Matching Familiar Figures Test (MFF, MFFT)

1. Abstract

The Matching Familiar Figures Test (MFF, MFFT) was developed by Jerome Kagan and his associates in 1964 at the Fels Research Institute to evaluate the bipolar cognitive style construct known as reflection-impulsivity. Operating as a performance-based visual discrimination and match-to-sample instrument, the MFFT presents an examinee with a standard stimulus figure and an array of highly similar comparison variants, exactly one of which matches the standard identically while the remaining distractors deviate by subtle, isolated graphic details. The original childhood test battery comprises two introductory practice trials followed by 12 standardized experimental test items administered via printed stimulus booklets or visual display cards.

The instrument generates two primary, interdependent behavioral metrics: the latency to the first response (measured in seconds with a stopwatch) and the total number of errors committed across trials before identifying the correct match. Traditionally categorized via a double median split methodology, test-takers are classified into four distinct cognitive processing quadrants: reflective (long latency, low errors), impulsive (short latency, high errors), fast-accurate (short latency, low errors), and slow-inaccurate (long latency, high errors). Psychometric investigations have demonstrated high internal consistency, with split-half reliability coefficients reaching .94 for response latency across large normative samples of children, alongside moderate test-retest stability across longitudinal intervals ranging from 6 to 12 months.

The MFFT has served as a foundational instrument across developmental psychology, clinical neuropsychology, and educational research, yielding vital empirical insights into executive functioning, visual search efficiency, attention-deficit/hyperactivity disorder (ADHD), and reading acquisition. This article provides an exhaustive psychometric exposition of the MFFT, detailing its historical development, operational administration, structural validity, empirical scoring debates, and modern clinical standing.

2. Keywords

Matching Familiar Figures Test, MFFT, Jerome Kagan, Reflection-Impulsivity, Cognitive Style, Information Processing, Visual Discrimination, Latency, Error Rate, Executive Functioning, Developmental Psychology, Perceptual Matching, Impulsivity Measurement

3. Authors

The Matching Familiar Figures Test was conceptualized, constructed, and standardized by an investigative team at the Fels Research Institute in Yellow Springs, Ohio, under the leadership of Jerome Kagan, Ph.D. The primary authors credited in the foundational monograph (Kagan et al., 1964) are:

• Jerome Kagan, Ph.D. (1929–2021): Principal Investigator at the Fels Research Institute; subsequently the Daniel and Amy Starch Professor of Psychology at Harvard University. Kagan was a pioneering figure in developmental psychology, temperamental theory, and cognitive differentiation.
• Bernice L. Rosman, Ph.D.: Research Associate at the Fels Research Institute; widely recognized for her later clinical contributions to developmental psychopathology and systemic family therapy at the Philadelphia Child Guidance Clinic.
• Deborah Day, B.A.: Research Assistant and Psychometrician, Fels Research Institute.
• Joseph Albert, M.A.: Psychometrician and Statistical Analyst, Fels Research Institute.
• William Phillips, B.A.: Research Assistant, Department of Psychology, Fels Research Institute.

Inquiries regarding historic archival data and development papers are preserved through the Department of Psychology at Harvard University and historical repositories of the Fels Longitudinal Study.

4. Purpose

The fundamental purpose of the Matching Familiar Figures Test is to provide an objective, behavioral measurement of individual differences in cognitive tempo and information processing under conditions of high response uncertainty. In situations where multiple plausible response alternatives exist and the correct answer is not immediately self-evident, individuals exhibit systematic variations in their propensity to pause and evaluate the validity of potential solutions before executing a behavioral response. Kagan termed this stable individual difference dimension “reflection-impulsivity” (R-I).

Rather than assessing general intellectual ability, vocabulary, or learned scholastic aptitude, the MFFT was created to isolate the strategic, stylistic decision-making processes an individual deploys during problem-solving. While traditional intelligence batteries (such as the Wechsler Intelligence Scales or Stanford-Binet) measure the terminal accuracy or capacity of cognitive operations, the MFFT captures the micro-structural dynamic between processing time and decision quality. Specifically, it addresses why two children possessing identical general mental ages or IQ scores may approach an ambiguous problem with entirely discordant strategies: one pausing deliberately to scan all available alternatives for minute discrepancy features, while the other acts precipitous upon the first marginally acceptable option.

In clinical and educational domains, the MFFT was implemented to achieve several distinct diagnostic and applied objectives:

• Identification of Learning Difficulties: Identifying children who struggle in early reading and mathematics not because of structural intellectual impairment, but due to an impulsive cognitive tempo that leads to premature perceptual closure, word-guessing, and erratic visual decoding.
• Neurodevelopmental and Behavioral Screening: Serving as an objective marker for behavioral disinhibition, hyperkinesis, and what is currently conceptualized as Attention-Deficit/Hyperactivity Disorder (ADHD), particularly in distinguishing inattentive from hyperactive-impulsive presentations.
• Evaluation of Cognitive Modification Programs: Providing a sensitive, pre- and post-intervention outcome metric for self-instructional training, cognitive behavioral interventions, and pharmacological regimens designed to enhance inhibitory control and sustained visual attention.
• Cross-Sectional and Longitudinal Research: Investigating the ontogenetic transition from the rapid, error-prone perceptual style typical of early childhood to the systematic, reflective, analytic processing characteristic of mature adolescence and adulthood.

5. Psychological Construct

The core psychological construct measured by the MFFT is reflection-impulsivity (also frequently referenced in cognitive psychology as cognitive tempo or conceptual tempo). Reflection-impulsivity represents a pervasive cognitive style that governs how an individual allocates attention, processes perceptual arrays, executes visual scanning strategies, and tolerates decision uncertainty. It is operationalized along a continuum anchored by two distinct phenotypic profiles:

The Reflective Style

Reflective individuals are characterized by prolonged response latencies accompanied by exceptionally low error rates. When confronted with a visual matching task involving six complex alternatives, a reflective individual systematically delays response execution. During this latency period, eye-tracking studies (e.g., Drake, 1970; Zelniker et al., 1972) demonstrate that reflective individuals engage in systematic, exhaustive feature-by-feature comparison across all stimulus variants. They actively formulate visual hypotheses, cross-reference fine-grained details against the target standard, identify subtle discrepancies, and eliminate distractors sequentially. Reflective children exhibit a high threshold of certainty before committing to a decision, prioritizing accuracy over speed.

The Impulsive Style

Impulsive individuals are characterized by remarkably short response latencies coupled with high error frequencies. Upon visual inspection of the target figure, the impulsive individual rapidly selects an alternative based on global, holistic perceptual similarity without conducting an exhaustive component analysis. Eye-tracking paradigms reveal that impulsive individuals exhibit fewer fixations overall, fail to scan every alternative, and direct their gaze predominantly between the target and one or two superficially attractive distractors before responding prematurely. The impulsive style is not governed by sensory impairment, but rather by an urgent need for quick closure, reduced tolerance for cognitive ambiguity, and failure to inhibit behavioral motor execution.

The Four-Quadrant Typology

Historically, researchers using the MFFT categorized respondents using a median-split approach based on group medians for latency ($Md_L$) and errors ($Md_E$):

• Reflective: Latency > $Md_L$ and Errors < $Md_E$. Characterized by deliberate scanning, high inhibitory control, and superior perceptual accuracy.
• Impulsive: Latency < $Md_L$ and Errors > $Md_E$. Characterized by premature behavioral execution, incomplete perceptual evaluation, and elevated error rates.
• Fast-Accurate: Latency < $Md_L$ and Errors < $Md_E$. Highly efficient information processors possessing superior visual discrimination speed and rapid parallel visual processing capacities.
• Slow-Inaccurate: Latency > $Md_L$ and Errors > $Md_E$. Individuals who delay responding yet continue to commit frequent errors, often indicating generalized information-processing deficits, low confidence, anxiety, or visual-spatial processing difficulties.

Although the reflective and impulsive classifications represent the theoretical anchors of interest, the existence of the off-diagonal groups (fast-accurate and slow-inaccurate) stimulated substantial psychometric inquiry into whether reflection-impulsivity is truly a unidimensional bipolar trait or an artifact of two distinct constructs: decision latency and cognitive ability.

6. Theoretical Framework

The Matching Familiar Figures Test emerged during the cognitive revolution of the 1960s, heavily informed by information processing theories, developmental stage paradigms, and early formulations of ego control and executive functioning. Jerome Kagan formulated the theoretical foundation of reflection-impulsivity by synthesizing three core psychological mechanisms:

1. The Cognitive Anxiety Model of Error Aversion

Kagan hypothesized that an individual’s position along the reflection-impulsivity continuum reflects their internal balance between two competing forms of anxiety: the anxiety of committing an error versus the anxiety of appearing slow or incompetent. Reflective children exhibit an acute aversion to making mistakes; their subjective cost of an error is exceptionally high. Consequently, they experience intense cognitive anxiety when tempted to respond prematurely, which acts as an internal braking mechanism that prolongs evaluation time. In contrast, impulsive children place far greater subjective cost on behavioral delay, remaining comparatively indifferent to potential errors, and thus prioritize rapid task completion over verification.

2. Microgenetic Information Processing and Visual Search

The task draws upon the microgenetic processing theories of visual perception articulated by Heinz Werner and Ulric Neisser. Visual identification occurs through a progression from global, syncretic feature recognition to differentiated, analytic feature extraction. Kagan postulated that impulsive children truncate this microgenetic sequence at the initial global stage, reacting to the overall “gestalt” of the familiar figure. Reflective children consistently transition into the analytic stage, breaking the holistic image down into constituent morphological sub-units (e.g., orientation of animal limbs, line curvatures, window placements on a house) and systematically executing hypothesis testing.

3. Developmental Maturation of Inhibitory Control

From an ontogenetic perspective, reflection-impulsivity is firmly rooted in developmental cognitive maturation. Normal cognitive development between ages 5 and 12 is characterized by a natural linear increase in response latency and a concurrent decrease in error frequency. As the prefrontal cortex matures, children acquire greater inhibitory control over motor impulsivity, expand their working memory capacity, and develop meta-cognitive strategies. Kagan’s framework conceptualized reflection as a developmentally superior, mature cognitive style that enables complex problem-solving, reading comprehension, and deductive reasoning.

7. Validity

The construct, convergent, discriminant, and predictive validity of the Matching Familiar Figures Test has been extensively investigated over six decades of empirical research.

Construct and Convergent Validity

Construct validity for the MFFT was primarily established through convergence with other cognitive and behavioral indicators of deliberation and self-regulation. Early studies demonstrated significant correlations between MFFT reflection and performance on tasks requiring delayed motor execution, such as the Walk-a-Line-Slowly and Draw-a-Line-Slowly tests (Maccoby et al., 1965). Children classified as reflective on the MFFT display significantly greater persistence on difficult cognitive puzzles, demonstrate superior performance on the Haptic-Visual Matching Test, and perform with higher precision on tests of field independence, such as Witkin’s Embedded Figures Test (Kagan et al., 1964; Messer, 1976).

Furthermore, convergent validity is robustly supported by eye-movement recordings during test administration. Drake (1970) and subsequent researchers confirmed that reflective subjects exhibit distinct gaze patterns: they make significantly more cross-referencing visual transitions between the standard and the comparison variants, distribute their fixations equally across all six variants, and systematically check feature points prior to verbalizing an answer. Impulsive subjects, conversely, display erratic visual search paths, fewer overall fixations, and prolonged gazes directed at single incorrect alternatives, confirming that the test scores accurately map onto structural visual information processing.

Discriminant Validity

A central theoretical premise asserted by Kagan was that reflection-impulsivity constitutes an independent cognitive style that cannot be reduced merely to general intelligence (IQ). Studies investigating the correlation between MFFT metrics and standardized IQ measures (e.g., WISC verbal and performance scores) generally show low to modest correlations between response latency and IQ ($r = .10$ to $.25$, often non-significant), confirming that the tendency to pause before responding is largely independent of psychometric intelligence. However, MFFT error scores exhibit moderate negative correlations with IQ ($r = -.30$ to $-.50$), an empirical reality that sparked extensive psychometric debate regarding whether error rates partly reflect general intellectual problem-solving ability rather than pure cognitive style (Block, Block, & Harrington, 1974).

Predictive and Clinical Validity

The MFFT demonstrates exceptional predictive validity across academic and clinical settings:

• Academic Achievement: Longitudinal studies show that reflective kindergarten and first-grade children significantly outperform their impulsive peers on standardized reading achievement tests, word recognition, and reading comprehension at grades two and three (Kagan, 1965; Messer, 1976). Impulsive children frequently commit word-reading errors characterized by initial-letter guessing (e.g., reading “horse” for “house”).
• Clinical Neuropsychology: The MFFT robustly differentiates children with ADHD from typically developing controls. Children with ADHD commit significantly more errors and exhibit truncated first-response latencies. Furthermore, double-blind clinical trials examining methylphenidate and atomoxetine administration have consistently used the MFFT as a sensitive functional measure, showing marked normalization of latency and significant reductions in error scores post-pharmacotherapy.

8. Reliability

The reliability of the Matching Familiar Figures Test has been evaluated across multiple parameters, including internal consistency (split-half reliability), inter-rater reliability, and longitudinal test-retest stability.

Internal Consistency

In the original standardization monograph by Kagan, Rosman, Day, Albert, and Phillips (1964), the split-half reliability coefficient for response latency was found to be exceptionally high. Based on an analysis of 300 test protocols of elementary school children in grades 1 through 4, the Spearman-Brown corrected split-half reliability coefficient for latency was $.94$. Split-half reliability coefficients for error scores are typically lower but remain acceptable, commonly falling within the $.62$ to $.85$ range across various age cohorts (Messer, 1976; Paulsen & Johnson, 1980).

Test-Retest Stability

Longitudinal stability exhibits notable differentiation between the two primary test metrics:

• Response Latency: Response time exhibits robust test-retest stability. Across 12-month retest intervals in primary school populations, latency stability coefficients typically range from $r = .60$ to $r = .75$. Over extended intervals spanning two to three years, latency continues to display moderate to high stability ($r = .48$ to $.65$), confirming that cognitive tempo operates as an enduring individual disposition.
• Error Scores: Total errors show moderate short-term stability ($r = .50$ to $.70$ over weeks to months), but lower stability over annual and multi-year spans ($r = .35$ to $.55$). Because error reduction is strongly linked to general cognitive maturation, scholastic instruction, and practice effects, errors demonstrate greater developmental plasticity than response latency.

Inter-Rater Reliability

Because the MFFT requires an examiner to monitor response timing using a manual stopwatch and visually record both the chosen variant number and total attempts, examiner precision is critical. Inter-rater reliability studies between simultaneous independent observers utilizing standardized stopwatch protocols report near-perfect agreement, with intraclass correlation coefficients exceeding $.98$ for latency and $.99$ for absolute error counts.

9. Factor Analysis

The underlying factor structure of the Matching Familiar Figures Test has been the subject of intensive psychometric scrutiny, focusing primarily on whether latency and errors reflect a single bipolar construct or two separable, oblique dimensions of performance.

Bipolar Unidimensionality vs. Two-Factor Independence

In early factor-analytic investigations of cognitive style test batteries, Kagan and colleagues found that MFFT latency and MFFT errors loaded heavily in opposite directions on a single primary factor, which they labeled “Reflection-Impulsivity.” Under this bipolar model, high latency loadings were mirrored by inverse error loadings of comparable magnitude.

However, subsequent exploratory factor analyses (EFA) and confirmatory factor analyses (CFA) conducted by independent psychometricians (e.g., Block, Block, & Harrington, 1974; Bentler & McClain, 1976) challenged this assumption. When MFFT items are factor-analyzed alongside broad psychometric batteries measuring non-verbal intelligence, verbal ability, field dependence, and motor inhibition, the two parameters frequently separate into two distinct factors:

• Factor 1: Decision Time / Cognitive Tempo: Primarily defined by first-response latencies across all 12 experimental trials, exhibiting factor loadings typically between $.70$ and $.88$. This factor accounts for 35% to 45% of the variance and is largely orthogonal to general intelligence.
• Factor 2: Perceptual Competence / Accuracy: Primarily defined by total errors across the 12 items, showing factor loadings between $.60$ and $.82$. This factor accounts for 15% to 25% of the total variance and frequently cross-loads moderately onto non-verbal reasoning and visual-spatial intelligence factors.

The Salkind and Wright Continuous $I$-Index Formulation

To reconcile these factor-analytic findings and overcome the statistical limitations of the median-split technique (which discards substantial variance and ignores the off-diagonal groups), Salkind and Wright (1977) developed a continuous psychometric model. Using standardized $z$-scores derived from normative samples, they formulated two continuous, orthogonal dimensions:

• Impulsivity Index ($I$): Calculated as $I = z_E – z_L$, where $z_E$ is the standardized error score and $z_L$ is the standardized latency score. A high positive $I$-score reflects classic impulsivity (high errors, short latency), while a high negative $I$-score reflects reflection (low errors, prolonged latency).
• Efficiency Index ($E$): Calculated as $E = z_E + z_L$. This index captures overall cognitive efficiency, cleanly distinguishing fast-accurate individuals (low $E$) from slow-inaccurate individuals (high $E$).

Structural equation modeling confirms that the $I$-index accounts for the shared cognitive tempo variance, while the $E$-index isolates the ability/competence component, validating the construct’s operational multidimensionality.

10. Instrument / Measurement Tool

The original standardized Matching Familiar Figures Test (MFFT) consists of an examiner-administered, match-to-sample perceptual discrimination task. Below is the operational overview of the instrument’s design, administration parameters, and scoring rules:

• Test Type: Performance-based behavioral test of cognitive style and perceptual discrimination.
• Target Population: Children ages 5 through 12 (primary school cohorts, Grades 1–4); subsequent adolescent and adult adaptations (e.g., MFFT-20) extend from age 13 into older adulthood.
• Test Materials:
  • Spiral-bound stimulus booklet or presentation binder containing two practice item sets and 12 numbered experimental test sets.
  • Each test set consists of two vertically or horizontally arranged display plates: a Standard Figure plate (containing a single target image) and a Response Array plate (containing six numbered variants arranged in two rows of three).
  • Standardized recording/scoring protocol sheet.
  • Mechanical or digital stopwatch (silent operation required).
• Item Count:
  • 2 Practice/Demonstration Trials (Item P1: Line drawing of a familiar object; Item P2: Familiar object).
  • 12 Experimental Test Trials (familiar, highly recognizable common objects, such as a house, scissors, cowboy, lamp, giraffe, telephone, ship, etc.).
• Variant Composition:
  • 1 Identical Match (exact duplicate of the standard figure in size, line width, orientation, and detailing).
  • 5 Distractor Figures (each distractor differs from the standard by exactly one specific, isolated graphic feature, such as a missing line, altered shape, modified curve, or inverted detail).
• Standard Administration Protocol:
  • The examiner places the stimulus booklet flat in front of the child.
  • Standard Verbatim Instructions: “I am going to show you a picture of something you know, and then some pictures that look almost like it. You are to look at the picture on top, and then find the one on the bottom that is just like it—an exact match. Let’s practice on this one.”
  • The child completes practice items P1 and P2 with corrective feedback to ensure task comprehension.
  • For experimental items 1 through 12: The examiner exposes the standard and variants simultaneously and discreetly starts the stopwatch immediately upon visual presentation.
  • The stopwatch is stopped the precise instant the child verbalizes or points to their first chosen figure. This duration is recorded as First Response Latency (to the nearest tenth or half second).
  • If the initial selection is correct, the examiner records “+” and advances to the next trial.
  • If the selection is incorrect, the examiner states: “No, that is not the one. Find the one that is just like it.” The examiner restarts the stopwatch, records the specific incorrect figure number, and continues the trial until the child selects the correct figure or reaches a maximum of six failed attempts per item.
• Derived Scores and Metrics:
  • Mean Response Latency: The sum of the first response latencies across all 12 items divided by 12 (in seconds).
  • Total Errors: The cumulative sum of all incorrect choices made across all 12 items before finding the correct match.
  • First-Choice Correct Items: The total number of items correctly solved on the initial attempt (range: 0 to 12).
  • Cognitive Style Classification: Determined via median split (Reflective, Impulsive, Fast-Accurate, Slow-Inaccurate) or via standardized continuous indices ($I$-score and $E$-score).

11. Permissions & Fee and Test Year

The original Matching Familiar Figures Test was developed in 1964 as part of federally funded research at the Fels Research Institute. The primary experimental procedures, item specifications, and theoretical findings were placed in the public domain via the American Psychological Association publication (Kagan et al., 1964).

Because the original 12-item MFFT was created as an experimental laboratory instrument rather than a commercial test kit, no single commercial test publisher holds an exclusive global proprietary lock over the historical 1964 research plates. Qualified academic researchers and licensed clinicians have historically obtained reproduction plates, testing booklets, and normative manuals directly from archival psychology departments, the Fels Institute, or through research reprints. Later revisions, such as the 20-item Matching Familiar Figures Test (MFFT-20) developed by Cairns and Cammock (1978), as well as computerized adaptations (e.g., the Cambridge Neuropsychological Test Automated Battery [CANTAB] match-to-sample analogs), are commercially licensed and subject to specific publisher fees and user qualification levels (Level B/C).

12. References

Bentler, P. M., & McClain, J. (1976). A multitrait-multimethod analysis of reflection-impulsivity. Child Development, 47(1), 218–226. https://doi.org/10.2307/1128302

Block, J., Block, J. H., & Harrington, D. M. (1974). Some misgivings about the Matching Familiar Figures Test as a measure of reflection-impulsivity. Developmental Psychology, 10(5), 611–632. https://doi.org/10.1037/h0037048

Cairns, E., & Cammock, T. (1978). Development of a more reliable version of the Matching Familiar Figures Test. Developmental Psychology, 14(5), 555–560. https://doi.org/10.1037/0012-1649.14.5.555

Drake, D. M. (1970). Perceptual correlates of impulsive and reflective behavior. Developmental Psychology, 2(2), 202–214. https://doi.org/10.1037/h0028787

Kagan, J. (1965). Reflection-impulsivity and reading ability in primary grade children. Child Development, 36(3), 609–628. https://doi.org/10.2307/1126830

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

Maccoby, E. E., Dowley, E. M., Hagen, J. W., & Degerman, R. (1965). Activity level and intellectual functioning in normal preschool children. Child Development, 36(3), 761–770. https://doi.org/10.2307/1126840

Messer, S. B. (1976). Reflection-impulsivity: A review. Psychological Bulletin, 83(6), 1026–1052. https://doi.org/10.1037/0033-2909.83.6.1026

Paulsen, K., & Johnson, M. (1980). Impulsivity: A discussion of its clinical and theoretical utility, measurement problems, and relationship to academic achievement. Journal of School Psychology, 18(2), 177–187. https://doi.org/10.1016/0022-4405(80)90040-5

Salkind, N. J., & Wright, J. C. (1977). The development of reflection-impulsivity and cognitive efficiency: An integrated model. Human Development, 20(6), 377–387. https://doi.org/10.1159/000271569

Zelniker, T., Jeffrey, W. E., Ault, R., & Parsons, J. (1972). Analysis and modification of search strategies of impulsive and reflective children on the Matching Familiar Figures Test. Child Development, 43(1), 353–368. https://doi.org/10.2307/1127768

13. Items of the Scale