Descriptive Statistics in SPSS: How to Run and Interpret the Results
When to use descriptive statistics
You use descriptive statistics at the beginning of a quantitative analysis to answer two basic questions: who is in your survey sample, and what do your variables look like. For demographic variables such as gender, age, occupation, or income, you usually need frequencies and percentages. For quantitative variables or variables measured with a Likert scale, you usually examine the number of valid observations, minimum, maximum, mean, and standard deviation.
Descriptive statistics do not test causal hypotheses. They do not tell you whether X affects Y or whether a regression model is significant. Their role is to help you check the data, describe the sample, and identify unusual values before running Cronbach's Alpha, EFA, regression, or PLS-SEM. You can also see related topics in the SPSS category, especially if you are unsure where this step belongs in the analysis process.
For a study with a hypothesis such as “characteristic X affects Y,” descriptive statistics are usually presented before the scale and model assessment. The mean of each item can show the response tendency, but you need to interpret it carefully. A high mean describes the sample's average level of agreement. It does not, by itself, prove that X causes Y to increase.
Preparing the data before running the analysis
Before opening the Frequencies or Descriptives dialog box, check your data file. Many difficult-to-read output tables begin with incorrectly coded variables, commas used instead of decimal points, or categorical variables stored as strings when SPSS requires numeric values.
First, open Variable View and check the Name, Type, Label, Values, Missing, and Measure columns. Variable names should be short and contain no spaces, such as GT1, TUOI, or HL3. The variable label should describe the content clearly. For gender, you might code 1 as male and 2 as female, then define both codes under Values. For a Likert scale, keep the coding direction consistent, such as 1 for strongly disagree and 5 for strongly agree.
If the questionnaire contains a reverse-coded item, deal with it before calculating the mean score for the construct. A reverse-coded item on a scale from 1 to 5 is commonly recoded using the formula new value equals 6 minus the old value. Save the recoded item under a separate name, record what you did, and do not overwrite the original data until you are certain the procedure is correct.
Next, check missing values. A blank cell, code 99, or code 999 may be treated by SPSS as a valid numeric value if you have not defined it as missing. Go to Variable View > Missing to define the appropriate missing code, or clean the data in Data View according to the rules specified in your research method.
Finally, identify the variable type. Gender, age group, and occupation are suitable for a frequency table. Age in years, mean scores, or total scale scores are suitable for Descriptives or Explore. If you are describing a scale, decide clearly whether you will report each item or a representative score for each construct.
Steps for running descriptive statistics in SPSS
Step 1: Run frequencies for categorical variables
Go to Analyze > Descriptive Statistics > Frequencies. Move variables such as gender, age group, education level, or occupation into Variable(s). Keep Display frequency tables selected if you want SPSS to print the frequency table.
Click Charts if you need a chart. For categorical variables, Bar charts are usually easier to read than pie charts when there are several groups. Select Continue, then click OK to run the analysis. In the output, you will see the columns Frequency, Percent, Valid Percent, and Cumulative Percent.
Step 2: Select statistics for quantitative variables
Go to Analyze > Descriptive Statistics > Descriptives. Move the variables you want to describe into Variable(s), then click Options. Select Mean, Std. deviation, Minimum, and Maximum. You can also select Sum if the study genuinely requires a total score, but this statistic is not commonly used to describe a Likert scale.
Click Continue, then OK. SPSS creates a Descriptive Statistics table with the columns N, Minimum, Maximum, Mean, and Std. Deviation. N is the number of valid observations used to calculate the statistics. It does not necessarily equal the total number of rows in the file.
Step 3: Examine the median and percentiles when the data are skewed
If you suspect outliers or a skewed distribution, go to Analyze > Descriptive Statistics > Explore. Move the variable you want to check into Dependent List, click Statistics, and keep Descriptives selected. Under Plots, you can select Boxplots and Histogram.
Explore provides the median, the confidence interval for the mean, and information about extreme values. If you need to understand more clearly how the median differs from the mean, see the article on the median in statistics. Do not automatically remove extreme values simply because they change the mean. Go back to the original questionnaire and your data-cleaning rules to determine whether a value is an entry error or a valid response.
Step 4: Save and check the syntax
In the output window, you can right-click a table and copy it into Word. A better approach is to open File > New > Syntax before running the analysis and paste the command from the dialog box using Paste instead of only clicking OK. Syntax allows you to repeat the same procedure and explain exactly which variables you ran.
After running the analysis, check the number of variables and observations in the output. If a table contains no result, check whether the variable is stored as a string. If N differs sharply between variables, check missing data and exclusion codes.
Reading the output
The Frequencies table is used for categorical variables. Frequency is the number of people in each group. Percent is calculated using all rows, including missing cases if SPSS recognises them as part of the data. Valid Percent is calculated using valid observations. When data are missing, use Valid Percent to describe the sample structure and state which denominator you used.
Cumulative Percent is the accumulated percentage and is more appropriate for ordered variables such as age groups or income levels. For gender, this column usually has little analytical value. Do not add rounded percentages and conclude that SPSS calculated them incorrectly, because the displayed total may differ slightly as a result of decimal rounding.
Read the Descriptive Statistics table row by row. N shows the number of valid observations, Minimum and Maximum show the actual value range in the file, Mean is the average, and Std. Deviation is the standard deviation. A larger standard deviation means that responses are more spread out around the mean in the sample, but there is no single value that applies to every study. You can read more about interpretation in the articles on standard deviation and variance.
The table below is illustrative output, not the result of a real study. The column names are kept as SPSS displays them so you can compare them with the file currently open.
| Descriptive Statistics | N | Minimum | Maximum | Mean | Std. Deviation |
|---|---|---|---|---|---|
| HL1 | 182 | 1 | 5 | 3.74 | 0.812 |
| HL2 | 182 | 1 | 5 | 3.61 | 0.901 |
| HL3 | 180 | 1 | 5 | 3.48 | 0.976 |
| HL4 | 182 | 1 | 5 | 3.82 | 0.765 |
| Valid N (listwise) | 180 |
From this illustrative table, you can say that HL4 has the highest mean among the four items, while HL3 has fewer valid observations than the other cases. You should not call HL4 the “best” item until you have identified the scale and clarified the construct meaning. You should also avoid concluding that HL3 performs poorly simply because its mean is lower, unless that difference is relevant to the research objective.
If you need to check variance, remember that variance is the square of the standard deviation when calculated using the same approach. The calculation and interpretation are explained more clearly in the article on the variance formula. When a boxplot shows an outlier, read about outliers in data before deciding whether to retain or remove the observation.
Evaluation criteria
Descriptive statistics do not have a pass threshold in the same way as Cronbach's Alpha or KMO. The question “how good should descriptive statistics be” is usually framed incorrectly. A Mean of 4 does not automatically make the data good, and a low Std. Deviation does not prove that a scale is reliable. Evaluate the output by checking whether the value range is valid, whether N is complete, whether the selected table matches the variable type, and whether the results are consistent with the questionnaire design.
| What to check | How to evaluate it in the SPSS file | Basis for the assessment |
|---|---|---|
| Likert scale | Minimum and Maximum fall within the coding range that was designed, such as 1 to 5 | Questionnaire coding convention |
| Number of valid observations | N matches the number of responses after cleaning, with an explanation if values are missing | Study data-cleaning procedure |
| Frequencies and percentages | No groups are missing, and percentages use the appropriate denominator | Study reporting rule |
| Mean | Used to describe the average tendency of a variable or scale, not used alone to test a hypothesis | Descriptive objective of the study |
| Std. Deviation | Used to describe dispersion and read together with Mean, Minimum, and Maximum | Descriptive objective of the study |
| Outlier | Check the original response, response logic, and data-cleaning rule before processing it | Data-checking procedure |
For a Likert scale, you can report the mean and standard deviation if this reporting approach has been established consistently in the study. Interpretation bands such as 1 to 1.8 for a low level or 4.2 to 5 for a high level should only be used when you have defined how the intervals are divided and have a clear methodological reason. Do not call a value acceptable simply because it exceeds a cutoff mentioned in an online guide.
How to write this in your thesis
Write in this order: introduce the table, state the sample size, describe the main groups, then comment on the Mean and Std. Deviation of the variables that matter. You do not need to copy the entire SPSS output into your thesis. Present a reformatted table, keep the variable names or replace them with clearer labels, and explain the numbers that relate to the research question.
You can adapt the following paragraph: “The descriptive statistics show that the study sample consists of [N] valid observations. The mean values of [variable group name] range from [lowest Mean] to [highest Mean], while the standard deviations range from [lowest SD] to [highest SD]. [Variable name] has the highest mean of [Mean], whereas [Variable name] has the lowest value of [Mean].”
For demographic variables, you can write: “Among [N] valid observations, [group name] accounts for the largest proportion at [Valid Percent]%, followed by [group name] at [Valid Percent]%.” If data are missing, add: “Some variables have a lower N than the total sample because of missing responses. The percentages in the table are calculated using valid observations.”
Do not write that descriptive statistics prove that hypothesis H1 is accepted. Accepting or rejecting a hypothesis requires an appropriate test such as regression, t-test, ANOVA, or a structural model. Descriptive statistics provide context so the reader can understand the data before reviewing those results.
Common mistakes when running descriptive statistics
The first mistake is selecting every data column and running Frequencies. The output becomes long and difficult to control, and it may include respondent identification codes in the table. Select variables according to your purpose, separate categorical variables from scale variables, and label each table in the analysis file.
The second mistake is reading Percent instead of Valid Percent when data contain missing values. The two columns may be identical when there are no missing cases, but they differ when some respondents leave an item blank. Identify the denominator before writing a percentage in Chapter 4.
The third mistake is treating the Mean as evidence of scale quality. Mean and Std. Deviation do not replace Cronbach's Alpha, EFA, or validity testing. If an item has a low Mean, check the item wording, coding direction, and data distribution before removing it.
The fourth mistake is deleting an outlier directly in Data View. A value of 1 or 5 on a Likert scale is not an outlier simply because it is at the boundary. If you have concerns, check the questionnaire, completion time, repeated response patterns, and the cleaning rules stated in the method section.
The final mistake is copying an entire SPSS table with too many decimal places. You can format the table in Word, use a consistent number of decimal places, and still keep the original output file for comparison when your supervisor asks about the result.
Frequently asked questions
What are descriptive statistics?
Descriptive statistics are techniques used to summarise data with frequencies, percentages, means, medians, standard deviations, minimum values, and maximum values. In SPSS, you usually run Frequencies, Descriptives, or Explore depending on the variable type and reporting objective.
How good should descriptive statistics be?
There is no fixed Mean or Std. Deviation that proves descriptive statistics are good. Check the coding range, number of valid observations, missing data, outliers, and whether the selected statistic fits the variable type.
Where do I run mean descriptive statistics in SPSS?
Go to Analyze > Descriptive Statistics > Descriptives, select the variables, click Options, select Mean, Std. deviation, Minimum, and Maximum, then click Continue and OK. The main results table is Descriptive Statistics.
When should I use frequency descriptive statistics in SPSS?
Use Analyze > Descriptive Statistics > Frequencies for gender, age group, occupation, education level, and other categorical variables. The Frequencies table shows the number and percentage of each group. Valid Percent is usually more appropriate when the data contain missing values.
Why has N decreased in the Descriptive Statistics table?
N decreases when one or more values for a variable are missing, defined as missing, or excluded through a data filter. Check Missing, Data > Select Cases, any active filter, and blank cells in Data View before running the analysis again.
Open the SPSS file again, separate categorical variables from scale variables, run Frequencies and Descriptives using the correct menu paths, and compare N with the number of valid responses before copying the table into Chapter 4. If you need to run the analysis on a .sav or .csv file and check the full workflow, M4 data analysis handles this step.