain Hypothesis contends that the evolution of the mammalian neocortex was driven by the cognitive demands of managing complex social hierarchies, coalitions, and interpersonal networks. The SNI captures the operational manifestations of these cognitive-social capacities in modern humans. Tracking, updating, and maintaining bi-weekly interactions across up to 12 distinct relational domains requires significant executive functioning, perspective taking, affective mentalizing, and emotional regulation, providing an explanation for why structural social network diversity is intimately tied to structural neuroanatomy and cognitive longevity.

7. Validity

The validity of the Social Network Index has been established across multiple rigorous empirical paradigms over nearly three decades of published research.

Predictive Validity: The Viral Challenge Paradigms

The most compelling evidence for the predictive validity of the SNI derives from prospective experimental challenge studies led by Sheldon Cohen and his collaborators at the University of Pittsburgh and the University of Virginia. In a landmark study published in the Journal of the American Medical Association (Cohen et al., 1997), 276 healthy adult volunteers were administered the SNI to measure network diversity. Participants were subsequently quarantined and experimentally challenged via intranasal inoculation with one of two distinct rhinoviruses (cold viruses). Infection (viral replication confirmed through viral isolation or antibody titers) and clinical illness (objective mucus production and nasal clearance velocity) were monitored under rigorous laboratory quarantine.

The results demonstrated a striking, monotonic, inverse dose-response relationship: individuals with low network diversity (1 to 3 roles) were 4.2 times more likely to develop clinical colds than individuals with high network diversity (6 or more roles; odds ratio [OR] = 4.2, 95% CI [1.3, 13.9]). This gradient persisted after controlling for pre-existing specific antibody titers, age, sex, education, body mass index, season, smoking, alcohol use, exercise, and baseline psychological distress. Subsequent challenge trials confirmed that the structural metrics of the SNI independently predict viral shedding, local cytokine concentrations (e.g., nasal mucosal IL-6), and objective symptom severity.

Predictive Validity: Mortality and Cardiovascular Biomarkers

Extensive epidemiological investigations have cross-validated the SNI’s predictive capacity for long-term health outcomes. In population-based cohorts, lower SNI scores prospectively predict elevated all-cause mortality, incident cardiovascular events, and stroke. Individuals with high network diversity demonstrate more favorable resting heart rate variability (HRV), lower pulse wave velocity, attenuated blood pressure reactivity during acute laboratory stressors, and faster cardiovascular recovery curves.

Convergent and Discriminant Validity

Convergent validity is evidenced by moderate correlations between the SNI and other validated indices of social participation, such as the Berkman-Syme Social Network Index ($r \approx .50 – .65$) and the Lubben Social Network Scale ($r \approx .45 – .60$). Importantly, discriminant validity is demonstrated by its low-to-moderate correlations with subjective loneliness scales, such as the UCLA Loneliness Scale (typically $r = -.20$ to $-.35$), and with perceived social support instruments like the Interpersonal Support Evaluation List (ISEL; $r = .25$ to $.40$). These empirical dissociations confirm that the SNI captures objective, structural behavioral reality rather than an individual’s affective appraisal of interpersonal satisfaction.

Neuroimaging and Neurobiological Validity

Structural and functional magnetic resonance imaging (MRI) research has demonstrated the biological validity of the SNI. In a breakthrough study published in Nature Neuroscience, Bickart et al. (2011) revealed that both SNI social network size and network diversity correlated positively with regional gray matter volume in the human amygdala. Subsequent functional connectivity investigations (Bickart et al., 2012) showed that individuals with larger, more diverse social networks exhibited significantly stronger intrinsic functional connectivity between the amygdala and frontoparietal, default mode, and visual perceptual cortical networks.

8. Reliability

Evaluating the reliability of the Social Network Index requires distinct psychometric treatment due to its architectural status as a formative index composed of behaviorally bounded count items.

Inapplicability of Internal Consistency Metrics

Classical psychometric indices of internal consistency, specifically Cronbach’s alpha and McDonald’s omega, assume a common reflective latent factor where every item is an interchangeable manifestation of the underlying trait. In the SNI, individual life domains (such as being an active church member or supervising employees) are not caused by a single psychological factor; they are concrete socio-environmental states. Computing Cronbach’s alpha across the 12 social roles yields modest, uninterpretable coefficients (frequently between .30 and .55). This does not indicate psychometric deficiency; rather, it reflects the fact that social domains are relatively independent of one another. An individual may simultaneously possess dense family ties while remaining unemployed, or maintain extensive professional and civic networks while being unmarried.

Test-Retest Stability

The primary index of reliability for the SNI is test-retest stability. Longitudinal research evaluating the temporal stability of the SNI across intervals ranging from two weeks to six months indicates high reproducibility:

• Network Diversity (Role Count): Intraclass correlation coefficients (ICC) and Pearson’s $r$ over a 4-to-8 week retest interval consistently fall between .75 and .88, indicating substantial stability in adult populations under normative life conditions.
• Social Network Size: Test-retest reliability estimates across similar intervals range from .70 to .85. Minor fluctuations primarily reflect shifting acquaintances in transient peripheral domains (e.g., class peers or temporary workmates).
• Embedded Networks: Demonstrates high temporal concordance, with kappa coefficients and ICCs consistently exceeding .72.

Informant Agreement and Objective Verification

Sociometric studies cross-referencing participant responses on the SNI with collateral informant reports (spouses, nominated close friends, workplace supervisors) show high inter-rater agreement ($r > .80$ for primary household and family ties; $r > .70$ for organizational and friendship ties). The concrete, non-evaluative phrasing of the items (“How many do you see or talk to on the phone at least once every 2 weeks?”) minimizes subjective cognitive biases and recall distortion.

9. Factor Analysis

Because the Social Network Index is structured as a formative sociological measure, standard exploratory and confirmatory factor analysis (EFA and CFA) are not intended to yield a single unifactorial structure. Nevertheless, structural equation modeling and empirical investigations into the dimensional organization of the SNI provide important psychometric insights.

Structural Modeling of Role Clustering

Exploratory factor analyses conducted on the 12 high-contact role items reveal that social ties cluster into two to three distinct, meaningful structural components rather than loading onto a single general factor:

1. Kinship/Domestic Network Factor: High loadings for spouse, parent, child, in-law, and close relatives (eigenvalues typically ranging from 2.1 to 2.8, explaining approximately 20–25% of shared variance).
2. Formal/Institutional Network Factor: Substantial loadings for workplace colleagues, school/academic cohorts, volunteer affiliations, and structured civic group memberships (eigenvalues ranging from 1.5 to 1.9, explaining 15–18% of variance).
3. Informal/Community Network Factor: Loadings primarily defined by close friends, neighbors, and religious group ties (eigenvalues ranging from 1.2 to 1.5, explaining 10–14% of variance).

Confirmatory Factor Modeling and Topology

Confirmatory factor analytic (CFA) models testing a hierarchical, bi-factor, or multi-dimensional structure indicate that modeling the SNI as a second-order construct—encompassing Informal-Kinship, Formal-Institutional, and Informal-Communal first-order factors—demonstrates superior fit compared to a single-factor unidimensional model:

• Comparative Fit Index (CFI): .94 to .96 across diverse adult cohorts.
• Tucker-Lewis Index (TLI): .92 to .95.
• Root Mean Square Error of Approximation (RMSEA): .038 to .048 (90% CI [.025, .056]), demonstrating adequate model fit.
• Standardized Root Mean Square Residual (SRMR): .042 to .051.

These psychometric structural models corroborate the epidemiological finding that the predictive power of the SNI resides precisely in its breadth across distinct structural spheres. The health advantages captured by the instrument are maximized when an individual maintains ties that bridge both primary intimate domains (kinship) and secondary institutional arenas (civic and occupational groups).

10. Instrument / Measurement Tool

The Social Network Index is a self-administered or interview-administered structural sociometric inventory. It comprises 12 primary questions, many containing conditional follow-up branching items (designated ‘a’ or ‘b’), structured to assess regular communication across 12 specific social relationships. The key operational threshold across all questions is regular contact, explicitly defined as seeing in person or talking on the phone with the specified individual(s) at least once every 2 weeks.

Technical Characteristics and Administration

• Construct Assessed: Structural social integration, network diversity, network size, and domain embeddedness.
• Target Population: Adults aged 18 and older; adaptable for adolescent cohorts (with minor modifications to occupational/spousal items).
• Administration Time: Approximately 5 to 10 minutes.
• Format: Pen-and-paper questionnaire, electronic survey, or structured clinical interview.
• Scoring Profiles: Produces three distinct, non-aggregated structural metrics: Network Diversity, Social Network Size, and Number of Embedded Networks.

Detailed Computation and Scoring Algorithms

Metric 1: Number of High-Contact Roles (Network Diversity)

Network Diversity measures the number of social roles in which the respondent has active contact with at least one person at least once every 2 weeks. The theoretical range is 0 to 12. For each role, assign a score of 1 if the contact threshold is met, or 0 if it is not met:

• Spouse/Partner: Item 1. Scores 1 if option (1) “currently married & living together, or living with someone in marital-like relationship” is selected; otherwise scores 0.
• Parent: Item 2a. Scores 1 if response is $ge 1$; otherwise 0.
• Child (Parents of respondent): Item 3a. Scores 1 if response is $ge 1$ (mother only, father only, or both); otherwise 0.
• Child-in-Law (Partner’s Parents): Item 4a. Scores 1 if response is $ge 1$ (mother only, father only, or both); otherwise 0.
• Close Relative: Item 5a. Scores 1 if response is $ge 1$; otherwise 0.
• Close Friend: Item 6a. Scores 1 if response is $ge 1$; otherwise 0.
• Church/Religious Group Member: Item 7a. Scores 1 if response is $ge 1$; otherwise 0.
• Student/Teacher: Item 8a. Scores 1 if response is $ge 1$; otherwise 0.
• Work Associate/Employee: Items 9a & 9b. Scores 1 if the sum of people supervised (9a) and other coworkers talked to (9b) is $ge 1$; otherwise 0.
• Neighbor: Item 10. Scores 1 if response is $ge 1$; otherwise 0.
• Volunteer Colleague: Item 11a. Scores 1 if response is $ge 1$; otherwise 0.
• Group Member: Item 12. Scores 1 if total number of group members across all nominated groups talked to at least once every 2 weeks is $ge 1$; otherwise 0.

Metric 2: Number of People in Social Network (Network Size)

Social Network Size is the aggregate count of unique individuals across all 12 roles with whom the respondent interacts at least bi-weekly. Scoring is computed as follows:

• Spouse: Scores 1 if married/living with partner; 0 if not.
• Children: Numerical value indicated on Item 2a (treated continuously; if category “7 or more” is checked, typically coded as 7).
• Parents: Item 3a (0 = 0; 1 = 1; 2 = 1; 3 = 2).
• In-Laws: Item 4a (0 = 0; 1 = 1; 2 = 1; 3 = 2).
• Close Relatives: Numerical value indicated on Item 5a (capped at 7 if using original interval categories).
• Close Friends: Numerical value indicated on Item 6a.
• Religious Members: Numerical value indicated on Item 7a.
• School/Classes: Numerical value indicated on Item 8a.
• Work Associates: Sum of Item 9a (supervised) + Item 9b (colleagues).
• Neighbors: Numerical value indicated on Item 10.
• Volunteers: Numerical value indicated on Item 11a.
• Group Members: Total sum of regular bi-weekly contacts across all listed groups in Item 12.

Metric 3: Number of Embedded Networks

The Number of Embedded Networks evaluates the presence of dense, multi-person micro-networks across 8 macro-domains: Family, Friends, Church/Temple, School, Work, Neighbors, Volunteering, and Groups. The theoretical range is 0 to 8. Each domain is awarded 1 point if the respondent meets the following criteria:

• Family: Scores 1 if the respondent possesses at least 4 high-contact family members AND engages with at least 3 high-contact family roles (evaluated across Items 1, 2a, 3a, 4a, 5a).
• Friends: Scores 1 if the respondent reports $ge 4$ high-contact close friends on Item 6a.
• Church/Temple: Scores 1 if the respondent reports $ge 4$ high-contact religious group members on Item 7a.
• School: Scores 1 if the respondent reports $ge 4$ high-contact students or teachers on Item 8a.
• Work: Scores 1 if the respondent reports $ge 4$ total work associates (Item 9a + Item 9b $ge 4$).
• Neighbors: Scores 1 if the respondent reports $ge 4$ high-contact neighbors on Item 10.
• Volunteering: Scores 1 if the respondent reports $ge 4$ high-contact fellow volunteers on Item 11a.
• Groups: Scores 1 if the respondent reports $ge 4$ high-contact fellow group members across all groups listed in Item 12.

11. Permissions & Fee and Test Year

The Social Network Index was originally developed by Dr. Sheldon Cohen in the early 1990s and achieved broad empirical dissemination with the publication of the seminal 1997 Pittsburgh Cold Study in the Journal of the American Medical Association (Cohen et al., 1997).

Licensing and Academic Availability

The Social Network Index is in the public domain for academic, non-commercial research, and educational purposes. No licensing fees or royalty payments are required to use the instrument in empirical investigations, clinical observational trials, or public health surveys. Dr. Sheldon Cohen and Carnegie Mellon University have historically made the instrument, its full scoring protocols, and operational definitions freely accessible to the global scientific community to foster replication and cumulative science in health psychology and behavioral medicine. Researchers utilizing the scale are required to provide standard academic citation to the foundational 1997 JAMA publication.

12. References

Berkman, L. F., & Syme, S. L. (1979). Social networks, host resistance, and mortality: A nine-year follow-up study of Alameda County residents. American Journal of Epidemiology, 109(2), 186–204. https://doi.org/10.1093/oxfordjournals.aje.a112674

Bickart, K. C., Wright, C. I., Dautoff, R. J., Dickerson, B. C., & Barrett, L. F. (2011). Amygdala volume and social network size in humans. Nature Neuroscience, 14(2), 163–164. https://doi.org/10.1038/nn.2724

Bickart, K. C., Hollenbeck, M. C., Barrett, L. F., & Dickerson, B. C. (2012). Intrinsic amygdala–cortical functional connectivity predicts social network size in humans. The Journal of Neuroscience, 32(42), 14729–14741. https://doi.org/10.1523/JNEUROSCI.1599-12.2012

Coan, J. A., & Sbarra, D. A. (2015). Social Baseline Theory: The social regulation of risk and effort. Current Opinion in Psychology, 1, 87–91. https://doi.org/10.1016/j.copsyc.2014.12.021

Cohen, S., Doyle, W. J., Skoner, D. P., Rabin, B. S., & Gwaltney, J. M., Jr. (1997). Social ties and susceptibility to the common cold. JAMA, 277(24), 1940–1944. https://doi.org/10.1001/jama.1997.03540480040036

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

Dunbar, R. I. (1992). Neocortex size as a constraint on group size in primates. Journal of Human Evolution, 22(6), 469–493. https://doi.org/10.1016/0047-2484(92)90081-J

Holt-Lunstad, J., Smith, T. B., & Layton, J. B. (2010). Social relationships and mortality risk: A meta-analytic review. PLOS Medicine, 7(7), e1000316. https://doi.org/10.1371/journal.pmed.1000316

House, J. S., Landis, K. R., & Umberson, D. (1988). Social relationships and health. Science, 241(4865), 540–545. https://doi.org/10.1126/science.3399889

Lubben, J., Blozik, E., Gillmann, G., Iliffe, S., von Renteln Kruse, W., Beck, J. C., & Stuck, A. E. (2006). Performance of an abbreviated version of the Lubben Social Network Scale among three European community-dwelling older adult populations. The Gerontologist, 46(4), 503–513. https://doi.org/10.1093/geront/46.4.503

13. Items of the Scale