How to Run Frequencies in SPSS and Read the Output Correctly

When to use a frequency table in SPSS

An SPSS frequency table counts how many times each value appears in a variable. With survey data, you usually use it to describe gender, age, occupation, income, education, customer groups, or responses on a Likert scale. The output shows how many respondents selected each option and what share of the sample they represent.

This is usually the first check after you import a .sav or .csv file. You can identify incorrect variable codes, values outside the scale, blank cells, and groups with an unusual number of observations before running Cronbach's Alpha, EFA, or regression. A frequency table does not test a causal relationship and does not answer a hypothesis such as “X affects Y.” It answers a descriptive question: who is in your sample, and how are the response options distributed?

If a variable contains many continuous values, such as an exact age, a mean score, or income, you may also need to examine standard deviation, variance, or the median instead of listing every value. Frequency tables are most suitable for categorical variables and Likert variables with a limited number of levels.

When you need to examine two categorical variables at the same time, such as gender by purchase intention, use Crosstabs. Therefore, an “SPSS Crosstabs table” is usually the next step after Frequencies, rather than another name for the same table.

Preparing your data before running the analysis

Open Variable View and Data View and check the data first. Each column should represent one variable, and each row should represent one respondent. Variable names should be short, contain no spaces, and match the codebook in your methodology chapter. For example, GIOITINH, TUOI, TN1, and TN2 are easier to check than long names or names containing special characters.

For categorical variables, check the Measure and Value Labels columns. For example, GIOITINH = 1 may mean Male and GIOITINH = 2 may mean Female. Frequencies will still count the data if you entered only numbers, but the output will be difficult to read when Value Labels are missing. For Likert variables, confirm that the codes run from 1 to 5 or from 1 to 7, depending on how you designed the questionnaire.

Checking variable type and coding

In Variable View, inspect the Type, Values, Missing, and Measure columns. A categorical variable can be set to Numeric with value labels attached. Text such as “Male” and “Female” also works with Frequencies, but numeric codes are usually more convenient for later analysis.

Do not combine codes such as “don't know” and “no answer” with valid responses if you want to calculate the actual response distribution. You can define these codes in Missing so that SPSS separates them from Valid Percent. If a variable uses 99 for a missing answer and you do not define it as missing, SPSS will treat 99 as a real value.

Checking missing data and unusual values

Quickly filter the demographic variables and indicators to find blank cells. A Likert item that should contain values from 1 to 5 but includes 0, 6, or 55 needs to be checked against the original file. If an unusual value resulted from a data-entry error, correct it according to the questionnaire or the cleaning rule you recorded earlier. Do not automatically replace every unusual value with the mean.

You should also examine outliers when analysing continuous variables. An outlier is not automatically a data-entry error, so deleting a row simply because it differs from most of the sample requires a reason. Save a version of the data before making that change.

Checking reverse-coded variables

A frequency table does not recode reverse-worded items automatically. If the questionnaire contains a statement with the opposite meaning, recode it before calculating the scale score. In SPSS, use Transform > Recode into Different Variables to create a new variable, keep the original variable, and check the value labels again. Then run Frequencies for the new variable to confirm that the recoded values remain within the correct range.

Steps to run a frequency table in SPSS

Step 1: Open the Frequencies dialog box

In SPSS, select Analyze > Descriptive Statistics > Frequencies. The Frequencies dialog box opens with the variables listed on the left. Select the variables you want to describe and move them to the Variable(s) box using the arrow button.

If you are checking gender, grouped age, and education, you can move all three variables at once. SPSS will create a separate table for each variable. You can also run indicators TN1 through TN5 together, but the output will be long. Select variables that serve the same checking purpose so that the output remains easy to read.

Step 2: Select the tables and charts to export

Keep Display frequency tables selected if you want SPSS to produce counts and percentages. Click Charts and select Bar charts for categorical variables or Pie charts when there are only a few groups. A Histogram is more suitable for a continuous variable and can show its distribution, but it is not necessary for every demographic variable.

Click Statistics if you want Mean, Median, Mode, Standard deviation, Minimum, and Maximum. For a gender frequency table, mean-based statistics usually have no practical meaning. For age or a score, they may be useful, but choose statistics that match your research question.

Step 3: Run the command and save the syntax

Click Continue and then select OK to run the analysis. The Output Viewer will produce a Frequency Table for each variable and a chart if you selected one. Save the syntax rather than saving only the output. This allows you to repeat exactly the same procedure after correcting the data.

To open the syntax window, select Paste instead of OK. SPSS will generate the corresponding command, usually in a form such as FREQUENCIES VARIABLES=GIOITINH TUOI /ORDER=ANALYSIS. Save the .sps file in the same folder as the .sav file, record the run date, and note which variables you checked.

Step 4: Run Crosstabs when you need to compare groups

Select Analyze > Descriptive Statistics > Crosstabs. Move the grouping variable to Row(s) and the variable you want to compare to Column(s), then click Cells to select Counts and Percentages. If you want percentages by row, select Row. If you want percentages by column, select Column.

For example, to examine purchase intention by gender, place GIOITINH in Row(s) and YD_MUA in Column(s). Decide in advance which percentage answers your question. If you want to know what percentage of each gender group selected each purchase-intention level, Row percentages are usually easier to interpret.

Crosstabs can also include a Chi-Square test, but that is a different analysis from a descriptive frequency table. Select Statistics > Chi-square only when the research question and test conditions are appropriate. Do not conclude that two variables are related merely because the percentages in the crosstab differ.

Reading the output

The main output table is called Frequency Table. Its usual columns are Frequency, Percent, Valid Percent, and Cumulative Percent. The example below is illustrative output, not the result of a real study.

Frequency is the number of observations in each group. In the example, 82 respondents are coded as Male and 118 are coded as Female. Percent uses all 204 cases as the denominator, including the four missing cases.

Valid Percent uses only the 200 cases with valid data. Therefore, the percentage of Male respondents is 82 divided by 200, or 41.0%, while the percentage of Female respondents is 118 divided by 200, or 59.0%. When reporting sample characteristics, state clearly that you used the valid sample if missing data are present.

Cumulative Percent is the accumulated percentage. It is more useful for an ordered variable such as income level or agreement level from low to high. For gender, this column has little substantive meaning because Male and Female have no natural order.

The Missing row shows the cases that SPSS excluded from Valid Percent. If there is no Missing row, the variable may have no missing data, or the missing data may have been defined in another way. Compare the total number of rows in Data View with the total number of valid observations.

Evaluation criteria

A frequency table is descriptive statistics, so it does not have one general “acceptable” threshold like Cronbach's Alpha or KMO. Evaluate it based on whether the codes are valid, the data are sufficiently complete, and the table matches the descriptive objective. The criteria below help you check parts of the output that are often misunderstood.

The sources in the table support data checking and presentation. They do not turn a frequency table into a hypothesis test. If one group contains only a few respondents, do not conclude automatically that the data are wrong. Compare the result with your sampling method and describe it as a characteristic of the sample.

When you need to report dispersion, do not call Percent “variance.” Percent is the proportion of observations, while variance measures the dispersion of numeric values. You can also review the variance formula if you are writing the descriptive-statistics section for a quantitative variable.

How to write this in your thesis

In Chapter 4, you do not need to paste the entire Output Viewer when the table contains many technical rows. Select the table you need, reformat it, keep the variable names and units clear, and explain one or two main points. For demographic variables, report the count and valid percentage in most cases.

You can adapt the wording as follows: “The descriptive statistics show that the study sample contains [total number] valid observations. Of these, the [group 1 name] group includes [n1] respondents, accounting for [p1]%, while the [group 2 name] group includes [n2] respondents, accounting for [p2]%. The percentages are calculated from [number of valid observations] cases after missing values were excluded.”

When reporting Crosstabs, state clearly whether the percentages are calculated by row or by column. For example: “Within the [group name] group, [p]% of respondents selected [option]. The percentages in the table are calculated by row.” Do not write that “gender affects purchase intention” based only on a crosstab. To conclude that one variable affects or is related to another, you need to run a test that matches the model and hypothesis.

You can also review how to report the median when a variable is ordered or has a skewed distribution. Place the table close to the paragraph that interprets it, number tables consistently, and identify the source as “Author's survey data analysis results” when the data are yours.

Common errors when running Frequencies in SPSS

Selecting the respondent ID instead of the variable you want to describe. If you run Frequencies for a questionnaire ID, each value will usually appear once and the table will have no descriptive value. Check the variable name again before clicking OK.

Using Percent instead of Valid Percent when data are missing. The two columns differ when one or more cases are missing. State the denominator used in your thesis.

Leaving 99 or 999 as a valid value. The output will show an additional, unusually large group. Check Value Labels and Missing before deciding whether to recode the value.

Using a pie chart for too many groups. The chart becomes difficult to read when a variable has many levels. In that situation, a table or Bar chart is usually clearer.

Running Frequencies for a mean score and reading it as a categorical variable. If you have calculated a scale mean, examine Mean, Standard deviation, Minimum, and Maximum instead. The articles on standard deviation and variance will be useful at this stage.

Editing the data directly without saving the original. Keep the original file in read-only mode, create a separate cleaned-data file, and record each change. When your supervisor asks why the number of observations decreased, you need to trace every decision.

Frequently asked questions

What is an SPSS frequency table?

A frequency table counts the observations at each value or group of values for a variable. The output usually includes Frequency, Percent, Valid Percent, and Cumulative Percent. Use it to describe the sample and check the data before later analyses.

Does an SPSS frequency table need a statistical test?

A frequency table is descriptive statistics, so it does not require a p-value by itself. If you want to test the relationship between two categorical variables, you can use Crosstabs with Chi-Square when the question and test conditions are appropriate. If you want to test effects between variables in a model, you need a different analysis.

Why are Percent and Valid Percent different?

Percent includes cases with missing data, while Valid Percent includes only cases with valid values. The two columns differ when Missing cases are present. In your report, state whether the percentage uses the total number of observations or the number of valid observations.

What should I do when 99 appears in the frequency table?

First, check whether 99 is the code for “no answer” or an error in the questionnaire. If it is a missing code, define it under Missing or recode it according to the cleaning rule you agreed on. If 99 is a real value, retain it and explain its meaning. Do not delete it simply because it differs from the other groups.

When should I use Crosstabs instead of Frequencies?

Use Frequencies to describe one variable by itself. Use Crosstabs when you want to examine the distribution of one variable across groups defined by another variable, such as agreement level by gender. In Crosstabs, select Row Percent or Column Percent carefully so that the output answers your research question.

Should I include the entire SPSS output in my thesis?

You do not need to include the entire Output Viewer when it contains many intermediate tables. Select the tables you need, reformat them for readability, and keep the syntax or original output file for reference. In the explanation, report the number of observations, the main percentages, and how missing data were handled.

Open your data file now, run Analyze > Descriptive Statistics > Frequencies for the demographic variables, and record Valid, Missing, and Valid Percent before moving to the next analysis step. If you need to run the procedure on your .sav or .csv file and check the output through the M4 workflow, see DoThesis M4 data analysis.