Z Score of -0.9: Meaning, Percentile, and Examples

Home » Blog » Z Score of -0.9: Meaning, Percentile, and Examples

A z score of -0.9 means a value is 0.9 standard deviations below the mean of its reference distribution. If the distribution is approximately normal, that position is around the 18th percentile: roughly 18% of observations fall below it and 82% above it.

What a z score measures

A z score expresses how far a value lies from a distribution’s mean in standard-deviation units. It is a standardized way to describe relative position, not a measurement with a fixed unit such as inches, points, or kilograms.

The usual calculation is:

z = (x − μ) / σ

  • x is the observation or score.
  • μ is the mean of the reference group or distribution.
  • σ is its standard deviation.

A negative z score indicates a value below the mean; a positive one indicates a value above it. A z score of zero is exactly at the mean. The size of the score tells how many standard deviations separate the observation from that mean.

How to interpret -0.9

For a z score of -0.9, the observed value is below average, but less than one standard deviation below average. It is not, by itself, evidence that something is wrong or unusually low. Whether it matters depends on what is being measured, the reference population, and the purpose of the comparison.

For example, if a class’s test scores have a mean of 70 and a standard deviation of 10, a score with z = -0.9 would be:

x = 70 + (-0.9 × 10) = 61

That score is 9 points below the class mean. In a different class, the same raw score could have a different z score because its mean and standard deviation might differ.

When a normal distribution is a reasonable model, the standard normal cumulative probability at -0.9 is about 0.184. This corresponds to approximately the 18th percentile. Percentiles describe the share at or below a position; they do not mean the value is 18% below the mean.

Percentile and normal-distribution assumptions

The 18th-percentile interpretation relies on the normal distribution, the familiar bell-shaped curve. In that model, about 18.4% of values are below z = -0.9 and about 81.6% are above it. Tables and calculators may round these figures slightly differently.

A z score can still be calculated for data that are not normally distributed, but it does not automatically translate to that percentile. In a skewed distribution, for instance, a value 0.9 standard deviations below the mean may have a different rank than the normal curve predicts. Outliers can also pull the mean and inflate the standard deviation, changing the interpretation.

So, distinguish two claims: “0.9 standard deviations below the mean” follows from the z score’s definition; “about the 18th percentile” is an estimate that follows if the normal model fits adequately.

How to calculate or check the score

If you know the observation, mean, and standard deviation, subtract the mean from the observation and divide by the standard deviation. Keep the subtraction in the stated order: observation minus mean. Reversing it changes the sign.

  1. Identify the value and the correct reference group.
  2. Find that group’s mean and standard deviation.
  3. Calculate the difference between the value and the mean.
  4. Divide the difference by the standard deviation.

Suppose a measurement is 46, the reference mean is 55, and the standard deviation is 10. Then (46 − 55) / 10 = -0.9. If the standard deviation is zero, the formula cannot be used: every reference value is identical, so there is no spread by which to standardize.

Some settings use a sample standard deviation, while others use a population standard deviation. The choice can slightly change the result, especially with small samples. Use the method specified by the source or analysis rather than mixing conventions.

What it does—and does not—tell you

A z score describes a position relative to a particular distribution. It does not reveal the practical importance of a difference, the cause of the result, or whether an individual needs treatment or intervention. A small shift may matter greatly in one context and very little in another.

It is also not a probability that the observation is “bad,” nor does -0.9 mean a 90% chance of being below average. The score is a standardized distance. A percentile or probability requires a distribution model and a clearly defined direction, such as the share below the score.

Always check the reference group. A score compared with one age group, school, laboratory population, or test norm may not be comparable to a score based on another. A change in the reference mean or standard deviation can alter the z score even when the raw value stays the same.

Comparing z scores with other scores

Unlike a raw score, a z score removes the original scale and expresses distance in standard deviations. This can help compare positions across measurements with different units, provided each score was standardized against an appropriate distribution. It does not make unlike tests equivalent in content or quality.

A z score also differs from a percentile. The z score reports standardized distance; the percentile reports relative rank. Under a normal model, z = -0.9 is near the 18th percentile, but the conversion is not universal for every dataset.

Some standardized tests report a transformed score, such as a scale with a chosen mean and standard deviation, rather than the z score itself. Such scores may be derived from z scores, but their units and interpretation depend on the test’s scoring rules.

Frequently Asked Questions

Is a z score of -0.9 good or bad?
It is below the reference mean by 0.9 standard deviations. Whether that is good, bad, or important depends on the measure and the context; the number alone does not make that judgment.

What percentile is a z score of -0.9?
For a standard normal distribution, it is approximately the 18th percentile. In data that are not normally distributed, the actual percentile can differ.

Is -0.9 considered unusual?
Usually it is not far from the mean in standard-deviation terms. Common rules for flagging unusually distant values often focus on scores around two or three standard deviations from the mean, but appropriate thresholds depend on the field and the distribution.

Does a negative z score mean below average?
Yes. A negative score means the observation is below the reference mean. The minus sign does not indicate a negative raw measurement.

How many standard deviations below the mean is -0.9?
It is 0.9 standard deviations below the mean. The standard deviation is the unit used to express that distance.

Related

Scroll to Top