A z table in statistics is a reference chart that gives probabilities for values in the standard normal distribution, using a z-score to locate each value. It lets you translate a standardized score into an area under the normal curve, such as the proportion below a value, above it, or between two values.
What a Z Table Represents
The standard normal distribution is a bell-shaped probability distribution with a mean of 0 and a standard deviation of 1. A z-score tells how many standard deviations a value lies from the mean: positive scores are above the mean, and negative scores are below it. A z table lists probabilities associated with those standardized positions.
The areas in the table are probabilities, so they range from 0 to 1. They can also be read as proportions or converted to percentages: an area of 0.90 represents 90% of the distribution. On a normal-curve diagram, probability corresponds to area, not to the height of the curve at a single point. For a continuous distribution, the probability of one exact value is zero; useful probabilities describe ranges.
Tables are printed in different formats, so there is no single layout that every z table follows. The most common version gives the cumulative area to the left of a z-score, written as Φ(z). Other versions show the area between the mean and z, or the area in the right tail. Before looking up a number, check the table heading or its notes to identify which convention it uses.
How to Calculate a Z-Score
For an individual observation from a population with mean μ and standard deviation σ, the standardized score is z = (x − μ) / σ. Here, x is the observed value. Subtracting the mean measures the distance from the center; dividing by the standard deviation expresses that distance in standard-deviation units.
For example, suppose test scores are modeled with a mean of 68 and a standard deviation of 4. A score of 74 has z = (74 − 68) / 4 = 1.5. That score is 1.5 standard deviations above the mean. A score of 62 would have z = (62 − 68) / 4 = −1.5, the same distance below the mean.
In practice, you may have to estimate the mean and standard deviation from a sample. A standardized value based on sample estimates is still often called a z-score, but the correct inferential procedure depends on the question and assumptions. In particular, a test about a mean with an estimated population standard deviation commonly uses a t distribution rather than treating the standard deviation as known.
How to Read a Typical Z Table
In a cumulative-left table, the row usually gives the first two digits of z and the column gives the hundredths digit. To look up z = 1.23, find row 1.2 and column 0.03. Their intersection is approximately 0.8907, meaning about 89.07% of a standard normal distribution lies below 1.23.
- Standardize the observation or identify the z-score already provided.
- Separate the score into the row part and the column part. For 1.23, those are 1.2 and 0.03.
- Read the table entry, then interpret it according to the table’s stated convention.
- Use subtraction or symmetry if the question asks for a different region of the curve.
Most printed tables round to four decimal places, so small differences can arise from rounding or from using a calculator. If the score falls between listed values, a common classroom approach is to use the nearest entry. For more precision, use a calculator or statistical software that evaluates the standard normal cumulative distribution.
Finding Areas Below, Above, and Between Scores
If the table gives cumulative-left probabilities, the entry for z is the probability that a standard normal variable Z is at most z. In symbols, P(Z ≤ z) = Φ(z). For z = 1.23, that probability is 0.8907. The area to the right is the complement, 1 − 0.8907 = 0.1093.
Negative scores can be handled directly if the table includes negative rows. If it lists only positive z-scores, use the symmetry of the normal curve: Φ(−z) = 1 − Φ(z). Since Φ(1.23) = 0.8907, Φ(−1.23) = 0.1093. This makes sense because a point 1.23 standard deviations below the mean leaves a relatively small area to its left.
To find the probability between two z-scores, subtract the cumulative area at the lower score from the cumulative area at the higher score. For example, the area between z = −1 and z = 1 is approximately 0.8413 − 0.1587 = 0.6826, or about 68.26%. More generally, for a ≤ b, P(a < Z < b) = Φ(b) − Φ(a). Whether endpoints are written as inclusive or exclusive does not change the result for a continuous distribution.
A Worked Example in Original Units
Suppose a measurement follows a normal model with mean 68 and standard deviation 4, and you want the proportion of measurements below 74. First convert 74 to z = 1.5. A cumulative standard normal table gives Φ(1.50) = 0.9332. Under the model, about 93.32% of measurements are below 74, and about 6.68% are above it.
This is a probability statement about the model, not a guarantee about a particular sample. If there are 1,000 independent observations drawn from the stated distribution, 933 below 74 would be a reasonable expected count, but the actual count will vary. The model also matters: if the measurements are strongly skewed or the stated mean and standard deviation are inaccurate, the normal-table estimate may not describe the real data well.
Using the Table Backward: Percentiles and Cutoffs
A z table can also help find a cutoff for a target percentile. Instead of starting with z and reading an area, start with an area and locate the closest table entry, then read the corresponding z-score. For example, the 90th percentile has cumulative area 0.9000 and z is approximately 1.28. A value at that percentile is about 1.28 standard deviations above the mean in a standard normal model.
For a central 95% interval, 2.5% lies in each tail. The cumulative area at the upper cutoff is 0.975, corresponding to z approximately 1.96; by symmetry the lower cutoff is −1.96. If the original variable has mean μ and standard deviation σ, the corresponding cutoffs are μ − 1.96σ and μ + 1.96σ. This describes the central 95% of a normal distribution; it is not automatically a confidence interval for every statistical problem.
Where Z Tables Are Used
Z tables appear in introductory statistics because they connect standardized values with probabilities. Common uses include:
- Finding the percentile rank of an observation under a normal model.
- Calculating a tail probability for a z-test or a standard normal approximation.
- Locating critical values for specified tail areas, such as those used in some hypothesis tests.
- Explaining how standard scores compare observations measured on different scales.
For instance, a z-score can compare a height relative to one population with a test score relative to another, provided each score is standardized using the relevant mean and standard deviation. The raw units differ, but both z-scores express distance in standard-deviation units. The comparison does not make the underlying measurements equivalent; it only puts their relative positions on a common scale.
Choosing Between a Z Table and a T Table
A z table describes the standard normal distribution. A t table describes a family of Student’s t distributions, each associated with degrees of freedom. The t distribution has heavier tails, especially when degrees of freedom are small, and approaches the standard normal distribution as degrees of freedom increase.
For inference about a population mean, a z procedure is typically appropriate when the population standard deviation is known and the model conditions are met. When the population standard deviation is unknown and must be estimated from the sample, a t procedure is generally used, particularly for small samples from an approximately normal population. In large samples the results may be close, but choosing a z table solely because the sample seems large can overlook the assumptions of the particular method. Follow the procedure specified by the statistical question or course, and check what quantity is known.
Common Mistakes and Misreadings
- Assuming every table reports the same area. A value may be cumulative-left, mean-to-z, or right-tail area. Identify the convention before interpreting it.
- Using the wrong tail. “Greater than” asks for a right-tail probability; a cumulative-left table requires subtracting its entry from 1.
- Ignoring negative signs. A negative z-score is below the mean. Treating it as positive can reverse the requested region.
- Confusing a z-score with a probability. A z-score such as 1.5 is a position; 0.9332 is the cumulative probability associated with it.
- Assuming all real data are normal. The z table gives exact areas for the standard normal model, not for any arbitrary set of observations.
A useful check is to ask whether the answer fits the geometry. A positive z-score should have more than half the distribution to its left; a negative one should have less than half. A right-tail probability and its corresponding left-tail probability should add to 1. If they do not, revisit the table convention and the subtraction.
Frequently Asked Questions
What does a z table tell you?
A z table gives an area, or probability, associated with a z-score under the standard normal distribution. Depending on the table, that area may be to the left of the score, to the right, or between the score and the mean.
What is the z-score for the mean?
The mean has z = 0 because its distance from itself is zero. In a standard normal distribution, half the area is below zero and half is above it.
How do I find the probability of a z-score greater than a value?
With a cumulative-left table, look up the score and subtract the entry from 1. For example, if the left area is 0.9332, the area above the score is 0.0668.
Is a z table the same as a normal distribution table?
Usually, yes: “z table” commonly means a table of probabilities for the standard normal distribution. Still, check the headings because different normal tables can report different regions.
When should I use a t table instead?
For many tests or intervals about a mean when the population standard deviation is unknown, use the t distribution and its degrees of freedom rather than automatically using a z table. The right method depends on the assumptions and design of the analysis.
Related
- Encrypt an Excel Spreadsheet: Steps and Security Tips
- Employee Time Tracking Template Excel: Setup and Tips
- X-Arcade Programming Software: What You Need to Know
Try it: the Date Difference Calculator: days, months, years and weekdays between two dates, with the DATEDIF formulas.
More in How-to basics
- Excel Import CSV with Line Breaks Without Splitting RowsTo import a CSV with line breaks inside a field without splitting its record across Excel rows, use…
- Excel Not Responding When Copying and Pasting? FixesIf Excel stops responding when you copy and paste, first wait a minute without clicking repeatedly; a large…
- How to Import CSV to Excel: Steps and Common FixesTo import a CSV into Excel for Microsoft 365 on Windows, open a workbook, select Data > From…
- How to Do Subscript in Excel With a ShortcutTo subscript text in Excel for Microsoft 365 on Windows, select the character or characters you want to…
- How to Find the Highest Average in ExcelTo find the highest average in Excel when you have several groups, calculate an average for each group…
- Excel Day of Year: Find the Day Number for Any DateTo find the day of the year in Excel, put a real Excel date in A2 and enter…
Sources: Function behaviour and menu paths: Microsoft Support documentation for Excel. Method & sources
Spotted an error in this guide? Let us know.