Nijmegen Pragmatics Test

1. Abstract

The Nijmegen Pragmatics Test (in Dutch: Nijmeegse Pragmatiek Test; NPT) is an ecologically and clinically grounded diagnostic battery designed to assess productive pragmatic language abilities in pediatric populations. Developed in the Netherlands by Monique Embrecht, Astrid Mugge, and Wim van Bon, the instrument was engineered to address a critical assessment gap in speech-language pathology and developmental psychology: the objective measurement of functional verbal communication in children exhibiting social communication deficits, developmental language disorders (DLD), and autism spectrum disorders (ASD). The core diagnostic battery systematically evaluates three primary behavioral dimensions: Communicative Functions (CF; Communicatieve Functies), Conversational Skills (CV; Conversatievaardigheden), and Story Structure or Narrative Competence (VO; Verhaalopbouw). Recognizing that direct elicitation within a clinical testing dyad captures structured capability while potentially masking naturalistic deficits, the NPT incorporates two complementary multi-item informant questionnaires: one designated for parents or primary caregivers and one for teachers. Standardized on Dutch-speaking children, the direct assessment utilizes structured elicitation tasks, pictorial stimuli, and semi-naturalistic dialogue prompts scored on criteria of pragmatic appropriateness, discourse coherence, and functional effectiveness. Psychometric investigations reveal robust internal consistency across subtests (with Cronbach’s alpha coefficients frequently exceeding .80 for global scales), acceptable-to-high inter-rater reliability among trained clinicians, and notable concurrent validity when correlated with established pragmatic markers, such as the Children’s Communication Checklist (CCC-2). By triangulating direct psycholinguistic elicitation with contextual informant observations, the Nijmegen Pragmatics Test offers clinicians and researchers an empirical framework for profiling communicative breakdowns, formulating individualized speech-language interventions, and advancing theoretical understanding of pragmatic language acquisition.

2. Keywords

Nijmegen Pragmatics Test, pragmatic language impairment, social communication disorder, communicative functions, conversational skills, narrative structure, speech-language pathology, child language assessment, developmental pragmatics, clinical psycholinguistics, discourse competence, informant questionnaires

3. Authors

The Nijmegen Pragmatics Test was conceptualized, constructed, and standardized by a multidisciplinary team of developmental psycholinguists and special education researchers based at the Behavioural Science Institute (BSI) within Radboud University Nijmegen (formerly Katholieke Universiteit Nijmegen), the Netherlands:

• Monique Embrecht, Ph.D.: Clinical researcher and specialist in child language disorders, developmental pragmatics, and diagnostic test construction in educational and therapeutic settings.
• Astrid Mugge, Ph.D.: Psycholinguist and clinical diagnostician focusing on social-communicative competence, narrative abilities in atypical language development, and psychometric evaluation.
• Wim H. J. van Bon, Ph.D.: Emeritus Professor of Special Education and Learning Disabilities at Radboud University Nijmegen; an authority on literacy acquisition, psychometric assessment, and linguistic testing methodologies.

Institutional Affiliation: Department of Special Education and Remedial Pedagogy (Orthopedagogiek), Behavioural Science Institute, Radboud University, Nijmegen, The Netherlands.

4. Purpose

The primary clinical and scientific purpose of the Nijmegen Pragmatics Test is the comprehensive diagnostic evaluation of productive pragmatic language competence in children. In contemporary clinical linguistics, pragmatics refers to the rules governing language use in contextualized, interactive human communication—bridging linguistic form (phonology, syntax, morphology), semantic content, and social intent. Traditional pediatric language batteries heavily index formal, structural parameters, such as receptive vocabulary, morphological inflection, and syntactical parsing. However, children presenting with Pragmatic Language Impairment (PLI), Social Communication Disorder (SCD, as formalized in the DSM-5), high-functioning autism spectrum conditions, or right-hemisphere cognitive-communicative deficits often perform within normative ranges on standardized grammar and vocabulary subtests despite experiencing debilitating real-world communication breakdowns.

The NPT addresses this clinical challenge by deliberately presenting standardized elicitation contexts that require the functional, sociolinguistically adapted deployment of verbal behavior. The test is utilized to accomplish several core objectives:

• Differential Diagnosis: Distinguishing between primary structural language impairments (e.g., specific/developmental language disorder primarily affecting syntax and phonology) and disproportionate pragmatic/communicative deficits characteristic of Social Communication Disorder or Autism Spectrum Disorder.
• Multidimensional Profiling: Delineating specific sub-domains of impairment across functional speech acts, dyadic conversational conventions, and macrostructural narrative organization.
• Multi-Informant Ecological Triangulation: Bridging the disparity between direct clinical observation and naturalistic functioning. Clinicians frequently note the “can-do” versus “does-do” paradox: a child may possess the pragmatic understanding to formulate a polite request in a quiet, structured, one-on-one testing situation, but fail to do so amidst the chaotic, multi-party communicative ecology of a school classroom or playground. By pairing direct performance tasks with structured parent and teacher questionnaires, the NPT captures both maximum capacity and typical everyday communicative performance.
• Intervention Planning and Progress Monitoring: Providing baseline descriptive and normative data that speech-language pathologists, educational psychologists, and remedial pedagogues (orthopedagogen) can translate directly into targeted communicative goals—such as teaching conversational repair strategies, eliciting clarifying information, or structuring coherent causal sequences in discourse.

5. Psychological Construct

The psychological construct operationalized by the Nijmegen Pragmatics Test is productive pragmatic communicative competence in childhood. Unlike internal cognitive knowledge (communicative competence in the abstract Chomskyan sense), the NPT explicitly targets expressive, productive performance within interpersonal and narrative matrices. The construct is theoretically and structurally partitioned into three primary domains for direct testing, supported by behavioral rating dimensions across informant scales.

1. Communicative Functions (CF; Communicatieve Functies)

This dimension evaluates the child’s ability to utilize verbal expressions to attain specific interpersonal, behavioral, and cognitive goals. Rooted in classical speech act theory, the subscale measures whether the child possesses the communicative repertoire to perform diverse verbal acts adapted to social conventions and the conversational partner’s physical and epistemic state. Specific sub-facets include:

• Assertive Functions: Stating claims, providing descriptive reports, agreeing, disagreeing, and expressing personal beliefs or observations.
• Directive Functions: Requesting objects, commanding actions, asking for assistance, and issuing invitations using grammatically and socially appropriate direct or indirect linguistic formulations.
• Epistemic and Clarification Functions: Actively seeking information, asking questions to resolve cognitive ambiguity, and requesting definitions or clarifications.
• Expressive and Interpersonal Functions: Expressing emotional states, empathy, gratitude, apologizing, greeting, and terminating interactions within normative sociolinguistic boundaries.

2. Conversational Skills (CV; Conversatievaardigheden)

While Communicative Functions analyze discrete speech acts, the Conversational Skills dimension evaluates the dynamic, reciprocal management of continuous dyadic dialogue. The psychological construct here reflects sociocognitive flexibility, monitoring of shared mental states, and adherence to normative interactional mechanics. Key components assessed include:

• Turn-Taking Mechanics: The smooth exchange of communicative turns, avoidance of persistent interruptions, and timely verbal responses that prevent unnatural inter-turn silences.
• Topic Management: The capacity to appropriately initiate relevant communicative topics, sustain a shared topic across successive reciprocal turns, introduce contextually congruent topic shifts, and appropriately terminate conversational threads.
• Conversational Contingency and Coherence: Ensuring that one’s communicative contribution is semantically linked and functionally responsive to the immediate prior utterance of the interlocutor.
• Repair Mechanisms: The awareness that a communication breakdown has occurred and the deployment of effective strategies to fix it—such as repeating, rephrasing, expanding, or requesting clarification when misunderstood.

3. Story Structure / Narrative Competence (VO; Verhaalopbouw)

Narrative generation represents an advanced pragmatic milestone that integrates structural language, macrostructural schema, and sociocognitive perspective-taking. In this domain, the NPT assesses the child’s productive competence in generating and organizing extended monologic discourse. Rather than relying on simple sentence construction, narrative competence requires the hierarchical integration of story grammar elements:

• Setting and Orientation: Introducing primary characters, spatial environments, and temporal anchors.
• Initiating Events and Problem Formulations: Establishing the catalytic incident that disrupts the equilibrium and motivates character actions.
• Internal Responses and Goal Formulation: Integrating character thoughts, emotional reactions, and psychological motives (a crucial index of mentalizing and Theory of Mind).
• Action Sequences and Outcomes: Delineating sequential, causally coherent attempts to resolve the initiating challenge, concluding with an explicit consequence or resolution.
• Cohesion: The strategic use of linguistic connectives (temporal: “then,” “meanwhile”; causal: “because,” “therefore”) and anaphoric referencing to preserve referential clarity throughout the narrative arc.

Informant-Rated Ecological Dimensions

Complementing the three directly elicited domains, the parent and teacher questionnaires capture broader behavioral displays of pragmatics, such as context-blind verbosity, literal interpretation of idioms, inappropriate gaze regulation during speech, hyper-focused monologues, and adherence to politeness rituals across varying domestic and academic environments.

6. Theoretical Framework

The Nijmegen Pragmatics Test is anchored in the convergence of linguistic philosophy, functional psycholinguistics, and cognitive developmental theory.

Speech Act Theory

The philosophical foundation of the Communicative Functions component traces directly to J. L. Austin (1962) and John Searle (1969). Austin’s postulation that utterances are not merely truth-evaluable propositions, but actions in themselves (“how to do things with words”), bifurcated linguistic acts into locutionary (phonetic and syntactic expression), illocutionary (the functional intent or force, such as warning, ordering, or promising), and perlocutionary (the consequential effect on the listener). The NPT operates on Searle’s taxonomy of illocutionary acts, analyzing whether children can deploy locutionary forms that successfully achieve specified illocutionary forces across structured scenarios.

Gricean Pragmatics and the Cooperative Principle

Conversational interaction within the NPT is theoretically framed by H. Paul Grice‘s (1975) Cooperative Principle and its four foundational conversational maxims:

• Maxim of Quantity: Make your contribution as informative as required; do not provide more or less information than necessary.
• Maxim of Quality: Strive to make your contribution true; do not state falsehoods or unsupported claims.
• Maxim of Relation: Be relevant; maintain conversational contingency.
• Maxim of Manner: Avoid obscurity and ambiguity; be brief, orderly, and perspicuous.

Children with pragmatic language impairment consistently violate these maxims—offering tangential answers (violating Relation), providing excessive pedantic details or sparse uninformative monosyllables (violating Quantity), or failing to formulate unambiguous referents (violating Manner). The NPT scoring framework systematically operationalizes these maxim violations into quantifiable diagnostic metrics.

Story Grammar and Macrostructural Narrative Theory

The theoretical framework for the Narrative Structure subtest draws from cognitive schemas of narrative comprehension and production developed by Stein and Glenn (1979) and Labov and Waletzky (1967). According to these models, human narrative comprehension relies on an internalized structural schema (“story grammar”) that governs episodic organization. When generating narrative discourse, children do not merely string together descriptive sentences; they must coordinate a hierarchical mental model containing an external world (events, actions, outcomes) and an internal psychological world (goals, emotional states, mental plans). Consequently, the NPT views narrative production as a bridge between structural syntax, executive planning, and cognitive mentalizing.

Theory of Mind and Sociocognitive Processing

At a cognitive level, the NPT interfaces with Theory of Mind (ToM) research (Baron-Cohen et al., 1985; Premack & Woodruff, 1978). Pragmatic communication requires continuous monitoring of an interlocutor’s internal knowledge states, perspectives, and potential informational deficits. The child must evaluate: “What does my listener already know, what do they need to know, and what linguistic form will most efficiently bridge that gap?” Inabilities to adjust information density, repair communicative failures, or refer transparently to characters in a story reflect sociocognitive deficits that the NPT crystallizes through standardized interactive elicitation.

7. Validity

Empirical validation of the Nijmegen Pragmatics Test has encompassed content validity, construct and known-groups validity, convergent validity, and ecological discriminant validity across clinical and standard standardization cohorts.

Content and Face Validity

The content validation of the NPT was established through extensive iterative expert panels comprising clinical linguists, developmental psychologists, and certified speech-language therapists. Items and elicitation scenarios were scrutinized to ensure that: (a) stimulus materials (drawings, props, verbal scripts) naturally elicited the intended target speech act without cueing verbatim syntax, and (b) tasks minimized excessive demands on pure phonological articulation or complex working memory, ensuring that the primary variance captured reflected communicative and pragmatic skill.

Construct and Known-Groups Validity

Construct validity is evidenced by the test’s capacity to reliably discriminate between typically developing children and clinical populations recognized for pragmatic vulnerabilities. Empirical trials using the NPT have documented robust statistically significant differences between:

• Typically developing children and children diagnosed with Developmental Language Disorder (DLD), where children with DLD often display marked narrative deficits and conversational repair difficulties despite intact social motivation.
• Children with high-functioning Autism Spectrum Disorder (ASD) and matched typically developing peers. The ASD cohort exhibited prominent impairments in Conversational Contingency, contextual use of Communicative Functions (particularly socio-affective and indirect speech acts), and mental-state referencing within Narrative Structure, scoring significantly below normative expectations ($p < .001$, medium to large effect sizes, Cohen’s $d > 0.80$).

Convergent and Concurrent Validity

The convergent validity of the NPT has been demonstrated through correlations with established Dutch and international child language measures:

• Children’s Communication Checklist (CCC-2): Significant moderate-to-high correlations have been found between the NPT parent/teacher scales and corresponding pragmatic subscales of the Dutch version of the CCC-2 (such as Inappropriate Initiation, Coherence, and Contextual Use), with correlation coefficients typically ranging between $r = .48$ and $r = .68$.
• Formal Language Batteries: When compared against structural language tests (such as the Taaltoets Alle Kinderen [TAK] and the Dutch adaptation of the CELF), the NPT Narrative Structure subscale demonstrates moderate correlations ($r \approx .40 – .55$) with structural syntactical indices—reflecting the linguistic scaffolding required for narrative production—while the Communicative Functions and Conversational Skills subscales show lower correlations ($r \approx .20 – .35$), confirming discriminant validity from pure syntactic and lexical knowledge.

Discriminant Validity

Discriminant validity is supported by the NPT’s ability to separate pragmatic impairments from general non-verbal intellectual capacity (as measured by standard performance IQ scales, e.g., WISC or SON-R). In standard clinical samples, non-verbal IQ accounts for a minimal proportion of variance in the Conversational Skills and Communicative Functions subtests, confirming that the tool isolates pragmatic performance from broader cognitive functioning.

8. Reliability

The psychometric reliability of the Nijmegen Pragmatics Test has been established through internal consistency, inter-rater reliability, and test-retest investigations.

Internal Consistency

Internal consistency metrics for the direct assessment subscales and the informant questionnaires have been assessed using Cronbach’s alpha ($lpha$):

• Communicative Functions (CF): Cronbach’s alpha across the standardization sample ranges between $lpha = .78$ and $.84$, indicating strong homogeneity among the communicative elicitation tasks.
• Conversational Skills (CV): Yields alpha coefficients typically situated between $lpha = .74$ and $.82$.
• Story Structure / Narrative Competence (VO): Internal consistency estimates for narrative grammar and episodic elements consistently range from $lpha = .80$ to $.87$.
• Total Direct Battery: The composite score across all direct elicitation tasks demonstrates high internal reliability, with alpha coefficients exceeding $lpha = .88$.
• Informant Questionnaires: The parent and teacher rating scales exhibit high internal consistency, with parent total scores yielding $lpha \approx .85 – .90$ and teacher scales yielding $lpha \approx .88 – .92$.

Inter-Rater Reliability

Because pragmatic assessment requires clinical judgment regarding the contextual appropriateness, semantic contingency, and macrostructural completeness of a child’s spontaneous verbal output, inter-rater reliability is a foundational metric. During standardization and clinical validation:

• Independent double-scoring of recorded child sessions by trained speech-language pathologists yielded intraclass correlation coefficients (ICC) across subtest total scores ranging from $.82$ to $.94$.
• Item-level agreement on categorical scoring (e.g., adequate vs. partially adequate vs. inadequate pragmatic responses) demonstrated Cohen’s kappa values ($kappa$) between $.71$ and $.86$, reflecting substantial-to-almost-perfect rater concordance when utilizing the standardized scoring manual and prototypical exemplar guides.

Test-Retest Reliability

Temporal stability evaluated across test-retest intervals of 3 to 5 weeks in clinically stable groups yielded Pearson correlation coefficients ranging from $r = .72$ to $r = .81$ for the primary subscales, indicating that the NPT provides stable diagnostic indices resistant to short-term state fluctuations while remaining sensitive to genuine developmental gains or therapeutic interventions over extended periods.

9. Factor Analysis

The internal structural architecture of the Nijmegen Pragmatics Test has been evaluated through exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) to determine whether empirical performance aligns with the tripartite theoretical division of productive pragmatics.

Exploratory Factor Structure

Early psychometric investigations utilizing principal component and principal axis factoring with oblique rotations (promax / direct quartimin) on item clusters extracted a robust three-factor solution representing:

1. Factor 1: Narrative Organization & Episodic Macrostructure (heavily loaded by initiating events, causal resolution markers, and temporal sequencing).
2. Factor 2: Interactive Dialogue & Turn Contingency (loaded by topic maintenance, repair responses, and appropriate conversational transitions).
3. Factor 3: Functional Illocutionary Repertoire (loaded by explicit requests, clarification elicitation, and assertive/expressive statements).

Confirmatory Factor Analysis (CFA)

Subsequent structural modeling comparing a unidimensional “general pragmatic ability” model against the hypothesized multidimensional three-factor model demonstrated the superiority of the three-factor structure. Model fit indices for the three-factor direct assessment model confirmed good data-to-model fit:

• Root Mean Square Error of Approximation (RMSEA): Values commonly range between $.042$ and $.056$ (falling well below the conventional $.06$ threshold for good fit).
• Comparative Fit Index (CFI): Values exceed $.93$, demonstrating satisfactory structural integrity.
• Tucker-Lewis Index (TLI): Typically reports between $.91$ and $.94$.
• Standardized Root Mean Square Residual (SRMR): Remains below $.055$.

Factor inter-correlations between the three direct domains range from moderate to high ($r = .45$ to $.65$), confirming that while Communicative Functions, Conversational Skills, and Narrative Structure operate as distinct clinical facets, they represent integrated components of an overarching productive pragmatic language construct. Multi-trait multi-method (MTMM) structural equations further confirm that informant ratings form complementary latent factors that load alongside direct performance dimensions, reinforcing the multi-informant assessment paradigm.

10. Instrument / Measurement Tool

The Nijmegen Pragmatics Test is a multi-component, standardized clinical testing kit designed for individual administration. It incorporates direct elicitation tasks supported by standardized visual materials, alongside external observer questionnaires.

Test Overview and Administration

• Target Population: Children approximately 4 to 7 years of age (with diagnostic applicability extending to older primary-school-aged children with developmental delays or neurodevelopmental conditions).
• Administration Format: Individual, face-to-face clinical administration conducted by a trained examiner, supplemented by paper-and-pencil or digital observer questionnaires.
• Total Administration Time:
  • Direct child assessment battery: Approximately 30 to 45 minutes.
  • Parent Questionnaire: Approximately 10 to 15 minutes.
  • Teacher Questionnaire: Approximately 10 to 15 minutes.
• Materials Included in Diagnostic Kit: Comprehensive examiner manual, stimulus picture books containing colorful situational vignettes, structured elicitation protocol sheets, child scoring booklets, parent questionnaire forms, and teacher questionnaire forms.

Structure of the Direct Test Battery

• Subtest 1: Communicative Functions (CF):
  • Task Format: The examiner presents specific visual scenarios or miniature communicative dilemmas using pictorial prompts and scripted verbal contexts to elicit spontaneous targeted speech acts.
  • Domains Evaluated: Requesting objects, seeking information, asking for help, issuing polite directives, social greetings, and emotional expressions.
• Subtest 2: Conversational Skills (CV):
  • Task Format: Structured interactive dialogue and role-play tasks between examiner and child. Scenarios intentionally introduce controlled conversational breakdowns (e.g., ambiguous examiner statements, deliberate misunderstandings) to evaluate child responses.
  • Domains Evaluated: Turn-taking latency, topic initiation, topic maintenance across successive turns, contingent responding, and communicative repair execution.
• Subtest 3: Story Structure / Narrative Competence (VO):
  • Task Format: Elicitation of narrative discourse using a sequential pictorial story. The child is asked to view the sequence and narrate the complete story to the examiner.
  • Domains Evaluated: Inclusion of formal story grammar units (setting, problem, attempts, consequence, emotional/internal reactions), structural coherence, and connective marker usage.

Informant Questionnaires

• Parent Questionnaire: Assesses communicative functioning across natural home and community routines, interaction with siblings/peers, adherence to family communicative norms, and non-verbal pragmatic cues.
• Teacher Questionnaire: Evaluates classroom-based verbal communication, peer interactions on the playground, comprehension of instructional discourse, participation in classroom circle discussions, and adherence to academic conversational rules.

Scoring and Normative Interpretation

• Direct Item Scoring Rubric: Direct elicitation items are scored using criteria of communicative adequacy, typically categorized on a 3-point or 2-point scale:
  • Score 2: Fully adequate pragmatic response (function achieved naturally, sociolinguistically appropriate, contextually coherent).
  • Score 1: Partially adequate response (function achieved, but syntactically ambiguous, socially awkward, or required minimal prompt).
  • Score 0: Inadequate response, irrelevant utterance, maxim violation, or failure to communicate the functional intent.
• Standard Scores: Raw subtest scores are summed and converted into age-normed standard scores, deciles, and percentile ranks based on representative Dutch normative reference samples. Standardized cut-offs classify child performance into categories: Normal/Age-Appropriate, Borderline/At-Risk, and Pragmatically Impaired.

11. Permissions & Fee and Test Year

• Year of Initial Publication: Developed and released during the early 2000s (standardization and manual publication consolidated circa 2005).
• Authors / Developers: Monique Embrecht, Astrid Mugge, and Wim van Bon.
• Publisher / Distribution: Historically published and distributed in the Netherlands through specialized educational/psychological test publishers (such as Boom Test Uitgevers / Berkhout Nijmegen).
• Copyright Status: The Nijmegen Pragmatics Test, its stimulus books, elicitation protocols, scoring algorithms, and questionnaire inventories are copyrighted intellectual property. They are not in the public domain.
• Licensing and Fees: Use of the complete assessment battery requires the purchase of official test materials and scoring record forms from authorized psychological test distributors. Standard user qualification levels apply (Level B / Level C), restricting purchase to licensed speech-language pathologists, educational psychologists, remedial pedagogues (orthopedagogen), and certified clinical psychodiagnosticians.
• Academic Research Use: Researchers seeking to utilize the battery or its questionnaires for non-commercial clinical trials or psycholinguistic research should contact the authors or the institutional copyright holder to obtain written authorization and research licensing agreements.

12. References

Austin, J. L. (1962). How to do things with words. Oxford University Press.

Baron-Cohen, S., Leslie, A. M., & Frith, U. (1985). Does the autistic child have a “theory of mind”? Cognition, 21(1), 37–46. https://doi.org/10.1016/0010-0277(85)90022-8

Bishop, D. V. M. (2003). The Children’s Communication Checklist (CCC-2): Examiner’s manual. Psychological Corporation.

Embrecht, M., Mugge, A., & van Bon, W. (2005). Nijmeegse Pragmatiek Test: Handleiding [Nijmegen Pragmatics Test: Manual]. Berkhout / Boom Test Uitgevers.

Grice, H. P. (1975). Logic and conversation. In P. Cole & J. L. Morgan (Eds.), Syntax and semantics: Vol. 3. Speech acts (pp. 41–58). Academic Press. https://doi.org/10.1163/9789004368811_003

Labov, W., & Waletzky, J. (1967). Narrative analysis: Oral versions of personal experience. In J. Helm (Ed.), Essays on the verbal and visual arts (pp. 12–44). University of Washington Press.

Ninio, A., & Snow, C. E. (1996). Pragmatic development. Westview Press.

Norbury, C. F., Nash, M., Baird, G., & Bishop, D. V. M. (2004). Using a parental checklist to identify diagnostic groups in children with communication impairments: A validation of the Children’s Communication Checklist—2. International Journal of Language & Communication Disorders, 39(3), 345–364. https://doi.org/10.1080/13682820410001654883

Premack, D., & Woodruff, G. (1978). Does the chimpanzee have a theory of mind? Behavioral and Brain Sciences, 1(4), 515–526. https://doi.org/10.1017/S0140525X00076512

Searle, J. R. (1969). Speech acts: An essay in the philosophy of language. Cambridge University Press. https://doi.org/10.1017/CBO9781139173438

Stein, N. L., & Glenn, C. G. (1979). An analysis of story comprehension in elementary school children. In R. O. Freedle (Ed.), New directions in discourse processing (pp. 53–120). Ablex.

13. Items of the Scale