How to Calculate Cumulative Frequency in Statistics and SPSS

What is cumulative frequency?

Cumulative frequency tells you how many observations have been added from the first group through the group currently being examined. For a variable arranged in order, you take the frequency of the current group and add all frequencies from the groups before it. This measure is commonly used to answer the question: how many observations fall from the lowest level up to a specific level?

For example, suppose a table has three score levels with frequencies of 12, 18, and 20. The cumulative frequency at the first level is 12. At the second level, it is 12 + 18 = 30. At the third level, it is 30 + 20 = 50. The final value is always equal to the total number of valid observations when you calculate it across all groups.

The general formula is:

F_i = f_1 + f_2 + ... + f_i

Here, f_i is the frequency of group i, while F_i is the cumulative frequency through group i. To calculate the cumulative percentage, divide the cumulative frequency by the total number of valid observations and multiply by 100:

P_i = (F_i / N) × 100%

N means the number of valid observations used in the table. It does not necessarily mean the original number of rows in the file. When the data contain missing values, the denominator for the cumulative percentage can be smaller than the total number of records you see in Data View.

Why cumulative frequency matters in quantitative research

Cumulative frequency gives you a quick view of how data are distributed along an ordered scale. For age, you can see how many people are no older than 25. For grouped income, you can see the proportion of people in income groups from the lowest level up to a specific threshold. For scores, the measure shows how many people fall from the lowest score through each score level.

In a questionnaire dataset, cumulative frequency is most useful when the values have a clear order. Education level, income level, agreement level, and age group can be arranged meaningfully. A nominal variable with no natural order, such as an arbitrarily assigned industry code, can still be added arithmetically, but the interpretation is weaker.

You can also use the cumulative percentage to locate a threshold. For example, if Cumulative Percent at the “from 5 to under 10 million” group is 68,4%, you can read this as 68,4% of valid observations belonging to this group or a lower group. This interpretation differs from saying that the group alone accounts for 68,4%.

Cumulative frequency does not replace other descriptive statistics. When you need to know how widely the data vary around the mean, see standard deviation. When you need the squared measure of deviation, see variance. Cumulative frequency mainly describes counts and proportions accumulated in an ordered sequence.

What cumulative frequency value is acceptable?

Cumulative frequency has no fixed “acceptable” threshold like Cronbach's Alpha, KMO, or factor loading. It is a descriptive calculation, so you check whether the group order makes sense, whether the total number of observations is correct, and whether the final value is reasonable. A correct table must show cumulative frequency staying the same or increasing from the first group to the last.

The table below gives practical checking rules. The rows concerning statistical principles identify their sources. The rows checking the arithmetic are internal checks of the table and are not attributed to an external source.

If the table ends at 94% instead of close to 100%, check whether SPSS has excluded missing values. Do not alter the percentages just to make the table look better. Identify the denominator, the number of missing observations, and the missing-value coding before you write the interpretation.

How to read cumulative frequency in the output

In SPSS, create the table through Analyze > Descriptive Statistics > Frequencies. Move the variable you want to inspect into Variable(s), keep the option to display the frequency table, and select OK. For a continuous variable with too many distinct values, the Frequencies table can become long and difficult to read. In that case, group the variable first or use a grouped table that fits your research design.

The output usually contains the columns Frequency, Percent, Valid Percent, and Cumulative Percent. Frequency is the number of observations in each group. Percent is calculated from all rows and usually includes missing cases in the denominator. Valid Percent is calculated from valid observations. Cumulative Percent is the cumulative percentage based on Valid Percent, so this is usually the column to use when interpreting the distribution of valid data.

The table below contains illustrative output, not the result of a real study. The figures are constructed so you can compare them with the SPSS table open on your screen.

The second row means that 31 people belong specifically to the income group from 5 to under 10 million, representing 31.6% of valid observations. When Cumulative Percent is 50.0%, read it as 50.0% of valid respondents having income under 10 million, which includes the first two groups.

You need to distinguish Percent from Valid Percent in this example. The Percent for the first group is 18.0% because the denominator is 100 rows. Valid Percent is 18.4% because the denominator is 98 people with valid answers. If your thesis describes the composition of people who answered the question, Valid Percent usually reflects the analysis population more accurately.

You can check the cumulative frequency by adding the Frequency column: 18, 18 + 31 = 49, 49 + 28 = 77, 77 + 21 = 98. The cumulative percentages are 18.4%, 50.0%, 78.6%, and 100.0%, respectively. If Cumulative Percent decreases in any row, check the display order or the way the table was exported.

What to do when cumulative frequency looks wrong

Because cumulative frequency has no passing threshold, “wrong” usually means that the table shows an unusual pattern. Follow the checks below in order so that you do not change the data arbitrarily.

Check coding and value labels

Open Variable View and inspect Values, Missing, and the measurement level of the variable. Some tables are wrong because code 99 was entered as a real income level, or because the groups “1”, “2”, and “3” were assigned incorrect labels. If 99 is a missing-value code, define it correctly before running Analyze > Descriptive Statistics > Frequencies again.

Check the group order

SPSS can sort groups by numerical code or alphabetical order. If the label “Over 15 million” appears before “Under 5 million”, the cumulative percentage becomes difficult to interpret even though the arithmetic remains correct. Check the numerical codes, not only the displayed labels.

Check duplicate and invalid rows

If the number of observations is unusually high, review duplicate entries, test records, or respondents who did not complete the questionnaire. Deleting data requires a criterion recorded in advance, such as excluding a record with no answer for the main variable or a record that violates the sample-screening condition. Do not delete a row only because it makes the cumulative table look less attractive.

Recode the variable when the table is too detailed

An age variable can contain dozens of distinct values, making Cumulative Percent difficult to use in the descriptive section. You can create grouped intervals with a meaningful research basis, retain the original variable, and document the grouping rule. If your goal is to locate the middle position of the data, also examine the median instead of trying to infer it from an excessively long table.

Handle outliers at the appropriate stage

A very high or very low value is not automatically a data-entry error. Before removing it, check the response record and the logic of the variable. You can consult the guide to outliers to distinguish a valid extreme value from a data error. If you retain the outlier, report a consistent treatment in the research methods section.

Distinguishing cumulative frequency from similar measures

Frequency is the number of observations in one individual group. Cumulative frequency is the running total from the first group through the group being examined. In the example above, the Frequency of the second group is 31, while the cumulative frequency through the second group is 49.

Percent is the proportion represented by one individual group. Cumulative percentage is the proportion represented by that group plus all preceding groups. Therefore, you cannot write “the group from 5 to under 10 million accounts for 50%” simply because Cumulative Percent is 50%. The accurate sentence is “groups with income under 10 million account for 50% of valid observations.”

Cumulative frequency also differs from the mean, median, and mode. The median is the position that divides ordered data into two parts, while cumulative frequency records the running addition process. From a cumulative table, you can estimate the group containing the median, but you should not call every value at the 50% point the exact median when the data have been grouped.

Finally, cumulative frequency is not variance or standard deviation. Those two measures describe dispersion around the mean. If you need to check the formula and interpretation of variance, see the variance formula. The appropriate measure depends on your research question, not on which table is easier to create in SPSS.

Common mistakes

The first mistake is adding frequencies according to the order of the data rows instead of the order of the variable. Cumulative frequency must follow a meaningful order, such as increasing age or agreement level from low to high. If you change the group order, your conclusion about a threshold also changes.

The second mistake is using Percent when missing values exist and then calling it the proportion of valid respondents. Look at the Missing row, then choose Valid Percent when your goal is to describe people with data for that variable.

The third mistake is calculating cumulative percentages from the original sample size while SPSS calculates them from the valid number. These two approaches can produce different results. Keep the denominator consistent across the table, the written interpretation, and the sample-description table.

The fourth mistake is treating cumulative frequency as a hypothesis test. This measure does not give you a p-value or establish a relationship between variables. If you want to analyse effects, choose a model and test that match your hypothesis instead of using cumulative percentages as evidence of causality.

How to write this in your thesis

When reporting the result in your thesis, you can write: “Descriptive statistics show that [x%] of valid observations have [characteristic] from [lowest level] to [threshold], as shown in the Cumulative Percent column of table [table name]. The total number of valid observations is [N], including [number of missing] missing cases.” Replace every bracketed part with the actual values from your output.

Frequently asked questions

What is cumulative frequency in statistics?

Cumulative frequency is the number of observations added from the first group through the group currently being examined. It tells you how many observations fall in the current group and every preceding group according to the selected order.

Do you need to sort the data to calculate cumulative frequency?

Yes. Arrange the groups in a meaningful order before interpreting the result. You do not necessarily need to sort the original data manually in Data View if SPSS already displays the values in the correct order, but you must check the codes and value labels.

What should the final cumulative frequency equal?

The final cumulative frequency must equal the total number of valid observations across the groups. If you are looking at cumulative percentage, the final value is usually 100% when the table includes all valid groups. If it is lower, check missing values, data filters, and rounding.

Is Cumulative Percent in SPSS the same as cumulative frequency?

Cumulative Percent is a cumulative percentage, not a cumulative count. You can calculate the cumulative frequency from the Frequency column, while Cumulative Percent shows the running proportion based on valid observations.

What should you do when cumulative frequency is wrong in SPSS?

Check the missing-value codes, value labels, group order, and duplicate or invalid records in sequence. After correcting the documented cause, run Analyze > Descriptive Statistics > Frequencies again and save the new output table. Do not manually change the figure in your thesis without correcting the data or the calculation rule.

Open your .sav file, run Frequencies, compare Frequency, Valid Percent, and Cumulative Percent, then write the interpretation using the actual denominator. If you need to run and check the full analysis on your file, DoThesis M4 analysis supports SPSS and SmartPLS with .sav or .csv data.