Written and maintained by CASRAI Editorial Board
Last updated
To calculate the median in SPSS, open Analyze > Descriptive Statistics > Frequencies, click Statistics, tick Median and run the procedure; the value appears in the Statistics table. A second route, Analyze > Descriptive Statistics > Explore, reports the median by default and can repeat it for every group in your data. This guide covers both routes, explains when each is the right one, and shows how to read the result.
The two routes are not interchangeable. Frequencies is the quick way to get one median (or one per variable), Explore is the way to get a median for each group, and the Descriptives procedure, despite being the most obvious menu item, does not offer the median at all. Most short tutorials show only the Frequencies route; the comparison below is what decides which one to open.
What it is used for in research
IBM’s documentation defines the median as “the value above and below which half of the cases fall, the 50th percentile,” and notes that it is not sensitive to outlying values. That is why researchers report it for skewed measures such as waiting times, counts and incomes, and alongside a spread measure such as the interquartile range. The reporting side is covered in median and interquartile range; this page is about getting the number out of SPSS.
A small example shows why the choice matters. For the five scores 3, 5, 7, 9 and 100, the middle value is 7 while the mean is 24.8. One extreme case drags the mean far from where most of the data sit and leaves the median untouched. When SPSS reports a mean well above (or below) the median, the distribution is probably skewed, and the median is usually the safer summary. With an even number of cases the median is the average of the two middle values: for 2, 4, 6 and 8 it is 5.
Install or access (official source only)
SPSS Statistics is an IBM product, and the menu names in this guide follow IBM’s documentation for SPSS Statistics 29.0.0 (checked 2026-10-29): the Frequencies Statistics page and the Explore Statistics page, both linked under Sources. Menu labels in other versions can differ slightly, so check the documentation for the version you run. If your institution does not provide SPSS, the free statistical software guide lists alternatives that can also produce a median.
Step-by-step: find the median in SPSS
Method 1: Frequencies (one median per variable)
- Choose Analyze > Descriptive Statistics > Frequencies.
- Move the numeric variable (or several) into the Variable(s) box.
- If the variable is continuous, untick Display frequency tables. Otherwise SPSS prints one row for every distinct value.
- Click Statistics. Under Central Tendency, tick Median. IBM lists the mean, median, mode and sum of all values as the available measures of central tendency.
- Click Continue, then OK. The median appears in the Statistics table in the Output Viewer.
The same Statistics dialog has a Percentile(s) box and a Quartiles option. IBM defines quartiles as the 25th, 50th and 75th percentiles, so the 50th percentile is the same number as the median. Ticking Quartiles gives you the median together with the 25th and 75th percentiles in one pass, which is the pair you need to report an interquartile range. See percentiles and quartiles for how those values are defined.
Method 2: Explore (a median for each group)
- Choose Analyze > Descriptive Statistics > Explore.
- Move the numeric variable into the dependent variable box.
- To get a separate median for each group, move the grouping variable (for example treatment arm or site) into the Factor List. Explore then repeats its output once per group.
- Click OK. In the Descriptives table, the median is listed with the mean and the 5% trimmed mean, because IBM states that these measures of central tendency and dispersion are displayed by default.
Explore’s Percentiles output, when requested, uses a default set of the 5th, 10th, 25th, 50th, 75th, 90th and 95th percentiles, and the 50th is again the median. Explore also supplies M-estimators, which IBM describes as robust alternatives to the sample mean and median for estimating the location of a distribution; they are a different statistic and not a replacement for the plain median. For a boxplot of the same groups, see box plots in SPSS.
Which route to use
| You need | Use | What the IBM documentation says |
|---|---|---|
| One median for each variable | Frequencies > Statistics > Median | Median is one of the central tendency options; the 50th percentile equals it |
| Median, 25th and 75th percentiles together | Frequencies > Statistics > Quartiles | Quartiles are the 25th, 50th and 75th percentiles |
| A median for each group | Explore with a Factor List | Median is in the default Descriptives output |
| Median of data already grouped into intervals | Frequencies > Statistics > Values are group midpoints | Estimates the median and percentiles for the original ungrouped data when your values are group midpoints |
For a wider comparison of the three Descriptive Statistics procedures, including why Descriptives has no median option, see descriptive statistics in SPSS.
Output and how to read it
With Frequencies, look for the table titled Statistics. Each variable you entered has its own column, with N (valid and missing) and the Median row you ticked. With Explore, the median is a row in the Descriptives table, one block per group when a Factor List was used.
- Compare it with the mean. If you also ticked Mean, a large gap between the mean and the median is a hint of skew or outliers. Check it with a histogram or boxplot before choosing a test.
- Report it with a spread measure. A median on its own says nothing about variability; pair it with the 25th and 75th percentiles from the same run.
- Check the N. The median is computed only on valid cases, so the N row tells you how many values it rests on.
Troubleshooting
There is no Median option in Descriptives
That is expected: the Descriptives dialog does not include the median. Use Frequencies or Explore instead.
Frequencies prints a huge table
Untick Display frequency tables in the main Frequencies dialog before you run it. The median stays in the Statistics table.
The median looks meaningless
Check the variable’s measurement level. A median of a nominal code (department number, group label) has no interpretation; the statistic is only meaningful for ordinal or scale variables. See choosing the right procedure by variable type.
Your data are intervals, not raw values
If each value in the data stands for a group, such as the midpoint of an age band, tick the group midpoints option in the Frequencies Statistics dialog so SPSS estimates the median for the original ungrouped data.
Free alternatives (table)
| Tool | What it is | Getting a median |
|---|---|---|
| R | A language and environment for statistical computing and graphics | After x <- c(2, 4, 6, 8), typing summary(x) prints the minimum, quartiles, median, mean and maximum |
| jamovi | A free and open statistical spreadsheet | Spreadsheet-style interface; powered by R |
| JASP | A free cross-platform program with a graphical user interface | Graphical interface offering classical and Bayesian procedures |
Installation sources and licences for each are in the free statistical software guide.
FAQ
How do I calculate the median in SPSS?
Go to Analyze > Descriptive Statistics > Frequencies, move your variable across, click Statistics, tick Median and run it. The result is in the Statistics table.
How do I find the median on SPSS for each group?
Use Analyze > Descriptive Statistics > Explore and put the grouping variable in the Factor List. Explore reports the median for every group in the Descriptives table.
Is the 50th percentile the same as the median?
Yes. IBM defines the median as the value above and below which half of the cases fall, the 50th percentile, and quartiles as the 25th, 50th and 75th percentiles.
Why does the median differ from the mean?
The mean uses every value, so extreme cases pull it; the median does not react to outlying values. A large gap suggests skew.
Sources
- IBM, SPSS Statistics 29.0.0 documentation, Frequencies Statistics: https://www.ibm.com/docs/en/spss-statistics/29.0.0?topic=frequencies-statistics (checked 2026-10-29).
- IBM, SPSS Statistics 29.0.0 documentation, Explore Statistics: https://www.ibm.com/docs/en/spss-statistics/29.0.0?topic=explore-statistics (checked 2026-10-29).
