Overview

A raw score is the original result before any scaling, such as 74 marks on a test. This shows how that score compares with the group, as a z-score and a percentile rank.

Enter the score, the group's mean and standard deviation. The target percentile field isn't used in this result.

Worked examples

InputsResult
Observed Value74Distribution Mean62Distribution Standard Deviation9Target Percentile (%)90Raw score z-standardization: z=1.33Percentile rank: 90.88%
Show 2 more examples
Observed Value85Distribution Mean70Distribution Standard Deviation10Target Percentile (%)90Raw score z-standardization: z=1.5Percentile rank: 93.32%
Observed Value45Distribution Mean50Distribution Standard Deviation5Target Percentile (%)90Raw score z-standardization: z=-1Percentile rank: 15.87%

Raw score z-standardization by observed value

With the calculator's starting values, the raw score z-standardization is 1.33. The table shows how it changes with the observed value.

Observed ValueRaw score z-standardization
40-2.44
50-1.33
60-0.22
700.89
74starting value1.33
802
903.11
1004.22
1105.33

Other inputs: Distribution Mean: 62 · Distribution Standard Deviation: 9 · Target Percentile: 90%. Change them in the calculator above for your own numbers.

How it works

The z-score is how many standard deviations the score is above or below the mean. The percentile rank comes from the normal distribution and is the share of the group expected to score lower. It assumes scores are roughly bell-shaped.

The formula

z = (raw score − mean) ÷ standard deviation
meanThe group's average score
standard deviationHow spread out the scores are

Practical tips

  • 74 with a mean of 62 and a standard deviation of 9 is a z of 1.33, around the 91st percentile.
  • To go back from a z-score to a raw score, use mean + z × standard deviation.
  • If scores bunch near the top, as on an easy test, the percentile will be less reliable.

Frequently asked questions

What's the difference between a raw score and a scaled score?

A scaled score is the raw score converted onto a fixed scale, often using its z-score, so different test versions can be compared.

Is percentile rank the same as percentage correct?

No. 80% correct could be the 50th percentile if most people did as well.

Can the percentile be 100?

Not quite. Even a very high score has a tiny share of the curve above it.