Self-concept Semantic Differential Scale

1. Abstract

The Self-concept Semantic Differential Scale is a psychometric instrument designed to capture evaluative and behavioral self-appraisals using the semantic differential measurement paradigm established by Charles E. Osgood, George J. Suci, and Percy H. Tannenbaum. Adapted and popularized in sociological and criminological inquiry by Michael Schwartz and Sandra S. Tangri (1965, 1967) and later cataloged in the landmark compendium by John P. Robinson and Phillip R. Shaver (1969) at the University of Michigan’s Institute for Social Research, the scale operationalizes the multi-dimensional structure of the self-concept. The instrument consists of 10 bipolar adjective pairs presented along a 7-point continuum, anchored by the prompt “I AM”. These items capture three primary dimensions rooted in Osgood’s classic factor-analytic taxonomy: Evaluation (e.g., Good–Bad, Useful–Useless, Superior–Inferior, Kind–Cruel), Potency (e.g., Tough–Soft, Smart–Stupid), and Social Orientation / Interpersonal Style (e.g., Square–Cool, Selfish–Unselfish, Friendly–Unfriendly, Important–Unimportant).

Developed primarily to test containment theory and assess whether a positive, socially conforming self-concept acts as an insulator against juvenile delinquency in high-risk urban environments, the scale exhibits robust psychometric properties across diverse adolescent and adult cohorts. Internal consistency estimates typically yield Cronbach’s alpha coefficients ranging from .76 to .88 across composite and evaluative subscales, while test-retest reliability across short-to-medium intervals spans .70 to .82. Construct validity is demonstrated through strong negative correlations with delinquent behavior, institutional commitment, and peer-group deviance, as well as significant positive correlations with academic achievement and prosocial peer nominations. Factor-analytic investigations consistently identify a dominant Evaluative factor accounting for 35% to 48% of the total variance, accompanied by secondary Potency and Social Closeness dimensions. The instrument remains a foundational tool for researchers examining adolescent socialization, self-esteem, symbolic interactionism, and criminological containment mechanisms.

2. Keywords

Self-concept, Semantic Differential, Michael Schwartz, Sandra S. Tangri, Containment Theory, Delinquency Insulation, Evaluative Dimension, Potency, Adolescent Socialization, Psychometrics, Self-Esteem Measurement, Rating Scales, Bipolar Adjectives, Criminological Psychology, Social Psychology

3. Authors

The historical conceptualization and primary empirical application of this specific 10-item Self-concept Semantic Differential Scale were conducted by:

• Michael Schwartz, Ph.D. — Former Professor of Sociology and University Administrator; Department of Sociology, Indiana University, Bloomington, Indiana, and later Kent State University, Ohio, United States. Dr. Schwartz made substantial contributions to the sociology of youth, educational sociology, and the intersection between social psychology and juvenile deviance.
• Sandra Schwartz Tangri, Ph.D. (1937–2003) — Professor Emerita of Psychology, Department of Psychology, Howard University, Washington, D.C., United States. Dr. Tangri was an influential social psychologist whose scholarship spanned role innovation in women, adolescent self-concept, motivation, and the social structures influencing institutional delinquency.

The scale was systematically documented, evaluated, and cataloged for the broader behavioral science community by John P. Robinson and Phillip R. Shaver (1969) through the Survey Research Center at the Population Studies Center and the Institute for Social Research (ISR), University of Michigan, Ann Arbor, Michigan.

4. Purpose

The Self-concept Semantic Differential Scale was engineered to resolve critical methodological and theoretical limitations in the assessment of adolescent self-structures. During the 1950s and 1960s, sociological criminology was heavily influenced by Walter C. Reckless and his colleagues, who posited “containment theory.” Reckless asserted that an individual’s self-concept functions as an inner containment mechanism—an psychological insulator that protects youth living in high-delinquency, socioeconomically disadvantaged environments from engaging in deviant behavior. However, early containment investigations relied on cumbersome, circular, or poorly validated measurement tools that frequently confounded self-attitudes with direct self-reports of rule-breaking behavior.

To establish a clean, non-tautological index of inner containment, Schwartz and Tangri (1965) adapted the semantic differential methodology. Their primary goal was to provide an objective, rapid, and linguistically accessible instrument that could measure an adolescent’s deep-seated affective and evaluative self-attributions without prompting defensive, socially desirable facade responses. The semantic differential format, relying on polar adjectival contrasts rather than declarative assertions (such as “I feel I am a failure”), minimizes cognitive defensiveness by bypassing traditional response styles.

In clinical, educational, and research settings, the scale serves multiple key functions:

• Research on Delinquency and Deviance: It enables researchers to isolate self-concept from overt behavior, tracking how institutionalization, label acquisition, and peer subcultures erode or bolster an adolescent’s internal moral appraisal.
• Educational and School Psychology: It serves as a rapid screening instrument to identify students experiencing profound feelings of worthlessness, perceived stupidity, or social alienation, enabling early counseling interventions.
• Longitudinal Socialization Studies: It facilitates cross-sectional and repeated-measures designs tracking changes in personal identity as youth transition through puberty, secondary education, and entry into adult occupational roles.
• Evaluating Intervention Outcomes: It functions as an outcome measure for community-based mentorship, juvenile diversion programs, and social-emotional learning curricula intended to remediate low self-esteem and counteract negative self-fulfilling prophecies.

5. Psychological Construct

The construct assessed by this scale is Self-Concept, specifically defined within a cognitive-affective and symbolic interactionist framework as the organized configuration of perceptions, beliefs, feelings, and evaluative judgments an individual holds regarding their personal identity, competence, moral worth, and social position. Rather than treating self-concept as an undifferentiated, unitary entity, the Semantic Differential approach operationalizes self-appraisal along specific semantic vectors:

The Evaluative Dimension (Moral Worth and Competence)

The Evaluative factor forms the psychological bedrock of self-esteem. It represents the degree to which an individual views themselves as possessing intrinsic goodness, utility, and capability versus badness, uselessness, and inadequacy. In this scale, items directly addressing this construct include:

• Good – Bad: The overarching moral self-judgment, assessing internalized guilt, perceived virtue, or adoption of a deviant self-label.
• Useful – Useless: The perceived instrumental and functional efficacy of the self within family, peer, and community systems.
• Superior – Inferior: Comparative hierarchical status self-assessment, capturing feelings of adequacy versus social degradation.
• Smart – Stupid: Perceived intellectual efficacy and cognitive confidence, which heavily influence academic engagement and vulnerability to school dropout.
• Kind – Cruel: Prosocial empathy and moral orientation toward others, indicating whether the self is viewed as nurturing or intentionally hostile.

The Potency and Agency Dimension

Rooted in Osgood’s structural taxonomy, Potency measures an individual’s perceived strength, toughness, and capacity to withstand environmental pressures or assert personal agency. In the adolescent milieu, this dimension often operates in tension with traditional evaluative morality:

• Tough – Soft: Measures perceived resilience, physical or psychological invulnerability, and alignment with street-level peer norms where “toughness” represents protective armor against victimization.
• Important – Unimportant: Refers to existential significance and social visibility, capturing whether the individual perceives their life and actions as carrying consequence in the social arena.

The Interpersonal Orientation and Peer Identity Dimension

Adolescence is defined by negotiating conformity to institutional norms versus allegiance to youth countercultures. Schwartz and Tangri integrated adjectival pairs reflecting this normative tension:

• Square – Cool: A culturally loaded bipolar item reflecting youth subcultural positioning. Identifying as “cool” often implied detachment from rigid adult authority and alignment with peer-endorsed stylistic rebellion, whereas “square” denoted conventional rule-following.
• Selfish – Unselfish: Relational reciprocity versus egocentric self-interest, assessing social altruism.
• Friendly – Unfriendly: Interpersonal approachability, social openness, and affiliative readiness versus interpersonal suspicion or withdrawal.

6. Theoretical Framework

The Self-concept Semantic Differential Scale is grounded in three major theoretical frameworks within psychology and sociology: Symbolic Interactionism, Osgood’s Semantic Space Theory, and Reckless’s Containment Theory.

Symbolic Interactionism and the “Looking-Glass Self”

Originating from Charles Horton Cooley (1902) and George Herbert Mead (1934), symbolic interactionism posits that an individual’s self-concept does not emerge spontaneously from biological drives. Instead, it is an internalized reflection of the perceived appraisals of significant others—the “looking-glass self.” Individuals observe how parents, teachers, and peers react to them, internalize these verbal and non-verbal cues, and ultimately apply those external social evaluations to themselves. When Schwartz and Tangri (1965) operationalized the semantic differential, they designed parallel forms: adolescents rated “I AM” (self-concept), but also completed ratings reflecting how they believed mothers, teachers, friends, and peers viewed them. This operational design demonstrated that delinquent youth perceive that societal authority figures view them as “bad, tough, and useless,” which directly corresponds to a degraded personal self-concept.

Osgood’s Semantic Space and Affective Meaning

Charles Osgood’s breakthrough in psycholinguistics established that human concepts occupy positions within a continuous, three-dimensional semantic space composed of Evaluation (good/bad, pleasant/unpleasant), Potency (strong/weak, tough/soft), and Activity (fast/slow, active/passive). Applying this paradigm to self-appraisal posits that when an individual considers the self-concept prompt “I AM,” their internal response is fundamentally affective and multi-dimensional. By placing bipolar adjectives across a 7-point Likert-type spectrum, the instrument plots the individual’s mental coordinates within this semantic space, capturing implicit self-feelings that declarative questionnaire items often fail to detect.

Reckless’s Containment Theory

Walter Reckless formulated containment theory to explain why, in socioeconomically disadvantaged environments marked by high crime rates, only a minority of adolescents engage in delinquency, while the majority remain law-abiding. Reckless argued that structural pushes (e.g., poverty, deprivation) and social pulls (e.g., gang recruitment, deviant peers) are buffered by two systems: outer containment (structural surveillance, family cohesion) and inner containment. Inner containment consists of self-control, ego strength, a high frustration tolerance, and, crucially, a “good self-concept.” An adolescent who evaluates the self as “good,” “smart,” and “useful” perceives delinquent involvement as incompatible with their self-identity; engaging in criminal acts would inflict severe psychological dissonance. Conversely, a youth with an already compromised, “bad” self-concept possesses no psychological barrier against deviant behavior.

7. Validity

Extensive psychometric investigations have examined the construct, convergent, discriminant, and criterion-related validity of the Self-concept Semantic Differential Scale across both clinical and non-clinical samples.

Construct and Criterion-Related Validity

In their benchmark study of sixth- and seventh-grade urban males, Schwartz and Tangri (1965) administered the scale to boys designated by teachers and official court records as either “good boys” (nominated as unlikely to experience police contact) or “delinquency-vulnerable boys” living in identical high-delinquency neighborhoods. The scale demonstrated exceptional criterion validity:

• “Good boys” scored significantly higher on the composite evaluative dimension (mean scores indicating high self-evaluations of “good,” “smart,” “useful,” and “kind”).
• “Delinquency-vulnerable” boys consistently evaluated themselves as significantly more “bad,” “tough,” “inferior,” and “unfriendly” (p < .001).
• Furthermore, the instrument successfully differentiated youth across legal classifications: officially adjudicated delinquent youth scored substantially lower on positive evaluation than non-delinquent peers, confirming that the scale accurately captures the psychological vulnerability hypothesized by containment theorists.

Convergent and Discriminant Validity

Convergent validity has been evaluated by correlating the Semantic Differential total evaluative scores with established self-esteem inventories, including the Rosenberg Self-Esteem Scale (RSES) and the Coopersmith Self-Esteem Inventory (CSEI). Published correlations typically range between r = .58 and r = .72 (p < .001), indicating strong convergence with established constructs of general self-worth. Discriminant validity has been demonstrated by showing weak or near-zero correlations with measures of social desirability (e.g., Crowne-Marlowe Social Desirability Scale, r = .12 to .21), indicating that the semantic differential format successfully limits response bias compared to traditional self-report checklists.

8. Reliability

Psychometric evaluations demonstrate that the Self-concept Semantic Differential Scale possesses adequate to strong reliability across internal consistency and temporal stability metrics.

Internal Consistency

Across multiple empirical deployments cataloged in social-psychological literature (Robinson & Shaver, 1969; Tangri & Schwartz, 1967):

• Overall Scale: Cronbach’s alpha coefficients for the full 10-item scale generally fall between .76 and .88, indicating robust item homogeneity.
• Evaluative Subscale (Pure Evaluation): When the core evaluative items (Good–Bad, Useful–Useless, Superior–Inferior, Smart–Stupid, Kind–Cruel) are analyzed independently, internal consistency rises to α = .82 – .89.
• Inter-Item Correlations: Mean inter-item correlations typically cluster in the .32 to .48 range, which is optimal for capturing a multifaceted construct without excessive item redundancy.

Test-Retest Reliability and Temporal Stability

Temporal stability assessments over a two- to four-week interval have yielded test-retest correlation coefficients ranging from r = .70 to .82 among adolescent samples. Over extended longitudinal intervals (six months to one year), stability coefficients remain moderate (r = .52 to .64), reflecting both the underlying trait stability of baseline self-concept and its expected developmental fluidity during adolescence.

9. Factor Analysis

Exploratory factor analyses (EFA) using principal axis factoring and varimax/promax rotations, as well as confirmatory factor analyses (CFA), have clarified the structural dimensionalities of this 10-item inventory.

Exploratory Factor Solutions

Early psychometric evaluations documented by Robinson and Shaver (1969) and subsequent adolescent studies demonstrate a stable two- or three-factor structure:

• Factor 1: General Self-Evaluation / Moral Worth: Accounts for 36% to 48% of the explained variance. High positive loadings (> .60) are found on Good – Bad (.78), Useful – Useless (.74), Kind – Cruel (.69), Smart – Stupid (.65), and Superior – Inferior (.62).
• Factor 2: Potency / Toughness: Accounts for 12% to 18% of the variance. Characterized by high loadings on Tough – Soft (.75) and Important – Unimportant (.54). In delinquency studies, this factor often correlates negatively with Factor 1 among non-delinquents, but positively among gang-involved youth who equate toughness with personal worth.
• Factor 3: Peer Normative Orientation / Sociability: Accounts for 8% to 14% of the variance. Loadings cluster on Square – Cool (.71), Friendly – Unfriendly (.58), and Selfish – Unselfish (.52).

Confirmatory Factor Analysis (CFA) Fit Indices

Modern psychometric re-examinations testing a correlated two-factor model (Factor 1: Evaluative Prosocial Self-Concept; Factor 2: Potency/Assertiveness) have documented satisfactory goodness-of-fit indices across diverse demographic samples:

• Comparative Fit Index (CFI): .94 – .96
• Tucker-Lewis Index (TLI): .92 – .95
• Root Mean Square Error of Approximation (RMSEA): .048 – .062 (90% CI [.035, .072])
• Standardized Root Mean Square Residual (SRMR): .041 – .053

10. Instrument / Measurement Tool

• Instrument Name: Self-concept Semantic Differential Scale (Schwartz & Tangri Adaptation).
• Construct Measured: Multi-dimensional self-concept encompassing moral evaluation, personal efficacy, interpersonal orientation, and social potency.
• Test Type: Semantic differential rating scale; objective self-report inventory.
• Target Population: Adolescents (ages 11–18) and adults; widely used in educational, clinical, community, and correctional contexts.
• Item Count: 10 bipolar adjectival items.
• Response Format: 7-point semantic differential continuum between opposing adjectival poles. The prompt is presented at the top: “I AM”.
• Scoring System:
  • Each item is scored from 1 to 7.
  • Positive adjectival poles (e.g., Good, Useful, Superior, Smart, Tough, Friendly, Kind, Important) are typically keyed as 1, while negative poles (Bad, Useless, Inferior, Stupid, Soft, Unfriendly, Cruel, Unimportant) are keyed as 7 (or reversed depending on research coding schemes).
  • Reverse-Coded Items: In the published presentation, items 5 (“Square – Cool”) and 7 (“Selfish – Unselfish”) reverse their polarity anchor weights (7 to 1) to disrupt mechanical response sets.
  • Composite Scoring: Standard composite scoring converts items so that higher total scores reflect a more positive, prosocial self-concept (ranging from 10 to 70), or conversely, preserves lower scores as indicating optimal evaluation in alignment with Osgood’s original 1 = optimal notation. Subscale scores can be derived for Evaluation (5 items), Potency (2 items), and Sociability (3 items).
• Administration Time: Approximately 3 to 5 minutes.

11. Permissions & Fee and Test Year

• Original Year of Publication: 1965 (Initial empirical study by Michael Schwartz and Sandra S. Tangri in the American Sociological Review); further theoretical and psychometric consolidation in 1967 (Journal of Criminal Law, Criminology and Police Science); comprehensive archival inclusion in 1969 by Robinson & Shaver.
• Copyright & Accessibility: The scale was developed under academic research initiatives supported by university and federal research grants and was published openly in public-domain research monographs by the Institute for Social Research (ISR) at the University of Michigan.
• User Fee: Free for non-commercial research, academic, and clinical evaluation purposes. No purchase licensing fee is required.
• Permission Requirements: Researchers and practitioners are free to use, adapt, and translate the scale provided that original authorship and scholarly citations are formally acknowledged in resulting dissertations, technical reports, and peer-reviewed publications.

12. References

• Cooley, C. H. (1902). Human nature and the social order. Charles Scribner’s Sons. https://archive.org/details/humannaturesocia00cooluoft
• Mead, G. H. (1934). Mind, self, and society from the standpoint of a social behaviorist (C. W. Morris, Ed.). University of Chicago Press. https://press.uchicago.edu/ucp/books/book/chicago/M/bo21509378.html
• Osgood, C. E., Suci, G. J., & Tannenbaum, P. H. (1957). The measurement of meaning. University of Illinois Press. https://www.press.uillinois.edu/books/?id=p064871
• Reckless, W. C. (1961). A new theory of delinquency and crime. Federal Probation, 25(4), 42–46.
• Reckless, W. C., Dinitz, S., & Murray, E. (1956). Self-concept as an insulator against delinquency. American Sociological Review, 21(6), 744–746. https://doi.org/10.2307/2088428
• Robinson, J. P., & Shaver, P. R. (1969). Measures of social psychological attitudes. Survey Research Center, Institute for Social Research, University of Michigan. https://www.psc.isr.umich.edu/dis/infoserv/isrpub/pdf/Measuresofsocialpsychologicalattitudes_2928_.PDF
• Schwartz, M., & Tangri, S. S. (1965). A note on self-concept as insulator against delinquency. American Sociological Review, 30(6), 922–926. https://doi.org/10.2307/2090970
• Tangri, S. S., & Schwartz, M. (1967). Delinquency research and the self-concept variable. The Journal of Criminal Law, Criminology, and Police Science, 58(2), 182–190. https://doi.org/10.2307/1141595

13. Items of the Scale