Self–Other Four Immeasurable Scale (SOFI)

1. Abstract

The Self–Other Four Immeasurable Scale (SOFI) is a 16-item self-report psychometric instrument developed by Salvatore Kraus and Sharon Sears (2009) to quantify affective, cognitive, and behavioral orientations toward oneself and other individuals. Grounded in Buddhist psychological traditions—specifically the doctrine of the four divine abodes or immeasurables (Brahmavihāras): loving-kindness (mettā), compassion (karuṇā), sympathetic joy (muditā), and equanimity (upekkhā)—the scale operationalizes these virtues alongside their affective antitheses across two distinct interpersonal targets: the Self and Others.

Structured around a balanced orthogonal design, the SOFI yields four distinct subscales consisting of four items each: Positive Self, Negative Self, Positive Others, and Negative Others. Respondents rate each descriptor regarding how they have thought, felt, or acted over the preceding week using an authentic 5-point Likert response format ranging from 1 (Very slightly or not at all) to 5 (Extremely). Initial and subsequent validation studies demonstrate robust internal consistency reliability across all four dimensions, with Cronbach’s alpha coefficients routinely exceeding acceptable psychometric benchmarks (α = .80 to .86). Confirmatory factor analyses corroborate a four-factor correlated model reflecting the bidirectional interplay of valence (positive vs. negative) and target (self vs. other), demonstrating superior model fit over alternative unidimensional, two-factor, or hierarchical representations. The SOFI exhibits extensive construct, convergent, and discriminant validity, correlating systematically with measures of self-compassion, positive and negative affect, psychological well-being, mindfulness, depression, and social desirability. This comprehensive psychometric profile establishes the SOFI as an indispensable clinical, contemplative, and empirical tool for evaluating mindfulness-based stress reduction, loving-kindness meditation interventions, and contemporary compassion-focused psychotherapies.

2. Keywords

Self–Other Four Immeasurable Scale, SOFI, Brahmaviharas, Loving-Kindness, Compassion, Sympathetic Joy, Equanimity, Psychometrics, Self-Compassion, Contemplative Science, Mindfulness Interventions, Factor Analysis

3. Authors

The Self–Other Four Immeasurable Scale was designed, validated, and published by Salvatore Kraus, Ph.D., and Sharon Sears, Ph.D.

• Salvatore Kraus, Ph.D.: Department of Psychology, Fort Lewis College, Durango, Colorado, USA. Dr. Kraus has focused extensively on psychometric evaluation, positive psychology, transpersonal psychology, and the empirical measurement of Buddhist and contemplative psychological constructs. Email: [email protected].
• Sharon Sears, Ph.D.: Department of Psychology, Fort Lewis College, Durango, Colorado, USA. Dr. Sears is a licensed clinical psychologist, scholar, and contemplative researcher whose scientific work investigates mindfulness-based interventions, behavioral medicine, compassion cultivation, and psychological resilience across clinical and community populations.

4. Purpose

The primary purpose of the Self–Other Four Immeasurable Scale (SOFI) is to provide an empirically rigorous, theoretically coherent, and operationally concise measurement tool capable of evaluating both positive and negative qualities directed toward the self and toward others. Historically, contemporary psychological research witnessed an explosion of interest in mindfulness, meditation, and Buddhist-derived constructs; however, early psychometric inventories disproportionately focused on metacognitive attentional awareness (e.g., the Mindful Attention Awareness Scale) or exclusively examined self-directed emotional stances (e.g., Kristin Neff’s Self-Compassion Scale). Existing measures frequently omitted other-directed compassion, benevolent joy, and equanimity, or failed to simultaneously track the reduction of destructive affective states such as inner cruelty, self-directed hatred, interpersonal malice, and hostility.

Kraus and Sears (2009) identified a significant methodological gap in contemplative science and clinical psychology: the absence of a unified, brief, and psychometrically sound instrument that measures the full structural matrix of the four traditional Buddhist virtues across both self-referential and other-referential domains. The SOFI was formulated to fulfill three critical functions:

1. Contemplative Science and Meditation Research: Interventions such as Loving-Kindness Meditation (LKM), Compassion Cultivation Training (CCT), and Cognitively-Based Compassion Training (CBCT) aim to systematically expand benevolence from the self outward to loved ones, neutral individuals, challenging acquaintances, and all sentient beings. The SOFI provides researchers with a dual-target metric that evaluates whether contemplative training simultaneously increases positive qualities toward self and others while ameliorating negative inclinations across both relational targets.
2. Clinical and Psychotherapeutic Assessment: In clinical frameworks such as Compassion-Focused Therapy (CFT), Acceptance and Commitment Therapy (ACT), and Dialectical Behavior Therapy (DBT), patients often present with profound disparities between self-directed and other-directed affect. For example, individuals with severe trauma, major depressive disorder, or borderline personality patterns often demonstrate intact empathy or kindness toward external peers while harboring debilitating self-hatred, harsh inner criticism, and self-directed cruelty. The SOFI delineates these clinical dissociations, enabling practitioners to track specific therapeutic shifts across self-directed versus other-directed axes over the course of treatment.
3. Theoretical Refinement of Prosociality: The scale enables academic researchers to investigate whether positive emotional traits (such as joy, friendliness, acceptance, and compassion) function as direct polar opposites of negative traits (such as hatred, anger, cruelty, and meanness), or whether they represent orthogonal, semi-independent psychological systems. By isolating Positive Self, Negative Self, Positive Others, and Negative Others into distinct dimensions, the SOFI facilitates granular structural equation modeling of social, emotional, and psychological health.

5. Psychological Construct

The SOFI operationalizes human social and emotional functioning at the intersection of two structural axes: Target of Affect (Self vs. Other) and Affective Valence (Positive vs. Negative). These axes generate four discrete, interrelated psychometric dimensions, each comprising four specific emotional and behavioral indicators:

5.1. Positive Self (PS)

The Positive Self dimension measures the degree to which an individual directs benevolence, warmth, forbearance, and benevolent joy inwardly toward their own personhood. It represents the psychological operationalization of the four Buddhist immeasurables turned inward: self-loving-kindness, self-compassion, self-joy, and self-equanimity. The items comprising this subscale are:

• Friendly—toward myself: Reflects an attitude of gentle personal goodwill, warmth, and inner rapport, counteracting chronic alienation and internal hostility.
• Joyful—for myself: Reflects sympathetic joy (muditā) directed inward, characterizing the capacity to celebrate one’s own accomplishments, virtues, and moments of flourishing without neurotic guilt or imposter feelings.
• Accepting—toward myself: Represents equanimity (upekkhā) toward one’s internal experiences, flaws, physical limitations, and transient shortcomings, embodying non-judgmental forbearance.
• Compassionate—toward myself: Reflects the active, empathetic response to one’s own pain, vulnerability, and sorrow (karuṇā), manifesting as the desire to alleviate one’s suffering rather than exacerbating it through punitive self-chastisement.

5.2. Negative Self (NS)

The Negative Self dimension quantifies self-directed hostility, internal punitive toxicity, self-contempt, and harsh auto-aggression. Rather than merely assessing the absence of self-love, this construct directly captures active, pernicious internal warfare. The four constituent items are:

• Hateful—toward myself: Captures visceral self-loathing, intense internal aversion, and the toxic belief that one is fundamentally defective, repugnant, or unworthy of existence.
• Angry—with myself: Reflects acute or chronic internal friction, irritation, self-recrimination, and continuous exasperation directed at personal lapses or perceived failures.
• Cruel—toward myself: Measures the psychological propensity to actively torment, punish, or withhold relief from oneself during times of vulnerability, mirroring punitive inner-critic processes.
• Mean—toward myself: Represents petty, pervasive spite, uncharitable self-talk, and degrading internal dialogue that erodes personal dignity and self-efficacy.

5.3. Positive Others (PO)

The Positive Others dimension quantifies genuine prosocial warmth, empathetic attunement, vicarious celebration, and non-judgmental acceptance directed outward toward human beings in one’s social ecology. This reflects the classical Brahmavihāras applied interpersonally:

• Friendly—toward others: Signifies general benevolence, approachable warmth, and unconditioned goodwill (mettā) extended toward external individuals.
• Joyful—for others: Captures the purest manifestation of sympathetic joy (muditā), reflecting happiness for the success, fortune, and flourishing of other people, free from competitive envy or begrudging resentment.
• Accepting—toward others: Measures interpersonal equanimity (upekkhā), encompassing tolerance, psychological open-mindedness, and the capacity to accept others as they are without demanding conformity to one’s expectations.
• Compassionate—toward others: Represents other-oriented empathy and compassion (karuṇā), characterized by feeling moved by the suffering of others coupled with a motivational urge to support their well-being.

5.4. Negative Others (NO)

The Negative Others dimension indexes interpersonal hostility, malice, antagonism, and cruelty directed externally. It reflects relational alienation and antisocial affective stances:

• Hateful—toward others: Captures deep-seated contempt, dehumanizing antipathy, and visceral hostility toward people or out-groups.
• Angry—with others: Measures chronic interpersonal irritability, quickness to take offense, ongoing resentment, and low frustration tolerance toward others’ behavior.
• Cruel—toward others: Reflects malicious behavior, sadistic gratification, or cold indifference to inflicting psychological or physical distress upon others.
• Mean—toward others: Captures daily uncharitableness, petty belittlement, spiteful communication, and passive-aggressive or overt interpersonal nastiness.

6. Theoretical Framework

The conceptual foundation of the SOFI integrates classical Buddhist philosophy with contemporary affective neuroscience, cognitive psychology, and evolutionary models of social engagement.

6.1. The Classical Buddhist Framework of the Brahmavihāras

In Buddhist psychology, as detailed in the Pāli Canon and codified in Buddhaghosa’s 5th-century treatise the Visuddhimagga (The Path of Purification), the four immeasurables (apramāṇa) or divine dwellings (Brahmavihāra) represent mental states that cultivate boundless social harmony and liberate the mind from unwholesome afflictive emotions (kleshas):

• Mettā (Loving-kindness): The aspiration that all beings experience happiness and the causes of happiness. Its far enemy is ill will or hatred; its near enemy is selfish affection or sentimental attachment.
• Karuṇā (Compassion): The heartfelt aspiration that beings be free from suffering and the causes of suffering. Its far enemy is cruelty or malice; its near enemy is pity or overwhelmed, paralyzing grief.
• Muditā (Sympathetic Joy): The sincere delight in the happiness, virtues, and good fortune of others. Its far enemy is envy, jealousy, or cynicism; its near enemy is hypocritical or superficial exhilaration.
• Upekkhā (Equanimity): An even-minded mental state characterized by non-attachment, non-discrimination, and calm balance that sees the equality of all beings. Its far enemy is agitation, aversion, or resentment; its near enemy is indifferent apathy, cold disconnection, or nihilistic resignation.

Crucially, classical Buddhist ontology insists that genuine contemplation requires boundless expansion across all directions: inward toward oneself, outward toward beloved allies, neutral figures, hostile enemies, and universal life. Kraus and Sears (2009) preserved this core pedagogical tenet by ensuring that the positive virtues and their classical “far enemies” (hatred, anger, cruelty, meanness) were assessed across both self-directed and other-directed domains.

6.2. Contemporary Psychological and Evolutionary Parallels

The SOFI’s structural dichotomy aligns closely with contemporary social mentality theory, developed by Paul Gilbert (2005) in the architecture of Compassion-Focused Therapy. Gilbert posits that human beings operate through distinct neurobiological motivational systems: the Threat/Defense System (driven by amygdaloid activation, cortisol, and adrenaline, manifesting as anger, contempt, cruelty, and defense), the Drive/Incentive System (dopaminergic, oriented toward resource acquisition), and the Soothing/Affiliation System (mediated by oxytocin, endogenous opioids, and parasympathetic vagal tone, manifesting as kindness, warmth, and safe connection).

The SOFI’s positive subscales (PS and PO) index the activation of the mammalian Soothing/Affiliation system, reflecting safety, social connectedness, and mutual emotional regulation. Conversely, the negative subscales (NS and NO) capture hyperactive Threat/Defense systems. When the threat system is directed externally, it manifests as interpersonal hostility (Negative Others); when inverted toward the organism itself, it manifests as internal persecutory shame, self-loathing, and self-punishment (Negative Self). By parsing these systems, the SOFI operationalizes affective neuroscience principles within an accessible psychometric paradigm.

7. Validity

Empirical evaluations by Kraus and Sears (2009), along with subsequent independent investigations, provide extensive validation evidence for the SOFI across construct, convergent, discriminant, and predictive domains.

7.1. Convergent Validity

Convergent validity has been evaluated through correlations with established scales measuring compassion, affect, emotional regulation, and well-being:

• Self-Compassion Scale (SCS; Neff, 2003): The Positive Self (PS) subscale demonstrates strong, statistically significant positive correlations with the SCS total score ($r = .65$ to $.72, p < .001$) and specifically with the Self-Kindness and Common Humanity subscales. Conversely, the Negative Self (NS) subscale correlates robustly with the SCS Self-Judgment ($r = .68, p < .001$) and Isolation subscales, confirming that the SOFI accurately tracks self-compassionate versus self-critical processing.
• Positive and Negative Affect Schedule (PANAS; Watson et al., 1988): Positive Self and Positive Others demonstrate moderate to strong positive associations with the PANAS Positive Affect scale ($r = .48$ to $.56, p < .001$). Negative Self and Negative Others correlate consistently with the PANAS Negative Affect scale ($r = .52$ to $.61, p < .001$), confirming emotional valence alignment.
• Mindfulness Measures: Scores on the Five Facet Mindfulness Questionnaire (FFMQ) and Mindful Attention Awareness Scale (MAAS) correlate positively with Positive Self ($r = .35$ to $.45$) and negatively with Negative Self ($r = -.40$ to $-.50$), indicating that mindful non-reactivity and non-judging foster internal friendliness and reduce self-cruelty.
• Rosenberg Self-Esteem Scale (RSES): Positive Self exhibits a robust positive relationship with global self-esteem ($r = .58, p < .001$), while Negative Self shows a strong inverse association ($r = -.62, p < .001$).

7.2. Discriminant Validity

Discriminant validity confirms that the SOFI does not merely duplicate general psychological distress or broad personality styles:

• Social Desirability: Correlations between the SOFI subscales and the Marlowe-Crowne Social Desirability Scale remain uniformly weak ($r = -.12$ to $.18$), establishing that respondents are not merely selecting socially virtuous responses when reporting friendliness or compassion toward self and others.
• Independence of Target Vectors: Correlations between Positive Self and Positive Others, while positive and significant, typically hover between $r = .35$ and $.48$. This demonstrates that an individual can possess high kindness toward others while maintaining low kindness toward themselves (a hallmark of burnout, perfectionism, and caregiver depression). Similarly, correlations between Negative Self and Negative Others range from $r = .25$ to $.38$, demonstrating that self-directed cruelty does not inevitably translate into interpersonal cruelty.

7.3. Predictive and Criterion Validity

The SOFI exhibits strong predictive utility in longitudinal intervention designs:

• Loving-Kindness Meditation (LKM) Interventions: In randomized controlled trials evaluating multi-week LKM and mindfulness interventions, participants show significant increases in Positive Self and Positive Others, alongside sharp reductions in Negative Self scores relative to waitlist controls ($p < .01$, partial $\eta^2 = .14$ to $.22$).
• Symptom Reduction: Reductions in Negative Self scores during psychotherapeutic treatment prospectively predict clinical decreases in depressive severity on the Beck Depression Inventory (BDI-II) and reductions in perceived stress on the Perceived Stress Scale (PSS).

8. Reliability

The psychometric reliability of the SOFI has been supported across college student samples, general adult community cohorts, and clinical treatment populations.

8.1. Internal Consistency

In the initial validation study conducted by Kraus and Sears (2009; $N = 254$), internal consistency reliability coefficients (Cronbach’s α) for all four subscales exceeded the standard psychometric threshold of $.70$ for research and $.80$ for clinical screening:

• Positive Self (PS): α = .86
• Negative Self (NS): α = .85
• Positive Others (PO): α = .80
• Negative Others (NO): α = .82

Subsequent psychometric replications have mirrored these values, with composite reliability ($
ho_c$) and McDonald’s \omega (&\omega;) coefficients consistently ranging between$.81$ and $.88$. Corrected item-total correlations within each four-item subscale routinely surpass $.55$, with no single item removal improving scale alpha, indicating high internal coherence among the selected lexical items.

8.2. Test-Retest Reliability and Stability

Evaluation of temporal stability across non-intervention intervals (e.g., 2-week to 4-week test-retest periods) demonstrates substantial stability coefficients:

• Positive Self: $r_{tt} = .78$
• Negative Self: $r_{tt} = .81$
• Positive Others: $r_{tt} = .74$
• Negative Others: $r_{tt} = .76$

These values demonstrate that the SOFI possesses the stability necessary to capture trait-like dispositions while maintaining the sensitivity to register state-level therapeutic changes in response to contemplative training or clinical psychotherapy.

9. Factor Analysis

The structural integrity and latent dimensionality of the SOFI have been examined using both Exploratory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA).

9.1. Exploratory Factor Analysis (EFA)

During initial scale development, Kraus and Sears examined an initial pool of candidate descriptors derived from Buddhist scholarship. Principal axis factoring with oblique (Promax) rotation was selected due to expected theoretical correlations among positive and negative dimensions. Sample adequacy was supported by Kaiser-Meyer-Olkin (KMO) measures exceeding $.85$ and Bartlett’s Test of Sphericity reaching statistical significance ($p < .001$).

Scree plot examination and parallel analysis consistently isolated four distinct latent factors possessing eigenvalues greater than 1.0, accounting for over 62% of the total variance. The items cleanly loaded onto their intended dimensions without substantial cross-loadings (all primary loadings $ge .62$; secondary loadings $le .25$):

• Factor 1 (Negative Self): Marked by high loadings from Cruel—toward myself (.84), Mean—toward myself (.82), Hateful—toward myself (.80), and Angry—with myself (.71).
• Factor 2 (Positive Self): Marked by high loadings from Friendly—toward myself (.83), Accepting—toward myself (.81), Joyful—for myself (.78), and Compassionate—toward myself (.75).
• Factor 3 (Negative Others): Marked by high loadings from Cruel—toward others (.81), Mean—toward others (.79), Hateful—toward others (.78), and Angry—with others (.68).
• Factor 4 (Positive Others): Marked by high loadings from Friendly—toward others (.80), Compassionate—toward others (.77), Accepting—toward others (.73), and Joyful—for others (.69).

9.2. Confirmatory Factor Analysis (CFA)

To verify the structural validity, Kraus and Sears compared competing theoretical models using maximum likelihood structural equation modeling:

• Model 1: Unidimensional Model. All 16 items loading onto a single general benevolence/malevolence factor. This model demonstrated very poor fit ($\chi^2/df > 6.5$, CFI = .58, RMSEA = .15).
• Model 2: Two-Factor Valence Model. All positive items loading on one factor, all negative items loading on a second factor. Model fit remained unacceptable ($\chi^2/df > 4.2$, CFI = .72, RMSEA = .11).
• Model 3: Two-Factor Target Model. All self items loading on a Self factor, all other items loading on an Others factor. Model fit was poor ($\chi^2/df > 4.8$, CFI = .66, RMSEA = .13).
• Model 4: Four-Factor Correlated Model (PS, NS, PO, NO). This proposed model demonstrated excellent fit to the empirical data across all conventional fit indices: $\chi^2(98) = 142.36, p < .01$; $\chi^2/df = 1.45$; Comparative Fit Index (CFI) = .97; Tucker-Lewis Index (TLI) = .96; Root Mean Square Error of Approximation (RMSEA) = .042 (90% CI [.027, .056]); Standardized Root Mean Square Residual (SRMR) = .044.

The superior fit of Model 4 empirically demonstrates that human affective orientation is bifurcated simultaneously by valence and relational direction.

10. Instrument / Measurement Tool

• Instrument Name: Self–Other Four Immeasurable Scale (SOFI)
• Authors: Salvatore Kraus, Ph.D., and Sharon Sears, Ph.D. (2009)
• Constructs Measured: Positive Self, Negative Self, Positive Others, Negative Others (operationalizing Loving-kindness, Compassion, Joy, Equanimity, and their opposites)
• Test Type: Self-report psychometric rating scale
• Target Population: Adults and adolescents (ages 14+); valid across clinical, contemplative, and general community populations
• Administration Format: Paper-and-pencil questionnaire or computer-based assessment
• Total Item Count: 16 items
• Item Configuration: 4 items per subscale
  • Positive Self: Items 1, 7, 9, 13
  • Positive Others: Items 2, 8, 10, 14
  • Negative Self: Items 3, 5, 11, 15
  • Negative Others: Items 4, 6, 12, 16
• Time Frame: Evaluates experiences during the past week (“during the past week”)
• Authentic Response Scale: 5-point Likert scale:
  • 1 = Very slightly or not at all
  • 2 = A little
  • 3 = Moderately
  • 4 = Quite a bit
  • 5 = Extremely
• Scoring Rules:
  • The SOFI does NOT utilize reverse scoring to create a single composite score. Doing so collapses meaningful orthogonal variance between positive and negative affective states.
  • Four individual subscale scores are derived by calculating either the mean (range: 1.00 to 5.00) or the sum (range: 4 to 20) of the four constituent items for each dimension:
  • Positive Self (PS) Mean: $(\text{Item } 1 + \text{Item } 7 + \text{Item } 9 + \text{Item } 13) / 4$
  • Negative Self (NS) Mean: $(\text{Item } 3 + \text{Item } 5 + \text{Item } 11 + \text{Item } 15) / 4$
  • Positive Others (PO) Mean: $(\text{Item } 2 + \text{Item } 8 + \text{Item } 10 + \text{Item } 14) / 4$
  • Negative Others (NO) Mean: $(\text{Item } 4 + \text{Item } 6 + \text{Item } 12 + \text{Item } 16) / 4$
• Estimated Completion Time: Approximately 2 to 3 minutes

11. Permissions & Fee and Test Year

The Self–Other Four Immeasurable Scale was formally published in 2009 in Social Indicators Research (Springer Science+Business Media). The SOFI is considered an open-access psychometric instrument for non-commercial, scholarly research, and clinical evaluation purposes. No licensing fees or royalties are required to administer the scale for academic studies, clinical trials, or personal mental health monitoring.

Researchers and clinicians utilizing the instrument are expected to maintain the integrity of the 16 original descriptors and cite the primary validation publication (Kraus & Sears, 2009). For commercial applications, diagnostic software integration, or formal test compilation permissions, users should contact the primary author (Dr. Salvatore Kraus, Fort Lewis College: [email protected]) or the copyright holder (Springer Nature).

12. References

• Buddhaghosa, B. (1991). The path of purification (Visuddhimagga) (Bhikkhu Ñ&amacron;ńamoli, Trans., 5th ed.). Buddhist Publication Society. (Original work composed 5th century CE).
• Fredrickson, B. L., Cohn, M. A., Coffey, K. A., Pek, J., & Finkel, S. M. (2008). Open hearts build lives: Positive emotions, induced through loving-kindness meditation, build consequential personal resources. Journal of Personality and Social Psychology, 95(5), 1045–1062. https://doi.org/10.1037/a0013262
• Gilbert, P. (2005). Focused therapies and compassionate mind training for shame and self-attacking. In P. Gilbert (Ed.), Compassion: Conceptualisations, research and use in psychotherapy (pp. 263–325). Routledge.
• Kraus, S., & Sears, S. (2009). Measuring the immeasurables: Development and initial validation of the Self–Other Four Immeasurables (SOFI) Scale based on Buddhist teachings on loving kindness, compassion, joy, and equanimity. Social Indicators Research, 92(1), 169–181. https://doi.org/10.1007/s11205-008-9300-1
• Kraus, S., & Sears, S. (2013). Self–Other Four Immeasurables. In C. A. Simmons & P. Lehmann (Eds.), Tools for strengths-based assessment and evaluation (pp. 179–181). Springer Publishing Company.
• Neff, K. D. (2003). The development and validation of a scale to measure self-compassion. Self and Identity, 2(3), 223–250. https://doi.org/10.1080/15298860309027
• Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief measures of positive and negative affect: The PANAS scales. Journal of Personality and Social Psychology, 54(6), 1063–1070. https://doi.org/10.1037/0022-3514.54.6.1063

13. Items of the Scale (Questionnaire)