A z score of 1.67 means a value is 1.67 standard deviations above the mean of its reference distribution. If the distribution is standard normal, the value is approximately at the 95.25th percentile: about 95.25% of observations fall below it and 4.75% fall above it.
What a z score tells you
A z score, also called a standard score, expresses how far a measurement lies from the mean in units of standard deviation. It makes values from different scales easier to compare, provided the comparison uses appropriate means and standard deviations.
The basic formula is:
z = (x − μ) / σ
- x is the observed value.
- μ is the population mean.
- σ is the population standard deviation.
A positive z score is above the mean; a negative one is below it. A score of zero is exactly at the mean. For example, if test scores have a mean of 70 and a standard deviation of 10, a score of 86 has z = (86 − 70) / 10 = 1.6. A score with z = 1.67 in that same distribution would be 16.7 points above the mean, or 86.7.
How unusual is a z score of 1.67?
For a standard normal distribution, the cumulative probability at z = 1.67 is approximately 0.9525. This is often written as Φ(1.67) ≈ 0.9525. It means a randomly selected value from that model has about a 95.25% chance of being at or below 1.67, and about a 4.75% chance of being above it.
Whether that is “unusual” depends on the question and context. A value above roughly 95% of a normal distribution is relatively high, but it is not extraordinarily rare. The corresponding upper-tail probability is about 0.0475, or 4.75%.
For a two-sided question—asking how often a value is at least this far from the mean in either direction—the probability is approximately 2 × 0.0475 = 0.095, or 9.5%. This calculation includes values at z ≤ −1.67 as well as z ≥ 1.67. It is not the same as the one-sided probability.
Percentile and probability interpretations
Under the normal model, z = 1.67 corresponds to about the 95th percentile. A percentile describes relative rank: approximately 95.25% of the modeled values are lower, while approximately 4.75% are higher. It does not mean the observation is 95.25% of the mean, nor does it tell you the size of the measurement in its original units.
For a concrete example, suppose adult heights in a particular population were modeled as normally distributed with a mean of 170 cm and a standard deviation of 8 cm. A height with z = 1.67 would be:
x = 170 + (1.67 × 8) = 183.36 cm
The percentile interpretation is meaningful only if the chosen mean and standard deviation describe the relevant population or reference group. The example is an illustration of the calculation, not a claim that all populations have those height statistics.
Finding the probability from a z table or calculator
A standard normal table usually reports the area to the left of a z score. Locate 1.67 in the table; the left-tail probability is about 0.9525. To find the area to the right, subtract that value from 1: 1 − 0.9525 = 0.0475.
Some tables use a different convention, such as showing the area between zero and the z score. Check the table heading before reading a result. A calculator or statistics program may label the left-tail result as a cumulative normal probability or normal CDF.
Because tables round values, answers can differ slightly in the final decimal place. For most practical uses, reporting 95.25% below and 4.75% above is suitably precise. More decimal places do not make an interpretation more reliable if the data or model assumptions are uncertain.
When the interpretation applies—and when it does not
The percentile and tail probabilities above rely on a standard normal distribution, or on data that are reasonably well modeled by a normal distribution after standardization. A z score can still be calculated for non-normal data, but the normal table’s probability may then be misleading. Skew, heavy tails, outliers, or a mixture of different groups can change the actual proportion above a given score.
There is also an important distinction between a z score for an individual observation and a z statistic for a sample mean. In hypothesis testing, a statistic of 1.67 may be compared with a reference distribution to assess evidence against a null hypothesis. Its p-value depends on whether the test is one-sided or two-sided and on the test’s assumptions. A p-value is not the probability that the null hypothesis is true, and a z score alone does not establish practical importance or causation.
Finally, a z score is only as sound as its reference values. If a standard deviation is zero, the formula is undefined. If a sample rather than an entire population provides the mean and standard deviation, the calculation may use sample-based estimates; in small samples, a t distribution may be more appropriate for inference.
Frequently Asked Questions
Is a z score of 1.67 good or bad?
It is above the mean by 1.67 standard deviations. Whether that is good depends on what is measured: it could be a high exam score, but a high score on a measure of risk or symptoms may be undesirable.
What percentile is z = 1.67?
Assuming a standard normal distribution, it is approximately the 95.25th percentile. In rounded terms, the value is around the 95th percentile.
What is the probability of getting a z score above 1.67?
The upper-tail probability is about 0.0475, or 4.75%, under the standard normal model.
What is the two-tailed probability for z = 1.67?
It is about 0.095, or 9.5%, when “two-tailed” means values at least 1.67 standard deviations from zero in either direction. The precise interpretation depends on the test being performed.
How do I convert z = 1.67 back to an original value?
Use x = μ + zσ. Multiply 1.67 by the relevant standard deviation, then add the mean. You need the correct mean and standard deviation for the population or dataset.
Does z = 1.67 always mean the 95th percentile?
No. That percentile follows from the standard normal model. For a non-normal dataset, the z score tells you the standardized distance from the mean, but its actual percentile must be determined from the data or an appropriate distribution.
