Skip to main content
v2026.11,986 entries · CC-BY 4.0

Tower of London Test: Administration and Scoring Guide

A comprehensive guide to the Tower of London test, covering administration rules, move scoring formulas, normative data, APA reporting, and code scripts.

Written and maintained by CASRAI Editorial Board

Published

On this page

The Tower of London test is a standardized neuropsychological assessment designed to evaluate executive functioning, specifically spatial planning, working memory, mental look-ahead, and rule-governed problem-solving in clinical and research settings. Originally developed by British neuropsychologist Tim Shallice in 1982 to examine planning impairments in patients with focal frontal lobe lesions, the instrument requires an examinee to manipulate three colored beads across three vertical pegs of varying heights to match a target configuration in the minimum number of moves. By quantifying problem-solving efficiency, planning latency, and rule adherence, the test provides empirical insights into prefrontal cortex function across developmental, neurological, and psychiatric cohorts.

This page describes products and general use; it is not medical advice.

Definition and Theoretical Background

The Tower of London (TOL) test assesses an individual’s capacity to formulate, execute, and monitor a multi-step sequence of goal-directed actions. In cognitive neuropsychology, planning is conceptualized as the ability to anticipate future consequences, organize subgoals, and select an optimal sequence of operations before initiating action. Shallice formulated the test within the framework of the Supervisory Attentional System (SAS), a theoretical model distinguishing automatic, schema-driven behavioral routines from conscious, non-routine executive control. In novel or complex tasks where automatic action schemas conflict or fail to reach an objective, the SAS—anatomically grounded in the prefrontal cortex—must intervene to inhibit prepotent responses, generate alternative goal hierarchies, and evaluate intermediate problem states.

Although structurally related to the classic Tower of Hanoi puzzle, the Tower of London incorporates structural constraints specifically engineered to isolate planning from simple perceptual recursion. In the Tower of Hanoi, three equal-height pegs accommodate all three disks, permitting cyclic or recursive subroutines that can be solved using rote perceptual heuristics. In contrast, Shallice’s Tower of London introduces asymmetrical peg heights that accommodate exactly one, two, and three beads respectively. This physical restriction severely constrains the problem space, forcing examinees to create temporary intermediate states where a bead must often be placed in a counter-intuitive holding position before moving toward the final target peg. Solving higher-level configurations demands robust working memory manipulation and forward look-ahead search, making the test an exceptionally sensitive measure of dorsolateral prefrontal cortex (DLPFC) integrity.

Apparatus Specifications and Administration Rules

Standard physical testing utilizes two identical wooden boards or apparatus frames placed on a flat examination surface. Each board features a horizontal base with three vertical pegs of graded lengths: the tallest peg holds a maximum of three beads, the middle peg holds two beads, and the shortest peg holds exactly one bead. Three spherical wooden beads—colored red, green, and blue—are mounted on the pegs. One apparatus serves as the examiner’s model (the goal state), while the identical second board serves as the examinee’s apparatus (the initial state).

Standardized administration protocols follow rigid experimental instructions to ensure psychometric uniformity across administrations:

  • Target Matching: The examiner arranges their board into a specified target configuration and presents the examinee board in a standardized start configuration. The examinee is instructed to transform their board to match the examiner’s target using the fewest possible moves.
  • Single-Bead Movement: The examinee may manipulate only one bead at a time using one hand. Two-handed manipulation or simultaneous transfers are prohibited.
  • Blocked Bead Restriction: A bead cannot be moved if another bead rests on top of it. Only the uppermost, exposed bead on any peg may be transferred.
  • Peg Capacity Limits: Peg height limits must be strictly respected. Placing a fourth bead on the tall peg, a third bead on the middle peg, or a second bead on the short peg constitutes a physical rule violation.
  • Placement Continuity: Beads must be transferred directly from peg to peg. Examinees cannot place beads on the table, drop them, or hold intermediate beads in their hands while moving others.
  • Pre-Planning Instruction: Examinees are explicitly directed to plan their entire move sequence mentally before making their first physical move.

In contemporary research, computerized implementations (such as the Cambridge Neuropsychological Test Automated Battery [CANTAB] One-Touch Stockings of Cambridge, the Psychology Experiment Building Language [PEBL] Tower of London, and Inquisit) simulate the two boards on touchscreens or computer monitors. While computerized variants eliminate physical manipulation errors and automatically record reaction times with millisecond precision, researchers must account for differences in motor interface demands when linking scores across modes, referencing established methodologies in test equating and linking.

Scoring Formulas and Mathematical Metrics

Performance on the Tower of London test is captured across multiple distinct quantitative dimensions rather than a single composite score. Evaluated dimensions include accuracy, move efficiency, planning and execution latencies, and executive inhibition failures.

Move performance across a test battery of k problems is governed by the following mathematical relations:

Let Mi represent the actual number of moves executed by the participant to reach the goal configuration on problem i, and let Mmin,i denote the theoretical minimum moves required for problem i. The Total Moves executed across the test is given by:

Mtotal = Σ Mi (for i = 1 to k)

The Excess Moves score (ΔMi) on problem i reflects deviations from the optimal solution path:

ΔMi = Mi – Mmin,i

where Mi ≥ Mmin,i, ensuring ΔMi ≥ 0. The Total Excess Moves score across all k problems represents the cumulative inefficiency score:

Emoves = Σ (Mi – Mmin,i) = Mtotal – Σ Mmin,i

In the standardized 12-problem protocol established by Krikorian, Bartok, and Gay (1994), participants are permitted up to three attempts per problem configuration. Accuracy is scored through a weighted points system based on initial trial success:

  • 3 points: Problem solved in the minimum number of moves on Attempt 1.
  • 2 points: Problem solved in the minimum number of moves on Attempt 2.
  • 1 point: Problem solved in the minimum number of moves on Attempt 3.
  • 0 points: Failure to solve the problem within three attempts or exceeding trial move ceilings.

The Krikorian Total Score (Stotal) is the sum of trial points across all 12 items:

Stotal = Σ si (for i = 1 to 12)

Because each of the 12 problems yields between 0 and 3 points, Stotal ranges from 0 to 36 points. A higher score reflects superior planning efficiency and immediate solution accuracy.

Temporal dynamics are divided into two distinct chronometric phases:

  • Initiation Time (Pre-Planning Latency, Tinit): The duration in seconds from initial problem presentation to the physical initiation of the first bead movement: Tinit = tfirst_move – tdisplay. This metric operationalizes deliberate, proactive mental look-ahead.
  • Execution Time (Movement Duration, Texec): The duration from the first bead movement to problem completion: Texec = tfinish – tfirst_move. This captures visuomotor coordination, online monitoring, and reactive adjustments during problem resolution.
  • Total Problem Time (Ttotal): The aggregate problem-solving duration: Ttotal = Tinit + Texec.

Inhibition and cognitive control are quantified through Rule Violations. Standardized protocols distinguish Type I violations (Ball Violations: moving two beads at once or lifting a bead trapped beneath another) from Type II violations (Peg Violations: exceeding peg height capacities). The Total Rule Violations metric serves as an index of behavioral impulsivity and working memory degradation under cognitive load.

To enable normative comparison against age-matched reference cohorts, raw performance scores are transformed into standard scores using the sample mean (μ) and standard deviation (σ). The standard deviation provides the unit scale for measuring deviation from expected developmental baselines, as outlined in our guide to standard deviation. The standard z-score is calculated as:

z = (X – μ) / σ

where X is the examinee raw score, μ is the normative mean, and σ is the normative standard deviation. Standard Scores (SS) with a mean of 100 and a standard deviation of 15 are then computed via linear transformation:

SS = 100 + 15 × z

For detailed properties and interpretation of standardized metrics, consult the comprehensive reference on z-scores.

Worked Example with Step-by-Step Scoring Arithmetic

To demonstrate how raw behavioural data translates into psychometric indices, consider an illustrative evaluation of an 8-year-old child (enrolled in Grade 2) administered the 12-problem Krikorian Tower of London protocol. The test spans two 2-move problems, two 3-move problems, four 4-move problems, and four 5-move problems, yielding a cumulative minimum move requirement of 42 moves.

The examinee completed the 12 problem sets with the following trial record:

Problem ID Minimum Moves (Mmin) Attempt Solved Moves on Final Attempt Points Awarded (si) Excess Moves (ΔMi) Initiation Time (s) Execution Time (s)
Problem 1 2 Attempt 1 2 3 0 3.2 4.1
Problem 2 2 Attempt 1 2 3 0 2.8 4.5
Problem 3 2 Attempt 1 2 3 0 3.5 5.0
Problem 4 3 Attempt 1 3 3 0 4.1 6.8
Problem 5 3 Attempt 2 3 2 2 5.6 9.4
Problem 6 3 Attempt 1 3 3 0 4.8 7.2
Problem 7 4 Attempt 1 4 3 0 6.2 10.5
Problem 8 4 Attempt 1 4 3 0 5.9 11.1
Problem 9 4 Attempt 2 4 2 2 7.4 14.2
Problem 10 5 Attempt 1 5 3 0 8.5 16.8
Problem 11 5 Attempt 3 5 1 4 9.1 22.4
Problem 12 5 Attempt 1 5 3 0 7.8 15.6

We execute the arithmetic calculations step by step:

Step 1: Calculate Total Score (Stotal)
Summing the individual points earned across all 12 items:
Stotal = 3 + 3 + 3 + 3 + 2 + 3 + 3 + 3 + 2 + 3 + 1 + 3 = 32 points.
The participant achieved 32 out of a maximum possible 36 points, successfully completing 9 of the 12 problems on the initial attempt.

Step 2: Calculate Excess Moves (Emoves)
Across the successful completion trials, excess moves accumulated on failed attempts prior to resolution:
Problem 5 (resolved on attempt 2 after a 5-move first trial): 5 – 3 = 2 excess moves.
Problem 9 (resolved on attempt 2 after a 6-move first trial): 6 – 4 = 2 excess moves.
Problem 11 (resolved on attempt 3 after two 7-move trials): (7 – 5) + (7 – 5) = 4 excess moves.
Cumulative Excess Moves: Emoves = 0 + 0 + 0 + 0 + 2 + 0 + 0 + 0 + 2 + 0 + 4 + 0 = 8 excess moves.

Step 3: Calculate Mean Initiation and Execution Latencies
Total Initiation Time = 3.2 + 2.8 + 3.5 + 4.1 + 5.6 + 4.8 + 6.2 + 5.9 + 7.4 + 8.5 + 9.1 + 7.8 = 68.9 seconds.
Mean Initiation Time = 68.9 / 12 = 5.74 seconds per problem.
Total Execution Time = 4.1 + 4.5 + 5.0 + 6.8 + 9.4 + 7.2 + 10.5 + 11.1 + 14.2 + 16.8 + 22.4 + 15.6 = 127.6 seconds.
Mean Execution Time = 127.6 / 12 = 10.63 seconds per problem.

Step 4: Standardize Total Score Against Grade 2 Normative Data
From the empirical normative dataset compiled by Krikorian et al. (1994, Table 2), the normative parameters for Grade 2 children (mean age 8.1 years, N = 25) are:
Mean (μ) = 29.0 points, Standard Deviation (σ) = 2.6 points.
Applying the z-score formula:
z = (32 – 29.0) / 2.6 = 3.0 / 2.6 = +1.1538 ≈ +1.15.
Converting to a Standard Score (SS):
SS = 100 + 15 × (+1.1538) = 100 + 17.31 = 117.31 ≈ 117.
Computing the Percentile Rank (PR) from the standard normal cumulative distribution function:
PR = Φ(1.1538) × 100 ≈ 87.6th percentile.

Interpretation and Normative Reading Guidance

Interpreting Tower of London metrics requires evaluating move efficiency, latency profiles, and error patterns simultaneously. A summary standard score provides a macro-level snapshot of planning proficiency, but qualitative task dynamics reveal the underlying cognitive mechanisms.

The following interpretation framework provides clinical and research benchmarks for standard scores and z-score ranges:

Z-Score Range Standard Score (SS) Percentile Rank Clinical Classification Cognitive Functioning Profile
≥ +2.00 ≥ 130 ≥ 98th Very Superior Exceptional mental look-ahead, rapid subgoal generation, zero excess moves.
+1.00 to +1.99 115–129 84th–97th High Average Robust planning accuracy, efficient working memory look-ahead, minimal errors.
-0.99 to +0.99 85–114 16th–83rd Average Developmentally appropriate planning capacity, occasional trial-and-error.
-1.00 to -1.49 78–84 7th–15th Low Average Mild planning vulnerability, elevated excess moves on 4- and 5-move problems.
-1.50 to -1.99 70–77 2nd–6th Borderline / Mild Deficit Frequent look-ahead failure, pronounced reliance on reactive motor trial-and-error.
≤ -2.00 ≤ 69 Under 2nd Impaired Severe planning breakdown, frequent rule violations, inability to formulate subgoals.

When analyzing individual profiles, clinicians distinguish four primary behavioral patterns based on the dissociation between initiation latency and move accuracy:

  • Reflective, Efficient Planners: Characterized by extended initiation times accompanied by optimal move counts (zero or near-zero excess moves). The examinee invests cognitive effort into pre-planning the full trajectory prior to action, typical of intact frontal systems.
  • Impulsive Responders: Characterized by brief initiation times (under 2 seconds) paired with elevated excess moves and rule violations. The examinee initiates movement immediately without generating subgoals, relying on haphazard online adjustments. This profile is frequently documented in Attention-Deficit/Hyperactivity Disorder (ADHD) and frontal disinhibition syndromes.
  • Sluggish or Inefficient Processors: Exhibited when prolonged initiation times fail to produce efficient moves, resulting in elevated excess moves despite lengthy deliberation. This suggests working memory overload or deficits in visualizing intermediate goal states, commonly observed in traumatic brain injury and vascular cognitive impairment.
  • Motor-Slowed Accurate Responders: Normal or high accuracy with prolonged execution times but normal initiation latencies. This dissociates primary motor slowing (such as in Parkinsonian syndromes) from cognitive planning deficits.

The data table below compares developmental normative parameters compiled from empirical school-age cohorts, illustrating the upward trajectory of planning accuracy across age groups.

Psychometric Properties: Reliability, Validity, and Methodological Rigor

Extensive psychometric investigations have evaluated the reliability and construct validity of the Tower of London across clinical and non-clinical samples. Internal consistency estimates for composite score indices range from acceptable to strong. Krikorian et al. (1994) reported split-half reliability coefficients of r = .78 to .82 across developmental cohorts, while computerized editions report Cronbach’s alpha values typically between α = .72 and .84 for total correct items. For a foundational exploration of internal consistency mathematics, see our guide on Cronbach’s alpha.

Test-retest reliability ranges from moderate (r = .55 to .75) over multi-week intervals. The moderate test-retest magnitude stems from the intrinsic nature of executive problem-solving tasks: once an examinee deduces the optimal holding-peg strategy for a specific configuration, subsequent administrations measure episodic memory retrieval rather than pure novel problem-solving. To preserve validity across longitudinal clinical trials, standardized alternate forms (such as the TOL-DX 2nd Edition) supply matched test versions with equivalent problem spaces.

Construct and criterion-related validity have been confirmed through neuroimaging and neuropsychological lesion studies. Functional magnetic resonance imaging (fMRI) and positron emission tomography (PET) consistently show robust bilateral activation of the dorsolateral prefrontal cortex (Brodmann areas 9 and 46), the anterior cingulate cortex (BA 24/32), and the posterior parietal cortex during Tower of London problem resolution. Furthermore, performance correlates significantly with established frontal battery measures, including the Wisconsin Card Sorting Test (WCST perseverative errors, r = -.42 to -.56) and the Trail Making Test Part B (r = -.38 to -.48), while displaying divergent validity against pure vocabulary knowledge tests.

Methodologists must remain attentive to potential psychometric artifacts. Simple 2- and 3-move problems exhibit pronounced ceiling effects in neurotypical older adolescents and adults, whereas complex 5- to 7-move problems may introduce floor effects in severe neurodegenerative disorders. Identifying and addressing these boundary artifacts is essential for preserving measurement precision, as detailed in our guide to floor and ceiling effects.

Assumptions and Common Testing Mistakes

Valid interpretation of Tower of London outcomes depends on several core experimental assumptions:

  • Intact Sensory and Motor Function: The examinee must possess sufficient visual acuity, color discrimination (to distinguish red, green, and blue beads), and manual dexterity to manipulate physical beads without physical impediment. If motor or visual deficits exist, computerized gaze-tracking or one-touch response paradigms should be considered.
  • Instructional Comprehension: The examinee must thoroughly understand the rules, particularly the negative constraints (cannot lift two beads simultaneously; cannot exceed peg capacity). Practice trials must be administered to verify comprehension before scoring begins.
  • Motivation and Sustained Effort: Executive planning tasks demand substantial mental effort. Fatigue, pain, or poor task engagement can mimic frontal lobe deficits.

Examiners frequently commit administration errors that compromise score reliability:

  • Interrupting Pre-Planning: Urging an examinee to hurry during their initial thinking phase disrupts initiation latency measurement and biases them toward impulsive errors.
  • Failure to Enforce Rule Violations: Overlooking subtle rule breaks—such as temporarily holding a bead while repositioning another—inflates move efficiency scores.
  • Misrecording Intermediate Moves: During physical administration, rapid bead movements can make accurate manual move counting difficult. Utilizing video recording or validated digital platforms eliminates counting errors.
  • Misinterpreting Motor Hesitation as Planning: Conflating bradykinesia (motor slowing) with pre-planning latency (Tinit). Initiation time must be clocked strictly until the hand begins reaching for the first bead.

How to Report Results (APA Style Example)

When reporting Tower of London findings in empirical manuscripts or neuropsychological evaluations, researchers should present descriptive parameters (means, standard deviations), exact sample sizes, standard scores, and effect sizes in accordance with American Psychological Association (APA 7th edition) conventions.

The following sample text illustrates proper reporting for a developmental study:

A standardized 12-problem Tower of London protocol (Krikorian et al., 1994) was administered to assess executive spatial planning and working memory efficiency. Second-grade participants (N = 25) achieved a mean total score of 29.00 points (SD = 2.60, range = 25–34), with an average initiation time of 5.74 s (SD = 1.32) and mean execution latency of 10.63 s (SD = 2.15). An independent-samples t test indicated that children in the targeted working memory intervention group performed significantly higher on total score (M = 32.10, SD = 2.10) than children in the active control group (M = 28.80, SD = 2.50), t(23) = 3.58, p = .002, d = 1.43, 95% CI [0.55, 2.28]. Total excess moves were significantly lower in the intervention group (M = 3.20, SD = 1.80) relative to controls (M = 8.40, SD = 3.10), t(23) = -5.09, p = .0001. Rule violations remained negligible across both groups (M = 0.20, SD = 0.41), confirming that gains reflected enhanced forward planning rather than changes in rule compliance.

How to Calculate Tower of London Scores in Excel, R, and Python

Automating Tower of London score computation ensures reproducibility and reduces scoring errors across large experimental batteries. Below are verified code implementations across three standard computing environments.

Calculating Scores in Microsoft Excel

To calculate metrics for a 12-problem dataset in Microsoft Excel, organize trial records with Problem IDs in column A, minimum moves in column B (cells B2:B13), actual moves executed in column C (cells C2:C13), and points awarded (0–3) in column D (cells D2:D13):

  • Total Raw Score (Cell D14): =SUM(D2:D13)
  • Total Moves Executed (Cell C14): =SUM(C2:C13)
  • Total Minimum Moves (Cell B14): =SUM(B2:B13)
  • Excess Moves (Cell E14): =C14 - B14
  • Z-Score (Grade 2 Norms, Cell F14): =(D14 - 29.0) / 2.6
  • Standard Score (SS, Cell G14): =100 + 15 * F14
  • Percentile Rank (Cell H14): =NORM.S.DIST(F14, TRUE) * 100

Calculating Scores in R

The following R script calculates total accuracy, excess moves, standardized scores, and summary chronometrics from trial data:

# Tower of London Scoring in R
tol_score = function(actual_moves, min_moves, points, init_times, exec_times, norm_mean = 29.0, norm_sd = 2.6) {
  # Input checks
  stopifnot(length(actual_moves) == length(min_moves))
  stopifnot(length(points) == length(actual_moves))
  
  # Core score aggregations
  total_score  = sum(points)
  total_moves  = sum(actual_moves)
  min_required = sum(min_moves)
  excess_moves = total_moves - min_required
  
  # Standardized psychometric conversions
  z_score    = (total_score - norm_mean) / norm_sd
  std_score  = 100 + 15 * z_score
  percentile = pnorm(z_score) * 100
  
  # Temporal summaries
  mean_init = mean(init_times)
  mean_exec = mean(exec_times)
  
  list(
    total_score = total_score,
    total_moves = total_moves,
    excess_moves = excess_moves,
    z_score = round(z_score, 2),
    standard_score = round(std_score, 1),
    percentile_rank = round(percentile, 1),
    mean_initiation_time = round(mean_init, 2),
    mean_execution_time = round(mean_exec, 2)
  )
}

# Example execution using worked example trial data
min_m  = c(2, 2, 2, 3, 3, 3, 4, 4, 4, 5, 5, 5)
act_m  = c(2, 2, 2, 3, 5, 3, 4, 4, 6, 5, 9, 5)
pts    = c(3, 3, 3, 3, 2, 3, 3, 3, 2, 3, 1, 3)
t_init = c(3.2, 2.8, 3.5, 4.1, 5.6, 4.8, 6.2, 5.9, 7.4, 8.5, 9.1, 7.8)
t_exec = c(4.1, 4.5, 5.0, 6.8, 9.4, 7.2, 10.5, 11.1, 14.2, 16.8, 22.4, 15.6)

results = tol_score(act_m, min_m, pts, t_init, t_exec)
print(results)

Calculating Scores in Python

The Python implementation below computes comprehensive Tower of London metrics and standard scores using base Python and standard library mathematics:

import math

def calculate_tol_metrics(trial_records, norm_mean=29.0, norm_sd=2.6):
    """
    Computes Tower of London summary scores, excess moves, and standard scores.
    Each trial record should be a dict with:
      'min_moves', 'actual_moves', 'points', 'init_time', 'exec_time', 'rule_violations'
    """
    total_points = sum(t['points'] for t in trial_records)
    total_moves = sum(t['actual_moves'] for t in trial_records)
    total_min = sum(t['min_moves'] for t in trial_records)
    excess_moves = total_moves - total_min
    total_violations = sum(t.get('rule_violations', 0) for t in trial_records)
    
    mean_init = sum(t['init_time'] for t in trial_records) / len(trial_records)
    mean_exec = sum(t['exec_time'] for t in trial_records) / len(trial_records)
    
    # Standard score calculation
    z_score = (total_points - norm_mean) / norm_sd
    standard_score = 100 + 15 * z_score
    # Standard normal cumulative distribution approximation
    percentile = 0.5 * (1.0 + math.erf(z_score / math.sqrt(2.0))) * 100
    
    return {
        'total_score': total_points,
        'total_moves': total_moves,
        'minimum_moves': total_min,
        'excess_moves': excess_moves,
        'rule_violations': total_violations,
        'z_score': round(z_score, 2),
        'standard_score': round(standard_score, 1),
        'percentile_rank': round(percentile, 1),
        'mean_initiation_time_s': round(mean_init, 2),
        'mean_execution_time_s': round(mean_exec, 2)
    }

# Example execution with trial records
sample_trials = [
    {'min_moves': 2, 'actual_moves': 2, 'points': 3, 'init_time': 3.2, 'exec_time': 4.1, 'rule_violations': 0},
    {'min_moves': 2, 'actual_moves': 2, 'points': 3, 'init_time': 2.8, 'exec_time': 4.5, 'rule_violations': 0},
    {'min_moves': 2, 'actual_moves': 2, 'points': 3, 'init_time': 3.5, 'exec_time': 5.0, 'rule_violations': 0},
    {'min_moves': 3, 'actual_moves': 3, 'points': 3, 'init_time': 4.1, 'exec_time': 6.8, 'rule_violations': 0},
    {'min_moves': 3, 'actual_moves': 5, 'points': 2, 'init_time': 5.6, 'exec_time': 9.4, 'rule_violations': 0},
    {'min_moves': 3, 'actual_moves': 3, 'points': 3, 'init_time': 4.8, 'exec_time': 7.2, 'rule_violations': 0},
    {'min_moves': 4, 'actual_moves': 4, 'points': 3, 'init_time': 6.2, 'exec_time': 10.5, 'rule_violations': 0},
    {'min_moves': 4, 'actual_moves': 4, 'points': 3, 'init_time': 5.9, 'exec_time': 11.1, 'rule_violations': 0},
    {'min_moves': 4, 'actual_moves': 6, 'points': 2, 'init_time': 7.4, 'exec_time': 14.2, 'rule_violations': 0},
    {'min_moves': 5, 'actual_moves': 5, 'points': 3, 'init_time': 8.5, 'exec_time': 16.8, 'rule_violations': 0},
    {'min_moves': 5, 'actual_moves': 9, 'points': 1, 'init_time': 9.1, 'exec_time': 22.4, 'rule_violations': 0},
    {'min_moves': 5, 'actual_moves': 5, 'points': 3, 'init_time': 7.8, 'exec_time': 15.6, 'rule_violations': 0},
]

output = calculate_tol_metrics(sample_trials)
print("Tower of London Summary Output:")
for key, value in output.items():
    print(f"  {key}: {value}")

Frequently asked questions

What is the difference between the Tower of London test and the Tower of Hanoi?

While both instruments measure executive problem-solving, the Tower of London uses pegs of varying heights (accommodating 1, 2, and 3 beads) and requires matching a specific visual goal configuration. In contrast, the Tower of Hanoi uses three equal-height pegs and requires transferring all graduated disks to another peg under a size-ordering rule (larger disks cannot sit on smaller disks). The Tower of London constraints prevent simple recursive or perceptual subroutines, forcing look-ahead planning and temporary counter-intuitive intermediate moves.

What cognitive domains does the Tower of London test measure?

The test evaluates higher-order executive functioning, primarily spatial planning, working memory maintenance and updating, mental look-ahead, response inhibition, and cognitive flexibility. Neuroimaging studies confirm strong activation of the dorsolateral prefrontal cortex, anterior cingulate cortex, and parietal areas during task performance.

How is the Tower of London test scored in clinical neuropsychology?

Standard scoring captures move accuracy (Total Correct Score and Total Excess Moves), latency dynamics (Initiation Time and Execution Time), and behavioral control (Rule Violations and Time Violations). Raw scores are converted to age-adjusted z-scores or standard scores based on normative reference tables to evaluate executive impairment.

What are the physical dimensions and rules of the Tower of London apparatus?

The standard apparatus consists of two identical wooden frames with three vertical pegs holding 1, 2, and 3 colored spherical beads (red, green, blue). Administration rules dictate moving only one bead at a time, moving only unblocked top beads, adhering strictly to peg height capacity limits, and keeping beads on the apparatus without holding or placing them on the table.

What does a high excess moves score indicate on the Tower of London?

A high excess moves score indicates inefficient problem-solving, failure of forward mental look-ahead, or difficulty maintaining subgoals in working memory. When paired with very short initiation times, high excess moves suggest impulsivity and poor response inhibition; when paired with long initiation times, they suggest working memory overload or failure to formulate viable intermediate problem states.

Can the Tower of London test be administered on a computer?

Yes, standardized digital and computerized adaptations (including CANTAB One-Touch Stockings of Cambridge, PEBL Tower of London, and TOL-DX computerized editions) are widely utilized. Automated administration ensures millisecond-accurate timing and standardized problem presentation, though examiners should account for motor interface differences when comparing physical and touchscreen results.

Sources

  • Shallice, T. (1982). Specific impairments of planning. Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences, 298(1089), 199–209. https://doi.org/10.1098/rstb.1982.0082 (Accessed: 2026-11-16)
  • Krikorian, R., Bartok, J., and Gay, N. (1994). Tower of London procedure: A standard method and developmental data. Journal of Clinical and Experimental Neuropsychology, 16(6), 840–850. https://doi.org/10.1080/01688639408402697 (Accessed: 2026-11-16)
  • Culbertson, W. C., and Zillmer, E. A. (2001). Tower of London-Drexel University (TOL-DX): Technical Manual. Multi-Health Systems. https://www.mhs.com (Accessed: 2026-11-16)
  • Lezak, M. D., Howieson, D. B., Bigler, E. D., and Tranel, D. (2012). Neuropsychological Assessment (5th ed.). Oxford University Press. https://global.oup.com (Accessed: 2026-11-16)
  • Berg, W. K., and Byrd, D. L. (2002). The Tower of London spatial problem-solving task: Enhancing clinical and research utility. Journal of Clinical and Experimental Neuropsychology, 24(5), 586–604. https://doi.org/10.1076/jcen.24.5.586.1006 (Accessed: 2026-11-16)
  • Mueller, S. T., and Piper, B. J. (2014). The Psychology Experiment Building Language (PEBL) and PEBL Test Battery. Journal of Neuroscience Methods, 222, 250–259. https://doi.org/10.1016/j.jneumeth.2013.10.024 (Accessed: 2026-11-16)
  • Anderson, V. A., Anderson, P., Northam, E., Jacobs, R., and Mikiewicz, O. (2002). Relationships between cognitive processes and executive function in children. Child Neuropsychology, 8(4), 231–240. https://doi.org/10.1076/chin.8.4.231.13506 (Accessed: 2026-11-16)

Data: Tower of London Developmental Normative Data (Krikorian et al., 1994)

Developmental normative data for the 12-problem Tower of London test across elementary and middle school grades, detailing sample sizes, chronological ages, raw score means, standard deviations, and score ranges compiled from Krikorian, Bartok, and Gay (1994).

Period covered: 1994. Coverage: United States.

Tower of London Developmental Normative Data (Krikorian et al., 1994)
School GradeSample Size (participants)Mean Age (years)Age SD (years)TOL Mean Raw Score (points)TOL Score SD (points)Minimum Score (points)Maximum Score (points)
Grade 12470.3927.83.71933
Grade 2258.10.3292.62534
Grade 32490.2630.32.72635
Grade 424100.3930.42.92535
Grade 524110.3530.52.82636
Grade 62612.20.3331.82.92536
Grade 72712.90.3431.62.32634
Grade 83113.80.3232.51.92835

Sources

Compiled by CASRAI from the sources above. Free to reuse under CC BY 4.0 with attribution.

Follow CASRAI

Research-administration guidance, standards updates and independent tool reviews.

Ask CASRAI · free to try

Ask about Tower of London Test: Administration and Scoring Guide

2 free questions a day, no account.

An AI assistant for research: funder rules, compliance, methods and publishing. It links the sources behind its answers and labels answers that come from a web search.

Answers draw on CASRAI's guides and dictionary plus the federal and funder documents we index: Federal Register, Grants.gov, Regulations.gov and UKRI.

Works on this site and inside Claude, Cursor and the AI tools you already use.

Everything CASRAI publishes — this page, the dictionary, the guides and the news — stays open to read, with no account and no card.

Referenced across the research world

University of Cambridge logoColumbia University logoCrossref logoUniversity of Edinburgh logoHarvard University logoUniversity of Oxford logoPrinceton University logoStanford School of Medicine logoUniversity College London logoORCID logoUniversity of Cambridge logoColumbia University logoCrossref logoUniversity of Edinburgh logoHarvard University logoUniversity of Oxford logoPrinceton University logoStanford School of Medicine logoUniversity College London logoORCID logo
  • University of Cambridge logo
  • Columbia University logo
  • Crossref logo
  • University of Edinburgh logo
  • Harvard University logo
  • University of Oxford logo
  • Princeton University logo
  • Stanford School of Medicine logo
  • University College London logo
  • ORCID logo

View CASRAI adoption →