Farnsworth-Munsell 100 Hue Test Scoring

A worked example: one cap's partial error score

Every cap is scored the same way: the distance to its left neighbour plus the distance to its right neighbour, minus 2. Cap 23 is used throughout; the neighbours change.

Farnsworth's score bands

The three bands printed on the original scoring sheet. They describe adults in their twenties and thirties best; use the age table for anyone else.

FM-100 Hue norms by age

Mean and 95th-percentile Total Error Score for 382 observers with normal colour vision, tested on the physical caps under standard lighting (Kinnear & Sahraie, 2002). The 95th percentile is the score that about one typical observer in twenty exceeds — a practical "upper limit of normal" for that age.

Transcribed from table 1 of the published paper. Ages 23–29 were not sampled in that study; the 20–22 and 30–39 rows bracket them.

Reading the error pattern

The total says how much; the pattern says what kind. The cap ranges are the approximate poles of each confusion axis on the 85-cap circle.

How is the Farnsworth-Munsell 100 Hue test scored?

Each of the 85 caps has a number that fixes its true place on the hue circle. After you arrange the four boxes, the score for one cap is the difference between its number and the number of the cap on its left, plus the difference to the cap on its right. A correctly placed cap sits between its true neighbours, so that sum is 1 + 1 = 2; Farnsworth's convention subtracts 2 so that a correctly placed cap scores 0. The result is that cap's partial error score, and the Total Error Score is the sum of all 85 partial scores.

This is why one badly placed cap costs so much. Slipping cap 23 in between caps 40 and 41 adds 33 to its own score and also raises the scores of caps 40 and 41, because their neighbours are now wrong too. The scoring is deliberately unforgiving of large transpositions and lenient about swapping two adjacent caps, which is the kind of error that lighting and screen quality produce.

Two conventions exist in software. Some programs report the raw sum without subtracting 2 per cap, so a perfect run reads 170 rather than 0; the X-Rite scoring software and this site both use the Farnsworth convention, where perfect is 0. If a score you were given is suspiciously high, check which convention was used before comparing it to any norm.

What is a good FM-100 Hue score?

Farnsworth's 1957 manual divided the population into three bands: a Total Error Score of 0–16 was called superior discrimination (about the top sixth of people), 16–100 average (about two-thirds) and over 100 low (about the bottom sixth). Those cut-offs are still printed on scoring sheets and are a reasonable first read for adults in their twenties and thirties.

They are a poor read for anyone else, because hue discrimination follows a U-shaped curve with age. Children improve every year through the teens, adults peak at about 19, and scores drift upward again from the forties as the lens yellows and the eye's optics change. A TES of 120 is ordinary for a 10-year-old and for someone in their sixties, and would be unusual for a 20-year-old. The age-norm table below is the fair comparison.

Why do clinicians use the square root of the TES?

Raw Total Error Scores are skewed: most people cluster near the low end while a long tail of high scores stretches to the right. Taking the square root pulls the distribution into a roughly normal shape, which makes means, standard deviations and percentiles behave, and lets an observer be placed against a norm in the usual way. Kinnear and Sahraie published their 2002 norms in both forms for that reason, and our result card reports √TES alongside the raw score.

A useful rule of thumb from the same data: a √TES of about 8.5 is the 95th percentile for adults aged 17 to 22, meaning roughly one typical young adult in twenty scores above it. Above that line the score is worth a second look — at the pattern, at the conditions, and if it persists, with an eye-care professional.

What are the red-green and blue-yellow partial error scores?

Because a colour vision deficiency produces errors in two opposite regions of the hue circle, the caps can be grouped by which confusion axis they lie on. Summing the partial error scores of the caps along the red-green axis gives a red-green partial score, and the same along the blue-yellow axis gives a blue-yellow score. Comparing the two is the simplest way to say whether a raised total is selective — pointing to a protan/deutan or a tritan pattern — or non-selective, which points to conditions or a general loss rather than a type.

The polar plot shows the same information visually. Even errors give a near-circular shape; a deficiency gives two bulges on opposite sides, and the axis through them estimates the type. On an uncalibrated screen that axis is an estimate, and the protan and deutan axes sit close enough together that mild cases are often a close call.

What does an online FM-100 score mean?

The scoring arithmetic online is exactly the printed test's, so a Total Error Score from this site is calculated the same way as one from a clinic. What differs is the stimulus: the physical caps are Munsell colour chips viewed under a standard daylight lamp, while a screen shows colours that vary with the display, its brightness setting, any night-shift filter, and the room. Those differences add errors that have nothing to do with your eyes, and they add them unevenly — blues suffer most.

Read an online score as relative rather than absolute. Our result card uses conservative bands (normal range below 80; graded above 200 and 400) that are wider than the printed test's, precisely because screens add noise. A high online score with a clear bipolar pattern is meaningful; a moderately high score with scattered errors usually means the conditions, not the observer. To attach a number to a licence, a job or a medical decision, take the physical test under standard lighting with an eye-care professional.

FM-100 Hue scoring — frequently asked questions

What is a normal Total Error Score on the 100 Hue test?

For a young adult, a Total Error Score up to about 76 is within the 95th percentile of the published norms, with a mean of about 44 at ages 18–21. Farnsworth's older bands call anything from 16 to 100 average. For children and for adults over 40 the normal range is higher — see the age table.

Is a lower FM-100 score better?

Yes. The score counts errors, so 0 is a perfect arrangement and every misplacement adds to it. A lower Total Error Score means finer hue discrimination.

What FM-100 score indicates color blindness?

No single total does. A colour vision deficiency shows as two opposite bulges on the polar plot — errors concentrated along one confusion axis — usually with a total well above the age norm. A high total with errors scattered evenly around the circle points to conditions, fatigue or a general loss rather than a type. The pattern decides; the total only says how much.

Does age affect the 100 Hue score?

Strongly. In Kinnear and Sahraie's norms the mean Total Error Score falls from about 364 at age 5 to 43 at age 19, then rises again through adulthood to about 150 in the seventies. Compare a score with the right age row rather than with the overall bands.

Why is my online score higher than expected?

Screens add error. An uncalibrated display, a night-shift filter, low brightness or a warm-lit room shift the colours of the caps, and blues in particular render poorly on many monitors. Retake it at full brightness with filters off in a neutral room; if the score stays high and the polar plot shows a clear axis, that is worth confirming with the physical test.

Does a very low score mean I am a tetrachromat?

No. A low score means excellent hue discrimination, which reflects attention, lighting and screen quality as much as biology. Tetrachromacy is a fourth cone class and cannot be demonstrated on any three-primary display, because the display cannot produce the pair of lights that would separate a tetrachromat from a trichromat.