Three-Minute Test

1. Abstract

The Three-Minute Test (known natively in Dutch as the Drie-Minuten-Toets or DMT) is an individually administered, standardized psychometric instrument designed to assess technical reading fluency and word identification speed in primary school children. Originally developed by Prof. Dr. Ludo Verhoeven and later calibrated and standardized in revised editions by I. Jongen, R. Krom, and the Dutch National Institute for Educational Measurement (Cito), the DMT constitutes an empirical cornerstone of the Dutch National Student and Educational Monitoring System (Leerling- en onderwijsvolgsysteem, or LOVS). The test measures the degree of automaticity attained in single-word decoding through a speeded performance paradigm across three distinct, graded word-reading cards, each presented for an exact duration of 60 seconds (total administration time: 3 minutes).

The three cards represent distinct levels of orthographic, morphological, and phonological complexity: Card 1 evaluates transparent monosyllabic consonant-vowel-consonant (CVC) configurations and basic consonant blends; Card 2 presents more complex monosyllabic words featuring complex consonant clusters, vowel digraphs, and inflectional markers; and Card 3 measures the decoding of polysyllabic, morphologically complex, and phonetically irregular words. The test records the total number of items correctly vocalized per minute, providing individual card scores as well as an integrated composite ability score (vaardigheidsscore). Psychometric evaluations conducted according to the guidelines of the Dutch Committee on Test Affairs (COTAN) consistently demonstrate high internal consistency, parallel-form reliability ($r > .88$), and robust criterion and construct validity. The DMT is calibrated using Item Response Theory (specifically, the One-Parameter Logistic Model / Rasch framework), enabling vertical scaling across elementary grades (grades 1 through 6; Dutch primary groups 3 through 8). As an indispensable diagnostic screening mechanism, it accurately identifies specific reading disabilities, informs Response to Intervention (RTI) trajectories, and differentiates lexical retrieval automaticity deficits associated with developmental dyslexia.

2. Keywords

Three-Minute Test, Drie-Minuten-Toets, DMT, reading fluency, word decoding, technical reading, psychometrics, LOVS, Cito, developmental dyslexia, Rasch modeling, orthographic automaticity, educational assessment, elementary education

3. Authors

The Three-Minute Test was conceptualized, researched, and normed through the collaboration of academic psycholinguists and test developers in the Netherlands:

• Prof. Dr. Ludo Verhoeven: Distinguished Professor Emeritus of Psycholinguistics and Learning Disabilities at the Behavioural Science Institute, Radboud University Nijmegen, the Netherlands. Principal architect of the theoretical underpinnings and empirical foundational designs of the initial DMT batteries.
• I. Jongen, MSc: Psychometrician and assessment developer at Cito (Centraal Instituut voor Toetsontwikkeling / National Institute for Educational Measurement), Arnhem, the Netherlands. Co-author of the 2009 combined normative standardization of the DMT and AVI systems.
• Dr. R. Krom: Senior research psychometrician and statistical methodology specialist at Cito, Arnhem, the Netherlands. Instrumental in the application of Item Response Theory, vertical equating, and normative standardizations within the LOVS framework.

Institutional Contact: Cito B.V., Postbus 1034, 6801 MG Arnhem, The Netherlands. Official website: https://www.cito.nl.

4. Purpose

The primary purpose of the Three-Minute Test is to quantify and monitor the development of technical word-level reading speed, decoding accuracy, and word recognition automaticity among school-aged children. Within both clinical neuropsychology and educational pedagogy, technical reading—defined as the rapid, accurate translation of graphemic representations into phonological units independent of contextual semantic scaffolding—serves as the critical baseline prerequisite for higher-order reading comprehension. When technical decoding requires excessive cognitive energy, working memory resources are consumed, leading to severe comprehension deficits. The DMT was purposefully engineered to isolate word-level decoding efficiency from textual guessing, linguistic context heuristics, and cognitive deduction.

In educational contexts, the DMT functions within a multi-tiered Response to Intervention (RTI) model. Administered biannually or triannually across primary schools as part of the Cito LOVS battery, it serves as a universal Tier-1 screening instrument to evaluate classroom-wide reading progress. For students performing below targeted normative benchmarks (e.g., scoring in the lowest 10th or 25th percentiles), the DMT identifies the precise linguistic boundaries at which decoding automaticity breaks down. This diagnostic function informs targeted Tier-2 pedagogical remediation, enabling educators to adapt phonics and word-study instruction to whether a child struggles with monosyllabic phoneme-grapheme conversions (Card 1), complex consonant clusters (Card 2), or polysyllabic morphological parsing (Card 3).

In clinical and psychological practice, the DMT serves as an essential component in the comprehensive diagnostic assessment of developmental dyslexia and specific learning disorders under DSM-5 and ICD-11 criteria. Dutch diagnostic consensus guidelines mandate objective documentation of persistent reading speed and accuracy deficits resistant to systematic, evidence-based intervention. By pairing the DMT with untimed reading assessments and pseudo-word reading inventories (such as the Klepel), clinicians can differentiate between direct lexical retrieval deficits and sublexical phonological assembly dysfunction, providing clear empirical profiles of reading impairment.

5. Psychological Construct

The psychological construct measured by the Three-Minute Test is technical word reading fluency (technische leesvaardigheid), operationalized as the rate and accuracy of vocalizing isolated printed words under time constraints. This construct comprises three interrelated cognitive-linguistic subcomponents:

1. Sublexical Phonological Decoding Automaticity

In the early phases of reading acquisition, children rely heavily on sequential phonological recoding. This process requires mapping graphemes to their corresponding phonemes, holding those phonological segments in working memory, and blending them to produce the target word. On the DMT, Card 1 tests this capability by presenting monosyllabic orthographically regular words. Efficient performance demands fast phonological assembly without conscious, effortful grapheme-by-grapheme sounding out. Automaticity in sublexical decoding reduces cognitive load, freeing attention for higher-order semantic processing.

2. Orthographic Lexical Retrieval and Sight-Word Recognition

As readers advance, they transition from serial phonological decoding to direct orthographic recognition, mapping complete visual-orthographic word patterns directly to entries in their mental lexicon. This stage, central to Linnea Ehri’s model of reading development, is evaluated primarily through Card 2 and familiar words on Card 3. The construct requires that orthographic representations stored in long-term memory are activated immediately upon visual fixation, bypassing the slower, serial phonological assembly route.

3. Morpho-Phonological Parsing and Polysyllabic Processing

Card 3 evaluates the child’s ability to parse complex multi-morphemic structures, compounding patterns, and phonotactically demanding multi-syllable constructions. This requires rapid identification of morphemic boundaries (roots, prefixes, suffixes) and the execution of complex stress patterns and vowel reductions. Deficits in this dimension reflect difficulties in transitioning from early rudimentary reading to advanced, structurally mature word identification.

Across all three cards, the unifying construct is speed of processing under cognitive linguistic constraints. Unlike untimed diagnostic batteries that evaluate whether a child can eventually decode a word, the DMT assesses cognitive automaticity—measuring how effortlessly and quickly lexical access occurs under strict one-minute limits per linguistic tier.

6. Theoretical Framework

The Three-Minute Test is grounded in well-established neurocognitive and psycholinguistic models of reading acquisition, particularly the Dual-Route Cascaded (DRC) model of reading aloud (Coltheart et al., 2001) and the Automaticity Theory of reading (LaBerge & Samuels, 1974).

The Dual-Route Cascaded Model

The DRC model posits two primary computational pathways for pronouncing written words:

• The Non-Lexical (Sublexical) Route: Visual graphemes are converted into phonemes via rule-based grapheme-to-phoneme conversion (GPC) mechanisms. This pathway is essential for reading novel words, regular words during early acquisition, and pseudo-words.
• The Lexical-Nonsemantic Route: Direct visual-orthographic input accesses the mental orthographic lexicon, which activates corresponding phonological representations in the phonological output lexicon, bypassing semantic processing. This pathway supports the rapid identification of familiar, irregular, and complex regular words.

The DMT is structured to reflect progression along these pathways. Initial cards evaluate whether non-lexical GPC rules have been mastered with sufficient speed, while advanced cards assess the functionality and access speed of the lexical-orthographic retrieval route.

Automaticity and Cognitive Load Theory

According to Automaticity Theory (LaBerge & Samuels, 1974), human cognitive processing operates under finite attentional and working memory capacities. Reading involves multiple hierarchical operations: sensory visual analysis, letter-string identification, phonological recoding, lexical retrieval, syntactic parsing, and semantic integration. If lower-level processes (letter and word identification) are not fully automatized, they consume cognitive capacity, leaving fewer mental resources for comprehension. By measuring word reading within a 60-second limit per card, the DMT directly indexes this automaticity threshold.

Ehri’s Phases of Word Recognition

The structural gradation of the DMT mirrors the phases of word reading development identified by Linnea Ehri: the pre-alphabetic, partial alphabetic, full alphabetic, and consolidated alphabetic phases. Card 1 corresponds closely to the transition from the partial to the full alphabetic phase, where grapheme-phoneme connections are established. Cards 2 and 3 correspond to the consolidated alphabetic phase, where larger multi-letter units, syllables, morphemes, and common orthographic patterns are stored as consolidated chunks, enabling fluent, rapid identification of complex language structures.

7. Validity

The Three-Minute Test has undergone psychometric validation studies conducted by Cito, academic universities, and the Dutch Committee on Test Affairs (COTAN). The instrument shows strong evidence across multiple validity domains:

Construct and Structural Validity

Construct validity has been established by demonstrating that performance on the DMT reflects developmental progressions in reading proficiency across primary school grades. Cross-sectional and longitudinal data from the LOVS standardization cohorts show marked increases in raw word counts and Rasch ability scores from Grade 1 to Grade 4, reaching a typical asymptotic leveling in Grades 5 and 6 as technical decoding approaches full automatization. Confirmatory factor analyses consistently support a single dominant dimension of technical word reading speed that accounts for the majority of common variance across all three cards.

Convergent and Criterion Validity

The convergent validity of the DMT has been evaluated against alternative Dutch standardized reading tests. Strong correlations have been found with:

• The Brus One-Minute Test (Eén-Minuut-Toets / EMT): Pearson product-moment correlation coefficients between the DMT composite score and the EMT range from $r = .82$ to $r = .91$, confirming that both tests capture the same underlying construct of isolated word decoding speed.
• The AVI Reading Assessment System (AVI-toetspakket): Correlations between DMT word scores and continuous text reading fluency on the AVI range from $r = .72$ to $r = .86$. This shows that word-level automaticity accounts for significant variance in running text fluency, while also confirming that continuous text introduces additional factors (e.g., syntactic context, semantic constraints, prosodic organization).
• The Klepel (Pseudo-word Reading Test): Significant positive correlations ($r = .74$ to $r = .83$) confirm that sublexical phonological assembly mechanisms are strongly engaged during DMT performance.

Predictive and Discriminant Validity

Longitudinal studies indicate that DMT performance in Grade 1 (Group 3) significantly predicts reading comprehension scores in later elementary grades (Grades 3 to 6), with regression models showing path coefficients between $\beta = .45$ and $\beta = .62$. Discriminant validity is supported by lower correlations with nonverbal intellectual capacity ($r = .20$ to $r = .35$), confirming that the DMT measures dedicated linguistic and decoding automaticity rather than general cognitive ability.

8. Reliability

The DMT demonstrates high reliability across multiple psychometric indicators:

Parallel-Form Reliability

To support repeated measurements within the LOVS without practice effects, the DMT provides alternate parallel forms (Forms A, B, and C). Across large national calibration samples of primary school children ($N > 2,500$ per grade level), alternate-form correlation coefficients consistently exceed acceptable psychometric thresholds:

• Card 1 Parallel-Form Reliability: $r = .89$ to $.93$
• Card 2 Parallel-Form Reliability: $r = .88$ to $.92$
• Card 3 Parallel-Form Reliability: $r = .90$ to $.94$
• Composite Score Parallel-Form Reliability: $r = .93$ to $.96$

Test-Retest Stability

Short-term test-retest reliability across a two-to-three-week interval ranges from $r = .87$ to $r = .93$, showing that individual score rankings remain stable across testing sessions while capturing expected developmental growth over longer instructional intervals.

Standard Error of Measurement and COTAN Evaluations

The Standard Error of Measurement (SEM) remains small relative to the variance of the raw scores across all grades. In independent reviews by the Dutch Committee on Test Affairs (COTAN), the DMT has consistently received positive ratings for reliability, internal consistency, the quality of its test materials, and the adequacy of its normative data sets.

9. Factor Analysis and Item Response Theory Calibration

The calibration and structural integrity of the Three-Minute Test rely on Item Response Theory (IRT), specifically the One-Parameter Logistic Model (OPLM) and the Rasch modeling framework utilized by Cito for the LOVS battery.

Unidimensionality and Factor Structure

Both exploratory (EFA) and confirmatory factor analyses (CFA) demonstrate that performance across Cards 1, 2, and 3 is explained by a single underlying latent trait: technical decoding speed ($ heta$). In structural equation models, single-factor solutions yield fit indices that meet standard criteria:

• Comparative Fit Index (CFI) $ge .98$
• Tucker-Lewis Index (TLI) $ge .97$
• Root Mean Square Error of Approximation (RMSEA) $le .045$

Card-to-factor standardized loadings systematically exceed $.82$, confirming that the three cards assess different difficulty tiers along the same latent cognitive dimension rather than isolated, distinct constructs.

Item Response Theory and Vertical Scaling

Using the One-Parameter Logistic Model (OPLM), item difficulties were estimated and aligned onto a single vertical scale (vaardigheidsschaal). In this IRT model, the discrimination parameters ($a_i$) are fixed at integer values, while item difficulty parameters ($b_i$) vary along a shared continuous logit scale. This approach provides several key measurement benefits:

• Cross-Grade Comparability: A child’s performance can be followed longitudinally on a continuous scale from Group 3 (Grade 1) through Group 8 (Grade 6), allowing direct measurement of growth over time.
• Card Equating: Differences in difficulty across Cards 1, 2, and 3, as well as between parallel versions (Forms A, B, and C), are modeled mathematically. Raw words-correct scores are transformed into unified ability scores (vaardigheidsscores) that adjust for card difficulty.
• Infit and Outfit Statistics: Rasch fit statistics (mean square infit between $0.85$ and $1.15$) show that individual items conform well to the modeled latent reading continuum, confirming measurement validity without significant multidimensional distortion.

10. Instrument / Measurement Tool

The Three-Minute Test is an individually administered diagnostic assessment designed for standardized educational and clinical use. Below is a structural breakdown of the test parameters, materials, and scoring rules:

• Assessment Type: Standardized, individual, criterion- and norm-referenced speed test of technical word reading.
• Target Population: Elementary school children from Grade 1 through Grade 6 (Dutch primary school groups 3 to 8; ages 6 to 12).
• Administration Format: Individual administration (examiner and student seated face-to-face in a quiet testing environment).
• Test Materials:
  • Three laminated student test cards (Card 1, Card 2, Card 3).
  • Parallel test card series (Versions A, B, and C) to prevent re-test familiarity.
  • Standardized examiner score recording sheet containing identical text columns with word-tracking indices.
  • Precision stopwatch or digital countdown timer.
  • Scoring manual with conversion tables for raw-to-ability scores and normative tables.
• Card Specifications and Item Layout:
  • Card 1: 150 regular monosyllabic words arranged in multiple columns (predominantly simple CVC, CCV, VCC, and simple CVCC structures).
  • Card 2: 150 complex monosyllabic words arranged in multiple columns (words with initial and final consonant clusters, digraphs, and complex orthographic patterns).
  • Card 3: 120 to 150 polysyllabic words (compound words, multi-morphemic structures, prefixed and suffixed words).
• Time Constraints: Exactly 60 seconds (1 minute) of reading time per card. Total testing time: 3 minutes (excluding transition instructions).
• Standardized Administration Protocol:
  • The examiner positions the specified card face down in front of the student.
  • Standardized instructions are delivered: the student is instructed to read the words aloud down each column as accurately and quickly as possible without skipping.
  • The card is turned over, and timing begins immediately as the student vocalizes the first word.
  • The examiner tracks the student’s reading on the recording sheet, marking errors unobtrusively.
  • At exactly 60 seconds, the examiner says “Stop” and notes the final word attempted.
  • The procedure is repeated for Cards 2 and 3 according to the grade-level protocol.
• Scoring Rules:
  • Raw Score per Card: Total number of words read correctly within 60 seconds (Total words attempted minus errors).
  • Counted as Errors: Mispronunciations, phoneme omissions, phoneme additions, word substitutions, and word omissions.
  • Self-Corrections: If the student spontaneously corrects an error before moving to the next word, the word is scored as correct.
  • Composite Score: Sum of raw correct words across applicable cards, or conversion to a standardized ability score (vaardigheidsscore) via Cito scoring software.
  • Normative Classifications: Ability scores are mapped to LOVS performance bands (A to E, where A represents the top 25% and E represents the lowest 10%; or quintile rankings I to V).

11. Permissions, Fee, and Test Year

The Three-Minute Test is a proprietary, copyrighted instrument owned and published by Cito (Centraal Instituut voor Toetsontwikkeling / National Institute for Educational Measurement), located in Arnhem, the Netherlands.

• Year of Initial Publication: The original DMT was developed by Prof. Dr. Ludo Verhoeven in the late 1980s and early 1990s.
• Major Standardized Revisions: A major standardized edition combining the DMT with the AVI system was released in 2009 by I. Jongen and R. Krom, with updated digital scoring adaptations introduced through the Cito Volgsysteem primair onderwijs (Leerlingvolgsysteem) in 2017 and subsequent years.
• Commercial Licensing and Fees: The test materials (physical test cards, scoring protocols, diagnostic administration manuals, and digital LOVS reporting licenses) are commercially distributed. Primary schools and licensed psychological clinics purchase test packages directly from Cito. Educational subscriptions are typically billed on an annual institutional basis per student cohort.
• Access Restrictions for Research and Clinical Practice: Qualified educational specialists, remedial teachers (remediële teachers), school psychologists, and clinical child neuropsychologists are authorized to purchase and administer the DMT. Academic researchers seeking to use the DMT in empirical studies must obtain written permission and license agreements directly from Cito B.V. The test cards and word lists are protected under international copyright law and may not be reproduced in research publications or distributed in the public domain.

12. References

Below is a list of foundational academic publications, psychometric standardization reports, and theoretical literature relevant to the Three-Minute Test:

• Coltheart, M., Rastle, K., Perry, C., Langdon, R., & Ziegler, J. (2001). DRC: A dual route cascaded model of visual word recognition and reading aloud. Psychological Review, 108(1), 204–256. https://doi.org/10.1037/0033-295X.108.1.204
• Commissie Testaangelegenheden Nederland (COTAN). (2010). Beoordeling Drie-Minuten-Toets (DMT) 2009. Nederlands Instituut van Psychologen (NIP). https://www.cotanbeoordelingen.nl
• Ehri, L. C. (2005). Learning to read words: Theory, findings, and issues. Scientific Studies of Reading, 9(2), 167–188. https://doi.org/10.1207/s1532799xssr0902_4
• Jongen, I., & Krom, R. (2009). Drie-Minuten-Toets en AVI Toetspakket: Handleiding voor de leerkracht. Cito.
• LaBerge, D., & Samuels, S. J. (1974). Toward a theory of automatic information processing in reading. Cognitive Psychology, 6(2), 293–323. https://doi.org/10.1016/0010-0285(74)90015-2
• Share, D. L. (1995). Phonological recoding and self-teaching: Phonological recoding and self-teaching: Sine qua non of reading acquisition. Cognition, 55(2), 151–218. https://doi.org/10.1016/0010-0277(94)00645-2
• Verhoeven, L. (1993). Drie-Minuten-Toets: Handleiding. Cito.
• Verhoeven, L., & van Leeuwe, J. (2009). Modeling the development of word reading fluency in a transparent orthography. Scientific Studies of Reading, 13(5), 438–456. https://doi.org/10.1080/10888430903162838
• Verhoeven, L., & van Leeuwe, J. (2012). The simple view of reading in a transparent orthography: A longitudinal perspective. Reading and Writing, 25(8), 1815–1835. https://doi.org/10.1007/s11145-011-9335-5

13. Items of the Scale