Measure of Original Thinking in Elementary School (MOTES)

1. Abstract

The Measure of Original Thinking in Elementary School (MOTES) (Acar et al., 2024) represents a significant advancement in computational psychometrics and educational assessment, specifically designed to capture the latent construct of original thinking in young learners. Traditional divergent thinking assessments have long faced psychometric and logistical constraints, including confounding fluency effects, labor-intensive manual scoring, and developmental mismatches when administered to pediatric cohorts. Developed to overcome these systemic challenges, MOTES is a computerized, gamified testing battery comprising 24 open-ended items distributed evenly across three distinct task subscales: Uses (8 items), Examples (8 items), and Sentences (8 items). The assessment prompts elementary school children to produce unique, creative responses within interactive digital environments, which can subsequently be evaluated through both expert human ratings and advanced automated natural language processing (NLP) pipelines. Psychometric validation on diverse samples of elementary students across the United States demonstrates exceptional structural and internal consistency reliability, yielding maximal reliability coefficients (Hancock’s H) of 0.82 for Uses, 0.85 for Examples, 0.88 for Sentences, and 0.89 for a general higher-order Originality factor (McDonald’s $\omega = .87$). Automated artificial intelligence scoring algorithms demonstrated high convergence with trained human raters, displaying correlations ranging from $r = .79$ to $.91$. Confirmatory factor analyses support a hierarchical model wherein individual task performance loads onto respective domain facets, which in turn define a overarching original thinking latent trait. MOTES demonstrates robust external and criterion validity through significant correlations with established divergent thinking metrics—notably the Torrance Tests of Creative Thinking (TTCT)—and standardized academic performance in mathematics and reading (Measures of Academic Progress; MAP). This assessment provides educators and researchers with a reliable, ecologically valid, and scalable methodology for diagnosing creative potential, evaluating instructional interventions, and cultivating 21st-century cognitive capacities in primary education.

2. Keywords

Measure of Original Thinking in Elementary School, MOTES, Divergent Thinking, Originality, Computational Psychometrics, Automated Creativity Scoring, Elementary Education, Natural Language Processing, Latent Semantic Analysis, Differential Item Functioning, Torrance Tests of Creative Thinking

3. Authors

The Measure of Original Thinking in Elementary School was conceptualized, designed, and psychometrically validated through a collaborative cross-institutional research initiative by leading scholars in educational psychology, psychometrics, and computational linguistics:

• Selcuk Acar, Ph.D. — Department of Educational Psychology, College of Education, University of North Texas, Denton, TX, USA. Specialist in gifted education, creativity assessment, and psychometric modeling of divergent thinking. (ORCID).
• Denis Dumas, Ph.D. — Department of Educational Psychology, Mary Frances Early College of Education, University of Georgia, Athens, GA, USA. Expert in quantitative methodology, latent variable modeling, and cognitive measurement. (ORCID).
• Peter Organisciak, Ph.D. — Department of Research Methods and Information Science, Morgridge College of Education, University of Denver, Denver, CO, USA. Specialist in computational linguistics, information retrieval, and automated text scoring algorithms. (ORCID).
• Kelly Berthiaume, Ph.D. — Educational researcher and assessment developer focusing on cognitive developmental measures and measurement invariance in childhood education.

4. Purpose

The primary purpose of the Measure of Original Thinking in Elementary School (MOTES) is to provide an empirically grounded, developmentally calibrated, and computationally scalable assessment of creative ideation in children between the ages of 6 and 12 years (grades 1 through 6). The cultivation of creativity and divergent thinking has emerged as an indispensable educational priority in the 21st century. Despite universal acknowledgment of its importance, systematic assessment of creative ideation in young populations has been fundamentally hindered by the limitations of legacy psychometric instruments. Historical measures—such as the classic Torrance Tests of Creative Thinking (TTCT) or the Alternative Uses Task (AUT)—were predominantly designed for older children, adolescents, or adults, requiring extensive text-generation capacities, open-ended narrative writing, and protracted manual scoring procedures that introduce severe human rater fatigue and subjective bias.

MOTES resolves these long-standing operational bottlenecks by merging modern computational psychometrics with child-centered human-computer interaction principles. Specifically, MOTES addresses four central theoretical, clinical, and pedagogical needs:

1. Mitigating the Fluency-Originality Confound: In classical divergent thinking batteries, respondents are instructed to list as many ideas as possible within a fixed timeframe. Consequently, indices of originality are heavily confounded by ideational fluency (the sheer volume of generated responses), confounding linguistic processing speed and fine-motor handwriting speed with underlying creative capacity. MOTES bypasses this confound by adopting a single-response-per-prompt paradigm across multiple tightly controlled prompts, directly isolating the statistical rarity and conceptual distance of the generated idea without penalizing slow typists or rewarding high-volume associative chaining.
2. Child-Centric Task Architecture: Young elementary students frequently struggle with the abstract or decontextualized prompts characteristic of adult creativity tasks. MOTES frames assessment prompts within interactive, gamified computerized scenarios that sustain engagement, reduce test anxiety, and match the executive functioning limitations typical of late childhood.
3. Integration of Automated Natural Language Processing (NLP): Evaluating thousands of qualitative responses has historically made large-scale creativity screening in public schools logistically impractical. MOTES was engineered from inception to interface with distributional semantic models (e.g., GloVe, fastText, Word2Vec, and transformer-based architectures), enabling instantaneous, objective, and reproducible originality scoring based on semantic distance calculations.
4. Educational and Diagnostic Applications: Clinically and diagnostically, MOTES serves as a dynamic screening battery for identifying gifted and talented students from non-traditional or underrepresented backgrounds who may fail standardized multiple-choice intelligence tests due to language or cultural barriers. Pedagogically, it provides classroom teachers and educational researchers with granular baseline and post-intervention metrics to evaluate the efficacy of creativity-enhancing curricula, project-based learning initiatives, and cognitive flexibility programs.

5. Psychological Construct

The psychological construct evaluated by MOTES is Original Thinking, conceptualized as a primary sub-component of divergent thinking and broader human creativity. Historically defined by J.P. Guilford and E. Paul Torrance, original thinking denotes the mental operation of producing responses, cognitive representations, or associations that are statistically infrequent within a reference population, conceptually novel, and contextually appropriate. In MOTES, original thinking is operationalized not merely as idiosyncratic or bizarre responding, but as meaningful semantic divergence—the cognitive capacity to navigate away from dominant, high-frequency mental schemas to retrieve, combine, and synthesize distant associative nodes in semantic memory.

To comprehensively sample the multifaceted expressions of original thinking across verbal and cognitive domains in childhood, MOTES bifurcates the overarching originality construct into three distinct task dimensions:

5.1. Uses (Alternative Conceptual Repurposing)

The Uses subscale represents a developmentally adapted variation of the Alternative Uses Task. In this dimension, children are presented with common everyday objects (such as an empty tin can, a wooden pencil, or a plastic bucket) and prompted to devise clever, unusual, and highly original functions for the item beyond its intended conventional utility. Cognitively, this subscale captures cognitive flexibility and the suppression of functional fixedness. An unoriginal response relies entirely on dominant affordances (e.g., using a bucket to hold water), representing a baseline associative retrieval. Conversely, a highly original response requires cognitive restructuring of the physical attributes of the object (e.g., turning the bucket upside down and drumming on it to create a percussive musical instrument, or decorating it with colored paper to use as a protective helmet in a space game). Performance reflects the child’s ability to inhibit immediate semantic prepotencies and explore distant associative networks.

5.2. Examples (Conceptual Categorization and Instance Generation)

The Examples subscale measures the generative breadth of a child’s category retrieval mechanisms. Students are presented with an abstract or broad descriptive category (e.g., things that make noise, things that are round, or things that are soft) and instructed to name an example that nobody else in their grade would likely think of, while still genuinely satisfying the category constraints. Whereas everyday cognitive tasks prioritize prototypical exemplars (e.g., naming a dog or a cat when asked for animals), creative originality demands searching peripheral conceptual domains. For instance, in response to “things that are cold,” responding “ice cube” represents minimal cognitive search and low originality, whereas responding “the shadow side of the moon” or “an abandoned winter campfire” indicates expansive semantic search, metaphorical bridging, and elite original ideation.

5.3. Sentences (Creative Synthesizing and Contextual Recombination)

The Sentences subscale evaluates original thinking within structured linguistic synthesis. Children are provided with pairs or triads of seemingly unrelated, semantically distant stimulus words (e.g., “cloud” and “shoe”) and asked to construct a coherent, meaningful sentence that links the concepts in an imaginative manner. This dimension challenges the child’s syntactical integration and narrative elaboration. A conventional, low-originality response attempts to resolve the tension through superficial juxtaposing (e.g., “I looked at a cloud while tying my shoe”). In contrast, an original response recombines the conceptual frameworks of the stimuli into an integrated metaphorical or narrative image (e.g., “The giant walked through the sky, leaving muddy shoe prints that floated away as thunder clouds”). The Sentences subscale bridges divergent ideation with integrative cognitive convergence.

6. Theoretical Framework

The conceptual and psychometric architecture of MOTES is grounded in an integration of classical psychometric creativity paradigms, contemporary cognitive associative network theories, and computational psychometric models.

6.1. Guilford’s Structure-of-Intellect and Torrance’s Psychometrics

MOTES traces its lineage to the pioneering psychometric formulations of J.P. Guilford (1950, 1967), who formally distinguished between convergent production (arriving at a single correct solution) and divergent production (generating a multiplicity of possibilities from given information). Guilford posited that divergent production comprises fluency, flexibility, elaboration, and originality. E. Paul Torrance (1966, 1974) operationalized these concepts into standardized testing batteries. However, contemporary psychometrics has revealed that Guilford and Torrance’s classic composite scoring suffers from multicollinearity: fluency often accounts for upwards of 80% of the variance in originality scores due to response-counting artifacts. MOTES directly resolves this theoretical dilemma by decoupling originality from fluency through a constrained “best-idea” or single-response prompt format, ensuring that the latent factor reflects cognitive novelty rather than mere processing rate.

6.2. Associative Theory of Creativity and Modern Executive Search

At the cognitive processing level, MOTES draws heavily upon Sarnoff Mednick’s (1962) Associative Theory of the Creative Process. Mednick hypothesized that individual differences in original thinking are governed by the organization of associative hierarchies in long-term memory. Less original individuals possess steep associative hierarchies, characterized by a rapid exhaustion of stereotypical associations with few idiosyncratic connections. Highly original thinkers possess flatter, broader associative hierarchies, allowing associative activation to spread to semantically distant concepts. Modern neuro-cognitive models (e.g., Beaty et al., 2014) expand this framework by demonstrating that original ideation requires a dynamic interplay between the Default Mode Network (responsible for spontaneous idea generation and episodic retrieval) and the Central Executive Network (responsible for cognitive control, task maintenance, and the inhibition of obvious, prepotent exemplars). In MOTES, tasks explicitly demand executive control to reject first-order associations in pursuit of distant semantic nodes.

6.3. Computational Linguistics and Distributional Semantics

The theoretical underpinning of MOTES scoring is rooted in the Distributional Hypothesis of linguistics (Harris, 1954; Firth, 1957), which asserts that words occurring in similar linguistic contexts share similar semantic meanings. By projecting child-generated textual responses into high-dimensional semantic vector spaces (e.g., using continuous vector space models like Word2Vec, GloVe, or transformer embeddings), the semantic relatedness between the prompt stimulus and the child’s response can be quantified geometrically using cosine similarity. The inverse of cosine similarity represents semantic distance. Computational psychometricians (e.g., Dumas, Organisciak, & Doherty, 2020) have demonstrated that semantic distance serves as a rigorous, objective proxy for original thinking, freeing creativity assessment from the historical subjectivity of human rater consensus.

7. Validity

The construct, convergent, discriminant, and criterion-related validity of the Measure of Original Thinking in Elementary School has been comprehensively evaluated across multiple empirical studies involving elementary student populations in the United States (Acar et al., 2024).

7.1. Construct and Convergent Validity

Construct validity was established by correlating MOTES scores with established, gold-standard measures of creativity. Most notably, MOTES task performance showed statistically significant, robust positive correlations with the Torrance Tests of Creative Thinking–Figural Form A (TTCT; Torrance, 1998b). Specifically, MOTES significantly predicted the Originality criterion scores on the TTCT ($r \approx .35$ to $.45, p < .001$) as well as the Abstractness of Titles scale, which measures conceptual synthesis and verbal metaphorical transformation. This demonstrates that while MOTES is primarily a verbal-computational assessment, it taps into domain-general creative ideational processes that transcend task modality.

Furthermore, convergent validity was substantiated through structural equation modeling examining relationships between latent originality and standardized academic achievement assessed via the Measures of Academic Progress (MAP). MOTES exhibited substantial, statistically significant standardized path coefficients with:

• MAP Reading Achievement: $\beta = .63, p < .001$
• MAP Mathematics Achievement: $\beta = .59, p < .001$

These findings align with cognitive developmental theory, which indicates that creative original thinking requires robust verbal knowledge networks, working memory capacity, and abstract relational reasoning—facets central to reading comprehension and mathematical problem-solving.

7.2. Discriminant Validity

To ensure that MOTES does not merely measure elementary cognitive processing speed, mechanical typing speed, or non-cognitive self-perception, discriminant validity analyses were conducted. MOTES originality scores demonstrated weak and non-significant or trivial correlations with mechanical reading and typing speed ($r < .15$), confirming that the tool measures cognitive divergence rather than physical fluency or reading velocity. Additionally, correlations between MOTES scores and children’s self-rated creative originality (measured via Likert-type creative self-efficacy scales) were modest ($r \approx .18$ to $.22$), reflecting the well-documented psychometric discrepancy between perceived creative competence and objectively verified creative production in pediatric populations.

7.3. Differential Item Functioning (DIF)

To evaluate measurement invariance across diverse demographic subgroups, item response theory (IRT) and logistic regression DIF procedures were executed across biological sex (male vs. female) and racial/ethnic groups. While minor, statistically significant DIF was detected at the isolated item level ($p = .01$), effect sizes were negligible ($R^2 \Delta < .02$), indicating that the overall MOTES composite and subscale scores provide equitable, unbiased assessments of creative potential across diverse student groups.

8. Reliability

The reliability of MOTES has been rigorously established utilizing both modern structural equation modeling reliability coefficients and automated-to-human scoring concordance analyses.

8.1. Internal Consistency and Maximal Reliability

Because traditional Cronbach’s alpha ($lpha$) relies on tau-equivalence assumptions that are routinely violated in multi-faceted creativity tasks, Acar et al. (2024) evaluated internal consistency utilizing Hancock’s $H$ (maximal reliability coefficient) alongside McDonald’s omega ($\omega$):

• Uses Subscale: Hancock’s $H = 0.82$
• Examples Subscale: Hancock’s $H = 0.85$
• Sentences Subscale: Hancock’s $H = 0.88$
• Higher-Order Originality Factor: Hancock’s $H = 0.89$ ($\omega = .87$)

These coefficients significantly surpass the established .80 threshold recommended for individual diagnostic assessments, confirming that the 24-item battery provides a tightly calibrated, psychometrically stable measurement instrument.

8.2. Inter-Rater Reliability and AI-Human Scoring Concordance

The manual evaluation of MOTES items was conducted by multiple independent, trained raters utilizing a standardized 1-to-5 holistic originality scoring rubric. Across human raters, the two-way random intraclass correlation coefficients (ICC) exceeded $.84$, indicating exceptional consensus. Even more critically for large-scale educational psychometrics, automated computational scoring algorithms (leveraging vector semantic spaces) demonstrated exceptional concordance with expert human judgment:

• Uses Subscale AI-Human Agreement: $r = .79$
• Examples Subscale AI-Human Agreement: $r = .91$
• Sentences Subscale AI-Human Agreement: $r = .85$

These findings confirm that automated semantic distance modeling can reliably substitute for manual human ratings, ensuring objective, rapid, and low-cost deployment in institutional environments without sacrificing measurement fidelity.

9. Factor Analysis

The internal structural architecture of MOTES was examined using extensive exploratory and confirmatory factor analytic (CFA) protocols across iterative development cohorts.

9.1. Exploratory Phase and Cognitive Labs

The initial prototyping of MOTES comprised four experimental task types with 10 draft prompts each (40 total items). Following qualitative cognitive laboratories with elementary students (evaluating developmental comprehension, digital interface interaction, and cognitive fatigue) and expert panel validation involving elementary educators, school psychologists, and creativity theorists, 16 psychometrically underperforming or developmentally confusing items were pruned. One task type was eliminated due to excessive collinearity and high task complexity, resulting in the final 24-item, three-subscale structure.

9.2. Confirmatory Factor Analysis (CFA) Models

Acar et al. (2024) tested competing structural models via CFA using robust maximum likelihood estimation (MLR) to account for slight non-normality in semantic distance distributions:

• One-Factor General Model: Posited that all 24 items loaded directly onto a single undifferentiated Originality latent variable. This model demonstrated mediocre fit ($ ext{CFI} < .90, text{RMSEA} > .08$), indicating that task modality variance cannot be ignored.
• Three-Factor Correlated Model: Modeled three distinct, inter-correlated latent factors representing Uses, Examples, and Sentences. This model exhibited an excellent fit to the empirical data ($\chi^2 / df < 2.0, \text{CFI} = .96, \text{TLI} = .95, \text{RMSEA} = .042 [90% \text{ CI: } .036, .048], \text{SRMR} = .038$). The inter-factor correlations were substantial (ranging from $r = .65$ to $.78$), indicating strong shared variance.
• Higher-Order (Hierarchical) Factor Model: Structured the three first-order factors (Uses, Examples, Sentences) under a single, second-order general latent factor of Original Thinking. This model displayed equivalent, exceptional statistical fit ($\text{CFI} = .96, \text{TLI} = .95, \text{RMSEA} = .043, \text{SRMR} = .039$).

The higher-order model was ultimately retained by the authors on both theoretical and pragmatic grounds: it formally recognizes the distinct cognitive operations required by different prompt formats while simultaneously providing empirical justification for calculating a single, unified composite score representing a child’s overall original thinking capacity. All 24 items demonstrated high and statistically significant standardized factor loadings onto their designated first-order factors (ranging from $lambda = .58$ to $.84, p < .001$), with first-order factors exhibiting robust loadings onto the overarching Originality construct ($lambda > .80$).

10. Instrument / Measurement Tool

The Measure of Original Thinking in Elementary School is configured as a standardized, interactive, digital assessment platform. Its formal structural specifications are summarized below:

• Instrument Type: Standardized, Computerized Performance Task / Divergent Ideation Battery.
• Target Population: Elementary School Children (Ages 6 to 12; Kindergarten through Grade 6).
• Total Number of Items: 24 discrete prompts.
• Subscales / Dimensions:
  • Uses Subscale: 8 open-ended prompts assessing alternative functions for everyday physical items.
  • Examples Subscale: 8 open-ended prompts requiring unique exemplar generation for abstract descriptive categories.
  • Sentences Subscale: 8 open-ended prompts demanding the creative linguistic combination of semantically distant noun/concept pairs.
• Administration Method: Electronic delivery via web browser or tablet interface. Tasks are presented within a game-like, visually engaging graphical layout featuring animated instructional tutorials and practice items.
• Response Modality: Single response per prompt. Students type their answers into an interactive text entry box or utilize speech-to-text voice recognition for younger participants (grades K–2).
• Administration Duration: Approximately 25 to 35 minutes for the entire 24-item battery, allowing completion within a standard single elementary school class period.
• Scoring Format: Responses are evaluated on a 5-point ordinal continuum ranging from 1 to 5:
  • 1 = Not Original: Highly conventional, prototypical, obvious, or dictionary-definition response.
  • 2 = Slightly Original: Low novelty; minor variation of a common idea.
  • 3 = Moderately Original: Demonstrates clear ideational departure from the most common responses; interesting connection.
  • 4 = Very Original: Highly unique, unexpected, imaginative, and semantically distant, yet conceptually appropriate.
  • 5 = Exceptionally Original: Rare, brilliant conceptual breakthrough, highly metaphorical, or striking narrative synthesis.
• Scoring Engine: Capable of dual-mode evaluation:
  • Human Scoring: Evaluated by trained educators/psychologists using rubric anchor exemplars.
  • Automated Computational Scoring: Integrated text-processing algorithms evaluate responses via vector-space semantic distance models (GloVe / text embeddings) calibrated against elementary corpus norms.

11. Permissions & Fee and Test Year

The Measure of Original Thinking in Elementary School was officially published in 2024. The empirical development, factor structure, and psychometric validation of the instrument were detailed in the Journal of Educational Psychology (Acar et al., 2024), published by the American Psychological Association (APA). The formal APA PsycTests database entry is indexed under DOI: 10.1037/t91308-000.

The underlying conceptual frameworks, item rubrics, and research methodologies are copyrighted by the original authors and the American Psychological Association. For legitimate non-commercial, academic, and scientific research endeavors, the instrument and scoring algorithms may be accessed by contacting the lead author (Dr. Selcuk Acar, University of North Texas) or via institutional access to the APA PsycTests repository. Commercial use, reproduction in digital learning management systems, or widespread institutional screening deployment requires formal written permission and appropriate licensing agreements from the copyright holders.

12. References

• Acar, S., Dumas, D., Organisciak, P., & Berthiaume, K. (2024). Measuring original thinking in elementary school: Development and validation of a computational psychometric approach. Journal of Educational Psychology, 116(6), 953–981. https://doi.org/10.1037/edu0000844
• Acar, S., & Runco, M. A. (2019). Divergent thinking: New methods, recent research, and extended theory. Psychology of Aesthetics, Creativity, and the Arts, 13(2), 153–158. https://doi.org/10.1037/aca0000231
• Beaty, R. E., Benedek, M., Wilkins, R. W., Jauk, E., Fink, A., Silvia, P. J., Hodges, D. A., Koschutnig, K., & Neubauer, A. C. (2014). Creativity and the default network: A functional connectivity analysis of the creative brain at rest and during divergent thinking. Neuropsychologia, 64, 92–98. https://doi.org/10.1016/j.neuropsychologia.2014.09.019
• Dumas, D., Organisciak, P., & Doherty, M. (2020). Measuring original thinking on open-ended tests: A computational psychometric approach. Creativity Research Journal, 32(1), 39–54. https://doi.org/10.1080/10400419.2020.1712160
• Firth, J. R. (1957). A synopsis of linguistic theory, 1930–1955. Studies in Linguistic Analysis, 1–32.
• Guilford, J. P. (1950). Creativity. American Psychologist, 5(9), 444–454. https://doi.org/10.1037/h0063487
• Guilford, J. P. (1967). The nature of human intelligence. McGraw-Hill.
• Harris, Z. S. (1954). Distributional structure. Word, 10(2–3), 146–162. https://doi.org/10.1080/00437956.1954.11659520
• Mednick, S. (1962). The associative basis of the creative process. Psychological Review, 69(3), 220–232. https://doi.org/10.1037/h0048850
• Runco, M. A., & Acar, S. (2012). Divergent thinking as an indicator of creative potential. Creativity Research Journal, 24(1), 66–75. https://doi.org/10.1080/10400419.2012.652929
• Torrance, E. P. (1966). The Torrance Tests of Creative Thinking: Norms-technical manual. Personnel Press.
• Torrance, E. P. (1974). Torrance Tests of Creative Thinking: Streamlined scoring guide, Figural A and B. Scholastic Testing Service.
• Torrance, E. P. (1998b). Torrance Tests of Creative Thinking: Norms-technical manual: Figural (streamlined) forms A & B. Scholastic Testing Service.

13. Items of the Scale