Trail Making Test — Standard Scores & Reference Values
How long should a trail making test take? This page collects the published reference values — the classic adult averages, the Reitan–Wolfson classification bands, and the age and education effects from the large Tombaugh normative study — with the source attached to every figure.
What the trail making test is — and what typical times look like
The trail making test (TMT) began as “Partington’s Pathways” in 1938, was adopted into the US Army Individual Test Battery in 1944, and was standardized for clinical use by Reitan in the 1950s — today it is one of the most widely used neuropsychological tests (https://www.millisecond.com/library/trailmakingtest). The standard paper form gives each part 25 circles: Part A is numbered 1–25; Part B mixes numbers 1–13 with letters A–L and requires alternating between the two sequences (https://ubccpd.ca/media/2762/download). Both parts are scored as seconds to completion, so higher scores mean slower performance.
The two parts probe different abilities. Part A is mainly visual search, sustained attention and processing speed; Part B adds task switching, working memory and cognitive flexibility — the reason Part B is treated as a quick executive-function screen (Sánchez-Cubillo et al. 2009: https://www.cambridge.org/core/journals/journal-of-the-international-neuropsychological-society/article/construct-validity-of-the-trail-making-test-role-of-taskswitching-working-memory-inhibitioninterference-control-and-visuomotor-abilities/453A48DA4B70F8BE793106A52BC0FE89; Salthouse 2011: https://doi.org/10.1016/j.intell.2011.03.001). A widely used clinical summary puts the adult average at 29 seconds for Trail A (deficient above 78; most people finish within 90) and 75 seconds for Trail B (deficient above 273; most within 3 minutes) — values drawn from Corrigan & Hinkeldey 1987 and Gaudino et al. 1995 (https://ubccpd.ca/media/2762/download).
The largest single normative dataset is Tombaugh 2004: 911 community-dwelling adults aged 18–89, with performance slowing steadily as age rises and education falls, published as norms stratified into 11 age groups and 2 education levels (https://www.sciencedirect.com/science/article/pii/S0887617703000398). For orientation, Reitan & Wolfson (1993) classify Trail A times of 0–39 s as normal, 40–51 s mildly impaired and 52 s or more moderately-to-severely impaired; for Trail B the bands are 0–85 s, 86–120 s and 121 s or more — with the caveat that age- and education-corrected scores are recommended (as summarized in Groth-Marnat’s Handbook of Psychological Assessment: https://www.nu.edu.ye/wp-content/uploads/2014/05/Groth-Marnat-The-Handbook-of-Psychological-Assessment.pdf). One more honest caveat: computerized versions systematically differ from the paper original, so online times should only be compared with online times (https://graphicsinterface.org/wp-content/uploads/gi2022-19.pdf).
Reference values, each with its source
| Measure | Reference value | Source |
|---|---|---|
| Test format (paper original) | 25 circles per part; Part A numbers 1–25, Part B numbers 1–13 + letters A–L; scored as seconds to completion | https://ubccpd.ca/media/2762/download |
| Adult average, Trail A / Trail B | ≈29 s (most within 90 s) / ≈75 s (most within 3 min); “deficient” above 78 s / 273 s (Corrigan & Hinkeldey 1987; Gaudino et al. 1995) | https://ubccpd.ca/media/2762/download |
| Reitan & Wolfson (1993) classification | Trail A: normal 0–39 s, mild 40–51 s, moderate–severe ≥52 s. Trail B: normal 0–85 s, mild 86–120 s, moderate–severe ≥121 s | https://www.nu.edu.ye/wp-content/uploads/2014/05/Groth-Marnat-The-Handbook-of-Psychological-Assessment.pdf |
| Age & education norms (Tombaugh 2004) | n = 911, ages 18–89: steady slowing with age and lower education; norms stratified into 11 age groups × 2 education levels | https://www.sciencedirect.com/science/article/pii/S0887617703000398 |
| Age-stratified Trail A screening cuts (16th percentile, from Tombaugh 2004) | Ages 18–24: 29 s · 25–34: 36 s · 35–44: 41 s · 45–54: 44 s | https://host.credim.u-bordeaux.fr/dnn-memento/Portals/0/Chercheurs/DOC/MEMENTO_Part4b_Instructions_Tests_Neuropsychologiques_Echelles.pdf |
| What Part A vs Part B measure | Part A: visual search, attention, processing speed. Part B adds task switching, working memory, cognitive flexibility | https://www.cambridge.org/core/journals/journal-of-the-international-neuropsychological-society/article/construct-validity-of-the-trail-making-test-role-of-taskswitching-working-memory-inhibitioninterference-control-and-visuomotor-abilities/453A48DA4B70F8BE793106A52BC0FE89 · https://doi.org/10.1016/j.intell.2011.03.001 |
| Test history | “Partington’s Pathways” (1938) → US Army Individual Test Battery (1944) → standardized for clinical use by Reitan (1950s) | https://www.millisecond.com/library/trailmakingtest |
| Times on this site (8/12/16 nodes, Dynamic mode) | No published norm exists for short online boards or moving-node versions | SchulteLab user aggregate (beta — sample still accumulating) |
Honesty note: the values above belong to the 25-circle paper-and-pencil test administered by a clinician. This site’s boards use 8–16 nodes and mouse/touch input, so the game scales these anchors by node count and marks every derived band with * — a derivation for orientation, never a clinical screen. If a time concerns you, consult a professional; this page is information, not a diagnosis.
Take or print the test
Measure your own time on the online Part A / Part B / Dynamic boards, or print fresh paper cards in the original format. Open the printable trail making cards →
Frequently asked questions
What is a good trail making test score?
On the 25-circle paper test, published clinical summaries put the adult average near 29 seconds for Part A and 75 seconds for Part B, and Reitan & Wolfson call up to 39 s (A) and 85 s (B) normal (sources in the table above). Faster is better, but a single slow round proves nothing — fatigue, distractions and device all move the number. Judge your trend, not one attempt.
How do age and education affect trail making times?
Both matter a lot. In the Tombaugh 2004 normative study of 911 adults aged 18–89, times rose steadily with age and were faster with more education, which is why the published norms are stratified into 11 age groups and 2 education levels. Always compare yourself with your own age band — and over time, with your own history.
Why is my online time so much faster than the paper norms?
Because the formats differ: this site uses 8–16 nodes instead of 25 circles, and tapping is faster than drawing pencil lines. Studies of digitized versions confirm they are systematically different from the paper original. The game therefore scales the paper anchors by node count and marks them as derivations — compare online times only with your own online history.
Is the trail making test a diagnosis?
No. It is a screening snapshot: clinicians interpret it together with age- and education-corrected norms, other tests and a full assessment. An online version with fewer nodes is training and self-tracking only. If you are worried about your cognition, bring your concerns — not just your seconds — to a qualified professional.