How to Interpret a Frequency Table in SPSS from A to Z

When to use a frequency table in SPSS

Use a frequency table when you want to describe a variable by its possible values or groups. For example, you may need to know how many men and women are in your sample, how many respondents belong to each age group, or how many people selected each point on a Likert scale. This is usually part of the opening of the results chapter, before you run Cronbach's Alpha, EFA, regression, or SmartPLS.

For a categorical variable, the research question usually asks which groups make up the sample and what percentage belongs to each group. For a discrete quantitative variable, such as years of experience or number of purchases, the frequency table shows how often each value occurs. When a variable has too many distinct values, consider grouping them into meaningful categories before interpreting the output.

A frequency table does not answer questions about causes or relationships between variables. It only describes the data you collected. A statement such as “women tend to purchase more because their income is lower” goes beyond what this table can demonstrate unless you have run an appropriate test.

If you need the full set of SPSS descriptive and testing topics, open the SPSS topic. A frequency table is often reported alongside a table of means, standard deviation, and descriptive charts.

Prepare the data before running the analysis

First, open the .sav file and inspect each variable in Variable View. The Name column should use short names without accents, such as GIOITINH, TUOI_NHOM, or HL1. Use the Label column for the full description so that you can identify the variable when reading the output. In the Values column, define the codes, such as 1 for Male and 2 for Female.

You also need to check Measure. Gender is usually Nominal, while an age group or education level with an order can be Ordinal. The average score of a scale is usually handled as Scale. The Measure setting does not change the data automatically, but it helps you choose more suitable charts and procedures.

Check coding and missing values

Look at Data View and check whether the response codes fall within the range used in the questionnaire. If a five-point Likert scale contains a value of 6, 99, or a text entry, determine whether it is a data-entry error or a missing-value code. Do not immediately treat 99 as a valid response.

In Variable View, use Missing to declare a missing code such as 99. SPSS will then exclude that code from some valid calculations, while the code remains in the file. Record your data-handling rule in the working file so that you can later explain why the total Valid count is smaller than the original number of questionnaires.

Check reverse-coded and grouped variables

If the questionnaire contains reverse-coded items, handle them before calculating the scale score. A demographic variable such as gender or age does not have a reverse-coded meaning. You only need to make sure that the value labels and groups have been entered correctly.

When you want to interpret age by group, create categories with a clear logic, such as under 25, from 25 to under 35, from 35 to under 45, and 45 or older. The intervals must cover the full sample without overlapping. If you need to describe extreme values before grouping the variable, see the guide to outliers.

Steps to run a frequency table in SPSS

Open Frequencies

From the menu bar, select Analyze > Descriptive Statistics > Frequencies. The Frequencies dialog box opens. The list of variables in your file appears on the left. Select the variables you want to describe and click the arrow to move them into Variable(s).

You can select several variables at once, such as GIOITINH, TUOI_NHOM, HOCVAN, and THUNHAP_NHOM. SPSS will create a separate frequency table for each variable. If the variable list is long, use Label to identify the content instead of relying only on the code names.

Select Statistics when you need the mean

If you only need counts and percentages, you can skip the Statistics button. If the variable is Scale and you want to see Mean, Median, Mode, Standard Deviation, Minimum, or Maximum, click Statistics and select the statistics you need.

To calculate the mean, add all valid observations and divide the result by the number of valid observations. In SPSS, Mean is calculated from valid values and excludes declared missing codes. The formula can be written as: Mean = sum of valid values / number of valid observations.

Mean is more suitable when the variable has a quantitative meaning and its values are not strongly skewed. A Mean for gender coded as 1 and 2 has no useful interpretation. For this type of variable, use Frequency and Percent.

Select Charts when you need a graph

Click Charts to choose a chart. A bar chart suits a categorical variable, a pie chart may be used when there are only a few groups, and a histogram is more suitable for a quantitative variable. In a thesis, the table provides exact figures, while the chart helps the reader see the sample structure quickly.

If you select a histogram for gender, the chart will be difficult to interpret even though SPSS can create it. Choose the chart according to the nature of the variable rather than choosing one simply because it looks easy to present.

Select Format and run the procedure

The Format button lets you sort values in ascending order, descending order, or by frequency. For age groups, ascending order is usually easier to read. For a service-choice variable, sorting by descending frequency can show which group is most common.

When the settings are complete, click Continue, then click OK. SPSS opens the Output Viewer. Save the output file with the run date in its name, such as tan_so_vong_1_2026_09_08.spv, so that you can distinguish it from later revisions.

Read the output

In the output, the Statistics table usually appears before the Frequency table. It reports the valid N, missing N, and sometimes Mean, Median, Mode, and standard deviation. Distinguish the valid N from the total number of rows in Data View.

The main table usually has four columns: Frequency, Percent, Valid Percent, and Cumulative Percent. Frequency is the number of observations in each category. Percent is the proportion of all cases, including cases with missing data. Valid Percent is calculated only from valid cases. Cumulative Percent is the accumulated proportion from the first value through the value currently being read.

For example, assume that you surveyed 200 people and 4 left gender blank. The valid total is therefore 196. If 110 respondents are women, Percent is 55.0% because 110 is divided by 200. Valid Percent is 56.1% because 110 is divided by 196. When describing the sample structure, state clearly which percentage you are using.

Illustrative output from SPSS

The table below is illustrative output, structured like an SPSS output table. It is not the result of a real study, and you must not copy these numbers into your thesis.

Based on the illustrative table, an accurate interpretation is: the sample contains 196 valid observations. There are 110 women, representing 56.1% of valid observations, and 86 men, representing 43.9%. There are 4 missing cases, equivalent to 2.0% of all observations.

You should not write “women account for 55% of the valid sample.” The figure 55.0% appears in the Percent column, whose denominator is 200. When you use Valid Percent, the percentage of women is 56.1%. Both figures can appear because they use different denominators.

For an ordinal variable such as age group, Cumulative Percent can be meaningful. If the “under 25” group has a Valid Percent of 30% and the “25 to under 35” group has a Valid Percent of 45%, the Cumulative Percent for the second group is 75%. This means that 75% of the sample falls within the first two groups. It does not mean that the second group alone accounts for 75%.

If you want to describe the dispersion of a quantitative variable, also read about variance and the variance formula. Those statistics address a different question from frequency.

Evaluation criteria

A frequency table does not have one general pass threshold like Cronbach's Alpha or KMO. Evaluate it by checking whether the data codes are valid, whether the observations are sufficiently complete, and whether the variable type matches the way you present it. Practical checks can be reported as follows.

This table does not turn a particular percentage into a pass or fail standard. For example, a sample that is 70% female is not automatically incorrect. Compare the distribution with the target population, sampling method, and research objective. If the study targets users of a service whose actual customers are mostly women, that distribution may be appropriate.

When interpreting the Mean of a Likert scale, state how many points the scale has and how it was coded. A Mean of 3.8 is meaningful only when the reader knows that 1 means strongly disagree and 5 means strongly agree. Do not call Mean an “agreement percentage,” because these are different types of information.

How to write this in your thesis

Write in this order: the number of valid observations, the groups with the highest or lowest percentages, the specific percentages, and then a note about missing data when applicable. Keep the interpretation close to the table and do not add causes that the table has not tested.

How to write this in your thesis: “The frequency statistics show that [N valid] observations were used to analyze the variable [variable name]. The [group name] group has the highest proportion, with [Frequency] observations, corresponding to [Valid Percent]%, while the [group name] group accounts for [Valid Percent]%. There are [N missing] missing cases, representing [Percent]% of all observations.”

If there are no missing data, you can use a shorter version: “Among the [N] valid observations, the [group name] group contains [Frequency] people, representing [Percent]%, which is higher than the other groups.” Replace every bracketed placeholder with the figures from your own output.

For Mean, use this template: “The variable [variable name] has a mean of [Mean] and a standard deviation of [Std. Deviation], with a minimum value of [Minimum] and a maximum value of [Maximum].” Include Minimum and Maximum only when they help the reader understand the range of the data.

After writing, compare the table in Word with the SPSS output again. A common error is to copy Frequency from one run and Percent from another, especially after removing questionnaires or changing the missing-value codes.

Common errors when running a frequency table

Using Mean for a nominally coded variable

Do not conclude that “the average gender is 1.56.” This figure only reflects the codes assigned to Male and Female. It does not represent an interpretable characteristic. Use Frequency and Valid Percent instead.

Confusing Percent with Valid Percent

These columns differ when missing data are present. If you use Valid Percent to describe the structure of valid respondents, use that column consistently for every group.

Grouping values after running the output

If you change the age-group codes in Data View, the old output does not update automatically. Run Frequencies again and check the date and time of the output. This is why you should save separate file versions rather than overwrite every previous run.

Treating an error code as a valid group

The value 99 can appear as the largest group if you forget to declare it as missing. When you see an unusual frequency, return to Variable View, check Values and Missing, and run the analysis again. Do not delete all 99 codes before you know what they represent.

Making claims beyond the descriptive table

A frequency table shows how many people belong to each group. It does not prove that one group is more satisfied, has higher income, or causes an outcome variable. Those conclusions require a crosstab, a statistical test, or an appropriate model.

Frequently asked questions

What should you include when interpreting a frequency table in SPSS?

State the number of valid observations, the counts and percentages for the main groups, the group with the highest or lowest percentage, and the number of missing cases when applicable. For an ordinal variable, you can also interpret Cumulative Percent.

Do you need to report both Frequency and Percent in a frequency table?

It is useful to report both the first time the result appears, for example, “110 people, representing 56.1%.” After the column headings are clear, you can focus on the percentage to keep the writing concise. Remember to distinguish Percent from Valid Percent when missing data are present.

How do you calculate the mean in SPSS?

Mean is calculated by dividing the sum of valid values by the number of valid observations. In SPSS, go to Analyze > Descriptive Statistics > Frequencies, select Statistics, check Mean, then click Continue and OK. Do not use Mean to interpret a nominal variable that is only coded with numbers.

What should you do if the frequency table does not show Valid Percent?

If the table has no Missing row, all cases may be valid, so Percent and Valid Percent are identical. You should still check the Statistics table and Data View to make sure that no missing codes have been left undeclared.

Should you include every frequency table in Chapter 4?

Include the tables that support the description of your sample and research variables. If there are many variables, you can combine tables with the same structure or move detailed tables to an appendix. The figures in your interpretation must still match the original output.

Do you need to create a chart together with the frequency table?

No. A table is suitable when exact figures are important, while a chart is useful when you want to show the sample structure visually. For a variable with many groups, a table is often easier to read than a pie chart.

Open your .sav file, run Analyze > Descriptive Statistics > Frequencies again for each descriptive variable, mark the rows with missing data, and replace the placeholders in the template with your actual figures before adding the results to Chapter 4. If you need to run the analysis on your own .sav or .csv file, M4 data analysis with SPSS and SmartPLS carries out this step.