Z-Scores in Psychology Tests and IQ Scores, Explained

Open almost any psychology test report and you'll find a small zoo of numbers: an IQ of 112, a T-score of 58, a percentile of 76. They look like three different measurements. They're not. Underneath nearly every score on that report sits one z-score, rescaled into friendlier outfits.

Once you see that, these reports stop being intimidating. This post is a translation guide. (General educational information only — interpreting a real report for a real person is a job for the psychologist who wrote it.)

IQ scores are z-scores wearing a costume

The most common IQ scales (the Wechsler tests are the best-known) are built on a mean of 100 and a standard deviation of 15. So converting any IQ to a z-score is one subtraction and one division:

z = (IQ − 100) / 15

Run the famous numbers:

• IQ 115 → z = (115 − 100) / 15 = +1.00 → 84.13th percentile. Better than about 84% of the norm group.
• IQ 130 → z = +2.00 → 97.72nd percentile. Only ~2.28% score higher, which is why 130 shows up so often around gifted programs.
• IQ 85 → z = −1.00 → 15.87th percentile. One SD below the mean — well inside the broad average range, despite how it sounds to anxious readers.
• IQ 70 → z = −2.00 → 2.28th percentile.

Reading the table the other way is just as useful: the 84.13th percentile always means z = +1 on any scale, whatever costume it's wearing. To check any conversion, the z score to percentile calculator takes two seconds.

One technical footnote worth knowing: not every test uses SD 15. A few older or alternative scales use 16 or even 24. If a report doesn't state the scale, the z-score math above may be off — another reason the interpretation belongs to the professional, not a web page.

T-scores: the clinical favorite

Many symptom inventories and personality questionnaires skip IQ-style numbers and report T-scores instead: mean 50, standard deviation 10, computed as T = 50 + 10z.

Why bother? Because T-scores have no negative numbers and no decimals, so a report saying "Anxiety: T = 68" feels less alarming than "z = +1.8" and reads cleanly for non-statisticians. The information is identical: T = 68 → z = 1.8 → about the 96th percentile, meaning the person's reported symptoms exceeded roughly 96% of the norm group. Many inventories flag the mid-60s and up as a range worth a closer look — but a high score on a questionnaire is a prompt for a professional conversation, not a diagnosis, and this post deliberately goes no further than the math.

Our full comparison in z-score vs T-score covers the conversions and the other meaning of "t" (the t-statistic) that students bump into in research methods classes.

Percentiles and stanines also appear

Percentiles show up because families understand them instantly: "scored better than 76% of same-age peers" needs no statistics vocabulary. Just remember a percentile is a rank, not a percentage of correct answers — a child at the 76th percentile did not get 76% on the test.

Stanines (1–9) appear in school testing reports. They're coarse z-score bands: stanine 5 covers the middle (z from −0.25 to +0.25, about 20% of students), and each step outward is half a standard deviation wide. Stanine 7, for instance, corresponds roughly to z between +0.75 and +1.25. Coarse on purpose — schools use them to avoid over-reading tiny score differences.

Putting it together: one child, three scores

Say a student's underlying performance on a reading test is z = +0.8. The same result could be reported as:

• Standard score (IQ-style): 100 + (0.8 × 15) = 112
• T-score: 50 + (0.8 × 10) = 58
• Percentile: the z score calculator gives the 78.81st percentile
• Stanine: 7 (z = 0.8 lands in the +0.75 to +1.25 band)

Four numbers, one performance. If two reports about the same student seem to disagree — 112 here, 58 there — convert both back to z before concluding anything. Nine times out of ten, they agree perfectly and were just wearing different clothes.

Frequently asked questions