Conceptual Style Test

1. Abstract

The Conceptual Style Test (CST), developed by Jerome Kagan, Howard A. Moss, and Irving E. Sigel in 1963 and elaborated by Kagan and colleagues in 1964, is a classic performance-based psychometric instrument designed to evaluate individual differences in cognitive categorization and information-processing strategies in children. Grounded in mid-twentieth-century cognitive style theory, the instrument assesses how individuals selectively attend to, organize, and classify visual stimuli when multiple competing classification rules are simultaneously viable. The CST consists of 30 stimulus cards, each displaying a triad of black-and-white line drawings representing common, familiar objects, animals, or people. The respondent is instructed to select the two stimuli that “are alike or go together” and to articulate the conceptual rationale underlying their grouping.

Responses are scored into three primary conceptual categories: (a) analytic-descriptive, which involves pairings based on shared objective, differentiated parts or visual attributes (e.g., pairing a watch and a ruler because both have numbers); (b) relational-contextual, characterized by thematic, functional, or temporal-spatial interdependence where neither object is an exemplar of the other (e.g., pairing a man and a watch because a man wears a watch); and (c) inferential-categorical, wherein items are grouped as members of a formal taxonomic or superordinate class (e.g., pairing a shirt and a coat because both are articles of clothing). Psychometrically, the CST demonstrated an exceptional corrected split-half reliability coefficient of .94 across a standardization protocol of 300 children. Longitudinal stability coefficients over a one-year interval from Grade 3 to Grade 4 yielded correlations of .70 for girls and .43 for boys. The test served as a foundational empirical bridge connecting perceptual differentiation, attentional deployment, and conceptual categorization, laying the structural groundwork for Kagan’s subsequent formulation of the reflection-impulsivity cognitive tempo dimension.

2. Keywords

Conceptual Style Test, Jerome Kagan, cognitive style, analytic conceptualization, relational conceptualization, inferential-categorical categorization, information processing, cognitive tempo, perceptual differentiation, developmental assessment, categorizing behavior, split-half reliability.

3. Authors

The Conceptual Style Test was collaboratively formulated and empirically validated by a team of prominent developmental psychologists:

• Jerome Kagan, Ph.D. (1929–2021): Serving at the time as Chairman of the Department of Psychology at the Fels Research Institute (Yellow Springs, Ohio) and later as Daniel and Amy Starch Professor of Psychology at Harvard University. Kagan was an internationally recognized pioneer in developmental psychology, temperamental reactivity, and cognitive style development.
• Howard A. Moss, Ph.D.: Research Psychologist at the Fels Research Institute and subsequently a senior investigator at the National Institute of Mental Health (NIMH), recognized for his landmark longitudinal studies on human behavioral stability from infancy to adulthood.
• Irving E. Sigel, Ph.D. (1921–2006): Distinguished Senior Research Scientist at the Merrill-Palmer Institute and later Educational Testing Service (ETS). Sigel devoted his career to exploring children’s representational competence, categorization behaviors, and distancing theory.
• Monograph Collaborators (1964 Validation): Bernice L. Rosman, Ph.D., Deborah Day, Albert J. Albert, and William Phillips, who collaborated with Kagan at the Fels Research Institute to comprehensively map the relations between analytic conceptualization, visual scanning, and reflective attitudes.

4. Purpose

The primary clinical and experimental purpose of the Conceptual Style Test is to assess habitual individual preferences in visual information processing, selective attention, and conceptual grouping in elementary school-aged children. In contrast to standard psychometric intelligence tests such as the Wechsler Intelligence Scale for Children (WISC) or the Stanford-Binet Intelligence Scales—which focus on maximum cognitive capacity, vocabulary breadth, and abstract logical ability—the CST assesses stylistic preference under conditions where multiple solutions are equally valid and correct.

Historically, developmental researchers required a standardized metric capable of disentangling how a child spontaneously parses complex visual environments. When encountering arrays of multifaceted stimuli, a child must choose whether to analyze the components into fine-grained visual parts, link the stimuli through a narrative or functional scene, or apply an abstract taxonomic label. The CST provides an experimental paradigm that elicits these spontaneous cognitive strategies without imposing penalizing constraints.

In educational and developmental research settings, the CST has been utilized for several interconnected applications:

• Delineating Cognitive Tempo: The test identifies children who exhibit an “analytic attitude,” characterized by the deliberate visual decomposition of an array and the extraction of subtle, objective sub-elements. Kagan and colleagues demonstrated that an analytic conceptual style is intrinsically tied to a reflective cognitive tempo (a disposition to withhold immediate judgment and examine visual alternatives carefully), whereas non-analytic (primarily relational) styles correlate with an impulsive cognitive tempo.
• Investigating Reading Readiness and Academic Performance: Because reading and orthographic decoding require children to isolate graphic components (e.g., discerning the visual difference between ‘b’ and ‘d’), the CST serves as a predictive marker for reading acquisition, visual-perceptual discrimination, and complex problem-solving efficacy.
• Assessing Categorization Development: The instrument provides empirical documentation of ontogenetic shifts in classification behavior, charting the transition from contextual-relational thought (predominant in early childhood) to analytic-descriptive and categorical-inferential thinking in middle childhood and adolescence.
• Investigating Clinical and Learning Variations: Clinical researchers have utilized the CST to study cognitive patterning in children with attention-deficit/hyperactivity symptoms, perceptual-motor delays, and academic underachievement, distinguishing between deficits in fundamental intellectual capacity and maladaptive, rapid information-processing styles.

5. Psychological Construct

The construct assessed by the CST is Conceptual Style, defined as an individual’s consistent, self-directed tendency to categorize, classify, and organize multifaceted visual stimuli along specific cognitive dimensions. Unlike cognitive abilities, which measure the upper limits of performance, cognitive styles represent pervasive, habitual modes of information processing that operate across disparate perceptual and intellective tasks. The CST operationalizes this construct across three mutually distinct categorization modes:

Analytic-Descriptive Conceptualization

Analytic-descriptive classification involves the isolation and conceptual pairing of stimuli based upon a shared, objective, observable sub-component or descriptive attribute that forms a constituent part of the total visual presentation. To form an analytic concept, the individual must intentionally segment the visual figure, ignore the salient holistic configuration, and identify an identical attribute embedded within two discrete entities.

Operational Example: Presented with a triad consisting of a man, a watch, and a ruler, an analytic response links the watch and the ruler because “they both have numbers on them.” Another classic example involves a triad containing a zebra, a shirt with vertical stripes, and a plain coat; grouping the zebra and the shirt because “both have stripes” represents an analytic-descriptive grouping. In these instances, the respondent disregards the broader taxonomic identity (living animal vs. inanimate object; measurement tool vs. garment) in favor of visual feature extraction.

Relational-Contextual Conceptualization

Relational-contextual classification occurs when items are grouped together based on a thematic, functional, temporal, or spatial relationship that exists between them within a shared experiential context. In this conceptualization mode, neither object is an independent exemplar or member of the other; rather, they derive their conceptual coherence from interactive, narrative, or contiguous coexistence.

Operational Example: In the same triad displaying a man, a watch, and a ruler, a relational-contextual response pairs the man and the watch on the grounds that “the man wears the watch to tell what time it is,” or “the man uses the watch so he is not late for work.” Similarly, presented with a dog, a bone, and an apple, pairing the dog and the bone because “the dog eats the bone” constitutes an unanalyzed, functional-relational linkage. This conceptualization relies heavily on global, episodic memory rather than structural segmentation.

Inferential-Categorical Conceptualization

Inferential-categorical conceptualization is characterized by the classification of items based on superordinate taxonomic categories or abstract inferential schema. In this mode, each of the paired items represents an independent exemplar of a formal categorical class, and the grouping requires an inferential leap beyond immediate perceptual features.

Operational Example: When presented with a triad consisting of a house, a tent, and a car, pairing the house and the tent because “they are both places where people live” or “both are forms of shelter” exemplifies an inferential-categorical response. Similarly, grouping a hammer and an axe because “they are both tools” represents a taxonomic classification. While abstract, this mode does not require the visual attribute segmentation characteristic of the analytic style, nor does it rely on the situational narratives of the relational style.

6. Theoretical Framework

The Conceptual Style Test was developed at the intersection of developmental psychology, ego psychology, and early cognitive information-processing theory during the early 1960s. The conceptual framework was influenced by several foundational psychological movements:

The Cognitive Control Tradition

During the 1950s and 1960s, researchers such as Herman A. Witkin (pioneering the field dependence-independence dimension) and George S. Klein and Riley W. Gardner (investigating “cognitive controls” such as leveling-sharpening and scanning) proposed that perception is not a passive sensory registration, but an active, personality-permeated cognitive organization. Kagan, Moss, and Sigel extended this paradigm by positing that categorization preferences reflect an underlying regulatory cognitive structure governing how individuals manage visual complexity.

Perceptual Differentiation vs. Global Processing

A core postulate of Kagan’s theoretical framework is the distinction between differentiated (analytic) and global (relational) processing. Analytic categorization demands a high degree of perceptual differentiation: the subject must visually fragment an organized visual gestalt, suppress the immediate prepotent narrative that connects the objects, and identify an isolated identical component across two distinct forms. Conversely, relational categorization reflects a global, unanalyzed response to the environmental field, prioritizing spontaneous contextual associations over visual part-whole segmentation.

The Emergence of Cognitive Tempo (Reflection-Impulsivity)

The theoretical investigation of the CST directly inspired Kagan’s conceptualization of Reflection-Impulsivity. Kagan observed that children who produced high frequencies of analytic-descriptive responses consistently exhibited longer decision latencies and made fewer perceptual errors when presented with complex visual discrimination tasks. Analytic processing necessitates cognitive delay—a period of visual scanning during which hypotheses regarding component features are systematically evaluated. Conversely, relational processing often emerges from immediate, impulsive judgments driven by the first thematic association that enters consciousness. Consequently, the theoretical architecture of the CST provided the empirical foundation for the subsequent design of the Matching Familiar Figures Test (MFFT).

7. Validity

The validity of the Conceptual Style Test as a measure of analytic conceptualization has been established across developmental, experimental, and psychometric investigations, although Kagan and his contemporaries recognized certain methodological nuances inherent in ipsative categorization measures.

Construct and Developmental Validity

Construct validity is supported by developmental progression. In longitudinal and cross-sectional investigations conducted by Kagan et al. (1964) and Sigel (1963), the frequency of analytic-descriptive responses increased steadily across the elementary school years (Grades 1 through 6). Young children (ages 5 to 7) exhibited a predominance of relational-contextual responses, reflecting reliance on concrete, narrative, and contextual cognitive representations. As children mature, cognitive maturation and formal schooling facilitate visual differentiation, leading to a marked increase in analytic-descriptive and inferential-categorical categorizations.

Convergent Validity

Convergent validity was demonstrated through systematic correlations with external measures of visual analysis, perceptual restructuring, and cognitive tempo:

• Perceptual Disembedding Tasks: Kagan et al. (1964) found statistically significant positive correlations between CST analytic scores and performance on the Children’s Embedded Figures Test (CEFT) and Gottschaldt Hidden Figures ($r$ values ranging from .30 to .52 across elementary cohorts). Children who favored analytic groupings on the CST were substantially more proficient at detecting simple geometrical figures embedded within complex backgrounds.
• Performance Intelligence Subtests: Moderate positive correlations ($r = .25$ to $.45$) were consistently observed between CST analytic scores and the Block Design and Object Assembly subtests of the WISC, both of which require spatial analysis and the perceptual breakdown of visual patterns into discrete components.
• Cognitive Reflection: CST analytic scores correlated significantly with longer response latencies and lower error rates on the Matching Familiar Figures Test (MFFT), confirming that an analytic conceptual style is intrinsically tied to a reflective cognitive orientation.

Discriminant Validity

To demonstrate that the CST does not merely replicate general psychometric intelligence, Kagan and colleagues controlled for verbal IQ (as measured by the WISC Verbal scale and the Stanford-Binet). In multiple regression and partial correlation analyses, the relationship between analytic conceptual style and visual-perceptual performance remained statistically significant after partialing out verbal IQ. This established that the CST captures a distinct cognitive style rather than raw verbal knowledge or semantic fluency.

8. Reliability

The psychometric evaluation of the Conceptual Style Test has focused primarily on internal consistency and longitudinal stability, with particular attention to the analytic-descriptive category.

Internal Consistency (Split-Half Reliability)

In the primary psychometric monograph by Kagan, Rosman, Day, Albert, and Phillips (1964), split-half reliability was evaluated using a standardization sample of 300 elementary school children. Protocols were divided into odd- and even-numbered stimulus triads:

• The raw correlation between split halves for the analytic-descriptive score was exceptionally robust, yielding a corrected split-half reliability coefficient of .94 using the Spearman-Brown prophecy formula.
• This high internal consistency confirmed that across the 30 visual triads, the propensity of children to isolate objective visual details operated as a coherent, unified cognitive behavioral pattern.

Test-Retest and Longitudinal Stability

Longitudinal stability was assessed across a one-year interval. A cohort of children initially evaluated in Grade 3 was readministered the CST in Grade 4:

• For girls ($N = 24$), the stability coefficient over the one-year interval was $r = .70$ ($p < .001$), demonstrating high stylistic stability.
• For boys ($N = 22$), the one-year stability coefficient was lower, yielding $r = .43$ ($p < .05$).
• Kagan attributed this sex difference to differential rates of behavioral maturation, socialization regarding attentional control, and fluctuating impulsivity patterns in boys during the middle elementary school period.

Inter-Rater Scoring Reliability

Because the CST relies on the examiner’s qualitative classification of the child’s verbal rationale, inter-rater reliability is critical. In independent investigations by Kagan et al. (1964) and subsequent replications (e.g., Davis, 1971), inter-scorer agreement across two independent raters coding verbal explanations into the three conceptual categories (analytic, relational, categorical) consistently exceeded 90% agreement (percentage agreement ranging from 92% to 96%; Cohen’s $kappa > .88$).

9. Factor Analysis

In the initial psychometric publications of the 1960s, formal confirmatory factor analysis (CFA) had not yet emerged in developmental psychology; empirical evaluations of the CST’s internal structure relied on exploratory factor analysis (EFA), inter-category correlation matrices, and principal component extractions.

Structural Nature of Categorization Scores

A fundamental structural characteristic of the CST is its ipsative nature. Because each of the 30 triad cards permits only a single classification choice, scoring more analytic-descriptive items mechanically constrains the total number of items available to be classified as relational-contextual or inferential-categorical. Factor analyses of the correlation matrices across CST indices consistently reveal the following structural dynamics:

• Bipolar Primary Factor (Analytic vs. Relational): Exploratory factor analyses of the CST response modalities typically yield a dominant bipolar factor accounting for the largest proportion of total variance. The analytic-descriptive score loads heavily in one direction (loadings typically exceeding $+.75$), whereas the relational-contextual score loads heavily in the inverse direction (loadings typically exceeding $-.70$). This confirms that the primary axis of individual variation on the CST is a continuum between perceptual-analytic differentiation and global-relational integration.
• Secondary Taxonomic Factor: The inferential-categorical score frequently isolates onto a secondary, independent factor or displays modest, positive loadings across both poles depending on the child’s age and verbal capability. In older elementary school children, a two-factor orthogonal model cleanly separates perceptual-analytic processing from verbal-categorical abstracting.

Multidimensional Considerations

Subsequent psychometric critics (e.g., Messick, 1976; Kogan, 1971) observed that when experimental variants of the CST allow multiple groupings per card (free-sorting paradigms rather than forced single-choice triads), the strong negative correlation between analytic and relational styles is attenuated. Under non-ipsative conditions, analytic capability and relational conceptualization can operate as partially independent cognitive modes, though the spontaneous preference elicited in forced-choice visual triads remains structurally bipolar.

10. Instrument / Measurement Tool

The Conceptual Style Test is an individually administered, performance-based cognitive test. Below is a structured summary of its administrative and structural parameters:

• Instrument Type: Standardized performance-based cognitive style test; triad picture classification.
• Format: Physical stimulus cards (traditionally $8 \times 10$ inches or $5 \times 8$ inches), each containing three black-and-white line drawings arranged horizontally or in a triangular configuration.
• Item Count: 30 stimulus triad cards (plus introductory practice items).
• Target Population: Human; male and female children (primarily ages 6 to 12; elementary school grades 1 through 6). Variants have occasionally been adapted for preschool children and adults.
• Administration Method: Individual, examiner-administered test. The examiner presents cards sequentially, records the child’s physical pointing choices, and transcribes the verbatim verbal explanation provided for each pair.
• Administration Protocol:
  • The examiner places each card before the child and delivers the standardized prompt: “Look at these pictures. Two of these pictures are alike or go together in some way. Show me the two that are alike or go together, and tell me why they go together.”
  • The child must point to two pictures and state the rationale.
  • Neutral probing (e.g., “In what way do they go together?”) is permitted only if the child’s initial explanation is ambiguous.
• Scoring System: Each of the 30 responses is scored into one of three mutually exclusive categories:
  • Analytic-Descriptive (A): Grouping based on an objective, shared visual part or concrete perceptual attribute (e.g., “Both have claws,” “Both have wheels,” “Both have a broken handle”). Total score = sum of A responses (0–30).
  • Relational-Contextual (R): Grouping based on functional, situational, or narrative interdependence where neither is a sub-part of the other (e.g., “The horse pulls the wagon,” “The boy flies the kite”). Total score = sum of R responses (0–30).
  • Inferential-Categorical (C): Grouping based on common membership in a taxonomic or superordinate category (e.g., “Both are animals,” “Both are furniture,” “Both are weapons”). Total score = sum of C responses (0–30).
  • Unclassifiable / Error: Failure to produce two items, bizarre rationales, or simple repetition of names without a conceptual link (scored as 0 or unclassified).

11. Permissions & Fee and Test Year

• Test Year: 1963 (initial formulation by Kagan, Moss, & Sigel); fully standardized and published in monograph form in 1964.
• Original Publication Source: Kagan, J., Rosman, B. L., Day, D., Albert, J., & Phillips, W. (1964). Information processing in the child: Significance of analytic and reflective attitudes. Psychological Monographs: General and Applied, 78(1), 1–37.
• Permissions and Availability: The Conceptual Style Test was developed under academic research auspices supported by the Fels Fund and the United States Public Health Service (National Institute of Mental Health). The test items and pictorial stimulus cards are protected by original historical copyright (American Psychological Association / Authors) and are not distributed as a commercial psychometric test kit.
• Commercial Status: Non-commercial instrument.
• User Fee: No licensing fees are charged for academic or educational research. Qualified researchers wishing to reconstruct or access archival versions of the pictorial cards must consult the original 1964 monograph or request archival materials through academic libraries and research repositories holding the Jerome Kagan archives.

12. References

• Davis, A. J. (1971). Cognitive style in preschool children: Educable mentally retarded, normal, and gifted. Perceptual and Motor Skills, 33(3_suppl), 1039–1045. https://doi.org/10.2466/pms.1971.33.3f.1039
• Kagan, J., Moss, H. A., & Sigel, I. E. (1963). Psychological significance of styles of conceptualization. In J. C. Wright & J. Kagan (Eds.), Basic cognitive processes in children. Monographs of the Society for Research in Child Development, 28(2, Serial No. 86), 73–112. https://doi.org/10.2307/1165657
• Kagan, J., Rosman, B. L., Day, D., Albert, J., & Phillips, W. (1964). Information processing in the child: Significance of analytic and reflective attitudes. Psychological Monographs: General and Applied, 78(1), 1–37. https://doi.org/10.1037/h0093830
• Kogan, N. (1971). Educational implications of cognitive styles. In G. S. Lesser (Ed.), Psychology and educational practice (pp. 242–292). Scott, Foresman.
• Messick, S. (1976). Individuality in learning: Implications of cognitive styles and creativity for human development. Jossey-Bass.
• Sigel, I. E. (1963). How intelligence tests limit understanding of intelligence. Merrill-Palmer Quarterly of Behavior and Development, 9(1), 39–56. https://www.jstor.org/stable/23034038
• Witkin, H. A., Dyk, R. B., Faterson, H. F., Goodenough, D. R., & Karp, S. A. (1962). Psychological differentiation: Studies of development. John Wiley & Sons.

13. Items of the Scale