METHOD GUIDE
Descriptive Statistics in SPSS
The Right Summary for Each Variable Type
The summary that fits depends on the type and shape of each variable, and a sample description is read closely by reviewers. Frequencies, Descriptives and Explore each print a different part of that picture, with denominators that do not always agree.
IN SHORT
What a Descriptive Summary Should Show for Each Variable
Descriptive statistics summarize the data observed in a sample, with counts and percentages for categories and a measure of center and spread for numeric variables, and they describe the sample without testing a hypothesis. The summary follows the variable type and the shape of its distribution: mean and standard deviation for roughly symmetric numeric data, median and interquartile range for skewed or ordinal data, and n (%) for categories. To get descriptive statistics in SPSS, open Analyze, Descriptive Statistics, then choose Frequencies for categorical variables, Descriptives for a brief numeric summary or Explore for numeric variables by group, and select the statistics you plan to report. OnlineSPSS, a specialist statistical consultancy for academic, clinical and professional research, carries out and reviews this analysis through an in-house PhD-level statistician.
ANALYSIS SUPPORT
Analysis Support for This Method
The summary table is where an analysis begins and where its reporting is first judged, which is where these two services come in.
Quantitative Data Analysis
The descriptive stage handled within the full analysis of your data, with variable types, distributions, missing data and the tests that follow agreed for your design.
Results Reporting & Interpretation
The sample description and results tables prepared in APA 7, with decimals, notation and table notes checked against your own output.
BEFORE YOU RUN IT
Settling Each Variable's Type and Summary First
One summary style applied to every variable is the commonest flaw in a sample description, so each variable's type and summary are decided first.
Measurement Type Nominal variables take counts and percentages, ordinal ones a median with quartiles or category percentages, and numeric ones a mean or a median with a spread. The type of each variable is listed before any table is requested.
Shape of the Distribution Skewed numeric variables, such as length of stay or income, are better summarized by the median and interquartile range, because a few extreme values move the mean and the standard deviation.
Groups to Describe Baseline tables usually describe the whole sample and then each group. Explore, or Split File, gives group summaries, and the grouping variable is fixed before the tables are built.
Gaps in Responses The n differs by variable when responses are missing. Each variable's valid n is reported, and the pattern of missing values is described before tests are run on incomplete variables.
IN SPSS
How to Run Descriptive Statistics in SPSS
Three procedures in IBM SPSS cover the sample description, and the order below starts from the variable settings they all rely on. Labels vary by version.
- Set each variable’s type and value labels in Variable View, and declare the missing-value codes.
- For categorical variables choose Analyze, Descriptive Statistics, Frequencies and keep Display frequency tables ticked.
- In Frequencies, open Statistics and request the mean, standard deviation, median, quartiles, minimum and maximum for numeric variables.
- For a compact numeric summary choose Descriptives and select the statistics to be reported.
- For numeric variables by group choose Explore, place the grouping variable under Factor List and request Descriptives under Statistics.
- Check the valid n of each variable in the output before copying any figure into your table.
READING THE OUTPUT
Reading the Frequencies, Descriptives and Explore Tables
SPSS Frequencies and Descriptives overlap in what they print but not in the denominators they use, so this table shows which output holds which figure and what to check before copying it.
| Output element | What it tells you | What to check |
|---|---|---|
| Frequencies: Statistics | The Valid and Missing N and the statistics requested for each variable | That the valid n of each variable is carried into the sample table |
| Frequency Table | Frequency, Percent, Valid Percent and Cumulative Percent for each category | That Valid Percent is used for n (%) when data are missing, and the denominator is stated |
| Descriptive Statistics (Descriptives) | The N, Minimum, Maximum, Mean and Std. Deviation of each numeric variable | That Valid N (listwise) at the bottom is not the n for each variable, and the minimum and maximum are checked for impossible values |
| Explore: Descriptives | By group, the Mean with its 95% Confidence Interval, Median, Std. Deviation, Interquartile Range, Skewness and Kurtosis | That Std. Error is not reported where the standard deviation is meant |
| Explore: Percentiles (if requested) | The percentiles under Weighted Average and under Tukey’s Hinges | That the definition used for the quartiles is consistent and stated |
| Skewness and Kurtosis with Std. Error | Numbers describing the shape of the distribution, each with its standard error | That they describe shape and are not used as a pass or fail test of any assumption |
| Explore: Extreme Values (if Outliers is ticked) | The five highest and lowest cases of each variable | That extremes are checked for entry errors before any are excluded, and the decision is reported |
| Explore: Case Processing Summary | The valid, missing and total cases for each group | That the group sizes appear next to the summary in the sample table |
REPORTING
Setting Out a Sample Description in APA 7
The sample table answers one question for each variable: which summary describes it honestly, and on how many cases.
Report numeric variables as M = ___ (SD = ___) when they are roughly symmetric, and as Mdn = ___ (IQR = ___) or the 25th and 75th percentiles when they are skewed, ordinal or bounded. Categories appear as n = ___ (___%), with the denominator named when data are missing. Keep the same number of decimals within a column and state the total N and any variable whose n differs. The standard error of the mean is not a stand-in for the standard deviation: it belongs with the confidence interval of the mean, which Explore prints. Present groups in columns beside an overall column. In a randomized trial, baseline differences are described and not tested, because chance alone produces them.
COMMON MISTAKES
Summary Choices That Misdescribe a Sample
Reviewers check whether the statistic matches the variable, and a mismatch is visible at a glance in a sample table.
Mean of a Category Code Averaging a code such as 1 and 2 for sex, or a diagnosis number, returns a figure without meaning. Nominal variables receive counts and percentages and nothing else.
Mean for Skewed Data Reporting M and SD for a long-tailed variable suggests a typical value that few participants have. The median with its IQR, or both with the shape stated, describes it more honestly.
SE Where SD Belongs The standard error describes the precision of a mean, not the spread of the data. In a sample table it makes the sample look far more homogeneous than it is.
Percent on the Wrong Base Percent in the Frequencies table includes missing cases, while Valid Percent does not. Quoting one while the text names the other gives figures that do not match the stated n.
Listwise N as the Sample Valid N (listwise) in the Descriptives table counts only cases complete on every variable listed. It is not the sample size of any one variable, and it is easily copied by mistake.
Impossible Values Left In A minimum age of 0 or a maximum of 999 usually marks a missing-value code or an entry error. The range is checked against what is possible before any summary is trusted.
Sample table already built? Independent Statistical Review checks the choice of summary, the denominators and the notation before you submit.
FREQUENTLY ASKED QUESTIONS
Questions About Descriptive Output in SPSS
Which procedure to open comes first, followed by the spread to report and a way to describe groups without leaving Split File on.
When should I use Frequencies, Descriptives or Explore in SPSS?
Use Frequencies for categorical variables, Descriptives for a compact numeric summary of the whole sample, and Explore when numeric variables must be summarized by group or with intervals. Frequencies also offers the median and quartiles for numeric variables, and Descriptives does not. Explore adds the confidence interval for the mean, the interquartile range, percentiles and extreme values, so it is the fuller choice for a variable you will report in detail.
Should I report the standard deviation or the standard error?
Report the standard deviation to describe the spread of your sample, and the standard error or a confidence interval only to describe the precision of an estimate. SPSS prints both beside the mean, which is why the wrong column is sometimes copied. In a sample description the standard deviation is expected, and the 95% confidence interval for the mean is added where precision matters.
Can I describe each group in one table without splitting the file?
Yes, Explore with the grouping variable under Factor List, or the Means procedure under Compare Means (Compare Means and Proportions in recent versions), gives group summaries without splitting the file. Split File also works, but it stays active until it is switched off, and later analyses then run separately by group without any prompt in the dialogs. If it is used, the Split File status is checked before every following analysis.
Can OnlineSPSS prepare and check my sample description table?
Yes. An in-house PhD-level statistician can check that each summary suits its variable, that the denominators and the valid n agree with the text, and that the notation follows APA 7, or prepare the tables from your own output. The descriptive stage can also be part of a full analysis of your data. The research question and the interpretation remain yours.
How do you protect my research data and confidentiality?
Your data and project details are handled confidentially, and project information is not published. Researcher names and institutions are not used publicly without permission. An NDA is available where required and can be arranged before sensitive materials are shared. Read more about confidentiality and data security.
What happens when I request an assessment, and how long does it take?
You share your research question, design and relevant materials, such as the dataset or the output you have, and a statistically trained Research Coordinator reviews the request and responds the same day. The analysis, deliverables, timeline and one fixed project fee are then confirmed before any work begins, so the timeline follows from the agreed scope.
Still have a question? Find more answers in our full FAQ.
NEXT STEP
Which Summary Fits
Each of Your Variables?
Send the variables with their types, the sample size and the tables SPSS printed. You hear back the same day, and the work on the sample table is confirmed before it starts.
No Commitment Required
Strictly Confidential
Fixed Project Fee
In-House PhD-Level Statisticians
Same-Day Response
Serving Researchers Since 2014