What Is a Frequency Chart? How to Read and Analyze It in SPSS

What Is a Frequency Chart

You open your SPSS file, run descriptive statistics, and see columns or bars showing how many people selected each response option. That is a frequency chart. It is commonly used to describe how values appear in a variable. For a nominal variable, such as gender or occupational group, the chart shows how many observations belong to each group. For an ordinal or numerical variable, it shows where the data are concentrated, how dispersed they are, and whether unusual values may be present.

The frequency of a value is the number of times that value appears in the data. If 80 people responded and 32 selected level 4 for SAT1, the frequency of level 4 is 32. The percentage is calculated by dividing the frequency by the total number of observations and multiplying by 100. When the data contain blank cells, SPSS separates the percentage based on all valid cases into Valid Percent, while Percent is usually calculated using all cases, including missing values.

In quantitative research, a frequency chart usually appears together with a Frequencies table. You can also review the Statistics section when you need to place this chart within a broader data analysis process.

Why Frequency Charts Matter in Quantitative Research

A frequency chart helps you check the data before running Cronbach's Alpha, EFA, regression, or a PLS-SEM model. It gives you a quick view of the raw data. You may spot a response option that almost nobody selected, a demographic group that is too small, an incorrectly coded value, or a variable with many blank entries.

For example, your questionnaire uses a 5-point Likert scale, but the chart for GIA1 contains the value 55. The problem is then in the coding or data entry, not in the analysis result. Check Variable View, the value labels, and the data column before continuing.

The chart also helps you describe the research sample in Chapter 4. If GIOITINH contains 146 people in group 1 and 104 people in group 2, you can report the sample structure using frequencies and percentages. For age or income, check how the categories were formed before interpreting them. A clear chart cannot fix a grouping scheme that lacks a sound basis.

You can also use the chart to look for duplicate entries, a respondent selecting the same level across most variables, or other unusual data patterns. A frequency chart does not by itself establish that an observation is an outlier. When you need to inspect unusual values, review outliers in the data using appropriate indicators.

What Frequency Chart Is Acceptable

There is no single cutoff for deciding whether a frequency chart is acceptable. It is a descriptive tool, so you evaluate it based on the research objective, variable coding, number of missing values, and whether the distribution is reasonable. The principles below help you check the right issues, but they should not become mechanical conditions for every chart.

Cutoffs such as Cronbach's Alpha at 0.7 or above belong to scale reliability assessment (Nunnally, 1978). A Corrected Item-Total Correlation at 0.3 or above is a commonly used reference point for an item (Nunnally and Bernstein, 1994). A KMO at 0.5 or above and a statistically significant Bartlett's Test are commonly used when assessing whether EFA can be performed (Kaiser, 1974). These are not frequency-chart cutoffs.

How to Read a Frequency Chart in the Output

In SPSS, run the frequency table through Analyze > Descriptive Statistics > Frequencies. Move the variable you want to inspect from the list on the left to the Variable(s) box. If you only need the numerical table, keep Display frequency tables selected. To create a chart, click Charts, choose Bar charts for a categorical variable or Histograms for a numerical variable, and then select the appropriate display option.

The output table usually contains the columns Frequency, Percent, Valid Percent, and Cumulative Percent. Frequency is the number of observations for each value. Percent is the percentage calculated across all cases, including missing cases. Valid Percent is the percentage calculated only across cases with valid data. Cumulative Percent is the cumulative percentage, which is appropriate only when the values have an order, such as agreement levels or age groups arranged from low to high.

Illustrative output

The table below is illustrative output, not the result of an actual study. Assume that the variable MỨC ĐỘ HÀI LÒNG has 100 respondents, including 4 missing responses.

In this example, 39 people selected level 4. The Percent for level 4 is 39.0% because the original total sample is 100. The Valid Percent is 40.6% because SPSS removes the 4 missing cases from the denominator. When describing the people who answered the item, use Valid Percent and state that 4 cases were missing.

If the variable uses a Likert scale from 1 to 5, Cumulative Percent can support the statement that 78.1% of valid responses were at level 4 or below. Do not use cumulative percentages for unordered variables such as province codes, industry codes, or gender codes. The value 2 in a gender variable is not higher than the value 1.

For a Histogram, the horizontal axis usually represents value intervals and the vertical axis represents frequency. A right-skewed shape, left-skewed shape, or multiple peaks is a signal to investigate further. If you need to assess regression assumptions, also review the residual chart, Normal P-P Plot, and appropriate diagnostic tables instead of relying on the frequency chart alone.

What to Do When a Frequency Chart Looks Wrong

First, determine whether the problem is in the data or in the way the chart is presented. If a value appears outside the questionnaire range, open Variable View and check Values, Missing, and the data type. Then compare several rows in Data View with the original response forms or collection file.

Next, inspect missing values. Do not replace a blank cell with level 3, the mean, or a convenient number by default. If the respondent skipped the item, keep it as missing and use a consistent treatment throughout the analysis. When missing values are frequent or concentrated in one group, report the issue to your supervisor before removing cases.

If one group has a low frequency, do not merge categories simply to make the chart look balanced. Merging is acceptable only when the groups are substantively similar, the grouping fits the research question, and the decision is recorded in the methods section. For a categorical independent variable in regression, you also need to check how SPSS creates dummy variables and selects the reference group.

If responses on a Likert scale are strongly concentrated at one level, describe that pattern accurately. Do not delete an item merely because the chart is asymmetric. Deletion should be based on the measurement objective and results such as Corrected Item-Total Correlation, Factor Loading, or Outer Loading, depending on the method. In PLS-SEM, an Outer Loading from 0.7 or above is a commonly referenced benchmark (Chin, 1998) or (Hair et al., 2022).

The recommended order is to check the coding, check missing values, compare the data with the original file, review the category-merging rule, and only then rerun the analysis. Record each change in a separate file. The committee may ask how the data were cleaned, so you need to explain the reason for every edit.

Frequency Charts and Descriptive Statistics

A frequency chart is a visual representation of how often values occur. A frequency table is the numerical information that accompanies the chart. Descriptive statistics cover a broader range and may include frequencies, percentages, Mean, Median, Mode, Standard Deviation, Minimum, and Maximum.

For gender, a frequency table and bar chart are usually more appropriate than Mean. If code 1 means male and code 2 means female, the Mean of that variable does not represent an average score. For age measured in years, you can review Mean and Standard Deviation and use a Histogram to observe the distribution.

A frequency chart is also easy to confuse with a normal distribution chart. A Histogram shows the shape of the data across value intervals. A normal curve placed over the histogram is a visual aid, not the only evidence for deciding whether the data are normally distributed.

You also need to distinguish a frequency chart from a crosstab. Analyze > Descriptive Statistics > Crosstabs shows the joint distribution of two or more variables, such as gender by age group. If your research question requires a comparison of proportions across groups, a crosstab may be more suitable than a univariate frequency chart.

A frequency chart is also different from a chart used to analyze moderating or a moderator. A moderator is a concept in a relationship model between variables, while a frequency chart only describes observed data and does not test a moderation effect. Similarly, when working with a questionnaire and items, you can review what a scale is to avoid confusing a measurement scale with a descriptive chart.

Common Errors

The first error is using Percent instead of Valid Percent when the data contain missing values. If you report the proportion of people who answered an item, check the denominator and state clearly which percentage is being used.

The second error is treating codes as quantitative variables. Codes 1, 2, and 3 for field of study are category labels when that is how the questionnaire defines them. Changing code 1 to 10 or code 2 to 20 does not make the second field larger than the first.

The third error is immediately deleting values with low frequencies. A small group may still be important to the study. Before removing it, determine whether it is a data-entry error, a valid respondent, or a group that is genuinely small in the population.

The fourth error is inserting a chart without reporting its figures in the discussion. The reader needs to know which group is largest, which is smallest, the total number of valid cases, and whether missing values exist. The image supports the explanation but does not replace it.

The fifth error is using a frequency chart to support a hypothesis conclusion. A chart does not show whether an independent variable affects a dependent variable, whether a p-value is significant, or how much variance the model explains. Those conclusions require an appropriate test or model. A chart also does not replace mediator analysis, although you can read more about a mediator when building the research model.

Frequently asked questions

What is a frequency chart in SPSS?

It is a chart that describes how many times each value or group appears in the data. You can create it through Analyze > Descriptive Statistics > Frequencies, usually together with the Frequency, Percent, Valid Percent, and Cumulative Percent tables.

What frequency chart is acceptable?

There is no universal cutoff for deciding whether a frequency chart is acceptable. Check the coding, number of valid observations, missing values, and whether the distribution fits the research objective. Cutoffs such as Alpha, KMO, or Factor Loading belong to other analyses and are not used to score a frequency chart by itself.

Should I use Percent or Valid Percent?

When there are no missing values, the two columns usually show the same result. When missing values exist, Valid Percent is calculated from valid cases and is usually appropriate when describing people who answered the item. In your thesis, report the number of valid cases so that the reader understands the denominator.

Why does the frequency chart show Missing values?

SPSS displays Missing when some rows contain no valid data for the variable being inspected, or when you have declared a code as missing in Variable View. Check the original file, coding, and data-entry procedure before deciding whether to retain or remove the cases.

Does a right-skewed chart mean that I should remove the data?

No. Skewness is a signal for further inspection, not an automatic reason to delete data. Consider the variable type, analysis objective, and assumptions of the method you are using. For regression, check the residuals and related indicators instead of looking only at the histogram of the original variable.

How to write this in your thesis

[Insert the variable name] had [insert valid frequency] valid observations, with [insert missing frequency] missing cases. The most common response was [insert value or category], accounting for [insert Valid Percent] of valid responses. The frequency chart showed [insert the main distribution pattern], so the data were checked for coding errors and missing-value treatment before the next analysis.

Open the .sav file again, run Analyze > Descriptive Statistics > Frequencies for the demographic variables and variables requiring inspection, then save the Frequency table and charts in the results folder. If you need the full analysis run on your own .sav or .csv file, M4 data analysis can take over this step.