APA descriptive statistics table: examples and rules
An APA descriptive statistics table lists each variable in the left-hand column and gives its statistics in the columns to the right: usually the mean (M) and standard deviation (SD), and, when your analysis needs them, n, the range, skewness, kurtosis and Cronbach's α. Survey means and standard deviations take one decimal, most other statistics two, and the sample size goes in the general note under the table. For the people in your sample, a second table gives n and % for each category.
- Put one variable per row in the left-hand column (the stub) and one statistic per column: M, SD, then only the extras your analysis uses.
- Write the statistical symbols in italics (M, SD, n, N) and leave Greek letters upright (α). Don't define them in the note.
- Round survey and Likert-scale means and standard deviations to one decimal; round correlations, alphas, skewness and kurtosis to two.
- No zero before the decimal point for values that can't exceed 1 (α = .89); keep it for values that can (SD = 0.9, skewness = −0.31).
- Give the sample size in the general note (Note. N = 412.), and say what the scale runs from and to.
See a complete descriptive statistics table, or the sample characteristics table.
01An APA descriptive statistics table for your variables
This is the table most quantitative theses open their results with. It comes from my example thesis, Brand trust and customer loyalty in subscription services: five scales, each the mean of several 7-point items, answered by 412 subscribers. Every number in it follows APA's rules for statistics, and the layout follows the APA table rules that I explain part by part in the APA table format guide.
| Variable | Items | M | SD | Range | Skewness | Kurtosis | Cronbach's α |
|---|---|---|---|---|---|---|---|
| Brand trust | 3 | 5.3 | 1.1 | 1.3–7.0 | −0.62 | 0.18 | .89 |
| Satisfaction | 3 | 5.6 | 1.0 | 2.0–7.0 | −0.85 | 0.71 | .91 |
| Perceived value | 4 | 4.9 | 1.2 | 1.0–7.0 | −0.31 | −0.27 | .84 |
| Loyalty | 4 | 5.0 | 1.3 | 1.0–7.0 | −0.54 | −0.12 | .87 |
| Word of mouth | 3 | 4.7 | 1.4 | 1.0–7.0 | −0.28 | −0.49 | .92 |
Note. N = 412. Scores are the mean of each scale's items, rated from 1 (strongly disagree) to 7 (strongly agree). Range gives the lowest and highest observed scores.
Read it the way your examiner will: the note says who answered and what a score of 5 means, the columns go from the essential (M, SD) to the diagnostic (skewness, kurtosis, α), and nothing is defined that a reader of statistics already knows. APA lists M, SD, N and n among the symbols you never define, together with Greek-letter symbols such as α.
02Which statistics to include
APA doesn't fix a list of columns. It asks for the statistics your reader needs, and nothing that repeats the text. In the theses I format, the columns that earn their place are these:
- M and SD: always, for every continuous variable. Give each its own column, so both line up on the decimal point.
- n: only if it differs between variables, for example because some respondents skipped a scale. If every variable has the full sample, give N once in the note instead of a column of identical numbers.
- Range, or minimum and maximum: useful when a reader needs to see floor or ceiling effects. Say in the note whether it is the observed range or the possible range of the scale; they are often confused.
- Skewness and kurtosis: only if your method chapter says you checked normality with them. If no test or decision depends on them, leave them out. If you report their standard errors, put them in the note when they are the same for every variable, as they are when n is the same.
- Cronbach's α (or ω): for multi-item scales, if you report reliability in this table rather than in the method chapter. Report it in one place only.
- Median and interquartile range: instead of M and SD for clearly skewed variables such as income or time on a platform. Write the median as Mdn and define IQR in the note.
Descriptive statistics by group
When you compare groups, keep the variables in the rows and give each group a pair of M and SD columns under a column spanner, such as "Monthly plan" and "Annual plan", with the group size in the spanner or in the note. The APA table format guide draws exactly this table, with the line that runs under each spanner. If the comparison is the point of the table, the test results (t, df, p, effect size) can go in further columns of the same table.
Correlations in the same table
Many theses add the correlations to the right of M and SD, making one table instead of two. The APA table format guide builds that layout step by step, so I won't repeat it here. A common convention, not an APA rule, puts each scale's alpha on the diagonal in parentheses instead of the dash. If you do that, say so in the note, and drop the separate alpha column.
APA asks you not to repeat statistics in both the text and a table. If Table 1 gives every alpha, the method chapter can say "reliabilities are in Table 1" instead of listing them again.
03A sample characteristics (demographics) table
Categorical variables, such as gender, age group or education, don't have a mean. For them the table gives the count (n) and the percentage of each category. I put the characteristics in the stub, with each group's n and % in separate columns under a column spanner, the device APA uses to group related column headings. I add a column for the full sample, and I set the category names (Gender, Age group) flush left and indent the answers under them, so each block reads as one group.
| Characteristic | Monthly plan | Annual plan | Full sample | |||||
|---|---|---|---|---|---|---|---|---|
| n | % | n | % | n | % | |||
| Gender | ||||||||
| Woman | 131 | 55 | 90 | 52 | 221 | 54 | ||
| Man | 101 | 42 | 80 | 46 | 181 | 44 | ||
| Another gender | 4 | 2 | 3 | 2 | 7 | 2 | ||
| Prefer not to say | 2 | 1 | 1 | 1 | 3 | 1 | ||
| Age group | ||||||||
| 18–24 | 74 | 31 | 31 | 18 | 105 | 25 | ||
| 25–34 | 92 | 39 | 63 | 36 | 155 | 38 | ||
| 35–44 | 45 | 19 | 47 | 27 | 92 | 22 | ||
| 45 and over | 27 | 11 | 33 | 19 | 60 | 15 | ||
Note. N = 412 (monthly plan n = 238; annual plan n = 174). Participants' mean age was 33.4 years (SD = 10.2). Percentages may not total 100 because of rounding.
- The % sign goes in the column heading, not after every number.
- Whole-number percentages are my default because they are easiest to read. Use one decimal only if your guide asks for it or the groups are large enough for the decimal to mean something, and use the same choice in every table.
- Rounded percentages rarely add up to exactly 100. Don't adjust a value to force it: say so in the note, as above.
- N is the whole sample, n a part of it. Give N once for the total and n for each group, as the note above does. Neither needs defining.
- Continuous characteristics, such as age in years, can be a row with M (SD) in the cells, or a sentence in the note. A note is tidier when it is the only continuous one.
In a thesis, this table usually belongs in the method chapter, where you describe the participants. The descriptive statistics of your study variables (Table 1) usually open the results.
04Decimals, zeros, minus signs and italics
These are APA's general rules for statistics, applied to the values that fill a descriptive table: skewness, kurtosis, alpha and percentages as well as means. The full rules for numbers in the text are in my guide to numbers in APA style.
- One decimal for means and standard deviations of data measured on whole-number scales, such as questionnaires: 5.3, 1.1.
- Two decimals for other means and standard deviations (reaction times, grade point averages, anything recorded with decimals), and for correlations, proportions and test statistics. I round skewness, kurtosis and alpha to two as well, the usual choice in published tables.
- Leading zero only when the value can exceed 1. Standard deviation 0.9, skewness −0.31, kurtosis 0.18: zero. Alpha .89, correlation .58, proportion .43, p = .006: no zero.
- Negative values: no space between the minus sign and the number: −0.62. In Word for Windows, type
2212and press Alt+X for a true minus sign; on a Mac, use Insert › Advanced Symbol. A plain hyphen is the common fallback, but be consistent. - Italics: Latin-letter symbols are italic (M, SD, n, N, Mdn, r, p); Greek letters stay upright (α). Words such as Skewness and Range in headings are not symbols, so they stay roman.
- Round, don't cut. SPSS shows 5.3172; the table shows 5.3. A value of −.618 becomes −0.62, not −0.61.
APA calls the decimal places "general guidelines" and says the deciding question is to round as much as possible while keeping the precision your reader needs. If your field or your university's guide asks for more decimals, follow it, the same way in every table.
05From SPSS output to an APA table
SPSS gives you all the numbers and none of the formatting. Here is its Descriptives output for three of the scales in Table 1, and the same results rebuilt for the thesis.
Get the numbers out of SPSS
- For scale scores, choose Analyze › Descriptive Statistics › Descriptives, move the variables into Variable(s), click Options and tick what your table needs: Mean, Std. deviation, Minimum, Maximum, and Kurtosis and Skewness if you report them. Mean, standard deviation, minimum and maximum are ticked by default.
- For categories, choose Analyze › Descriptive Statistics › Frequencies. Its table gives Frequency (your n), Percent and Valid Percent. If anyone skipped the question, Valid Percent leaves them out and Percent doesn't: use one of the two consistently and say in the note how missing answers were handled.
- Get alpha from the reliability analysis you ran for each scale, and correlations from the correlation output, if they go in the same table.
Build the table in Word
- Don't keep the pasted SPSS table: its fixed widths, fonts and shading fight every APA setting. Type the rounded values in the thesis instead, one row per line with a Tab between cells, select the lines and choose Insert › Table › Convert Text to Table (Word for Windows and Mac). Under Separate text at, choose Tabs.
- For spanners such as "Monthly plan", add a heading row, select the two cells above n and % and choose Table Layout › Merge Cells (on a Mac, Merge Cells is on the Layout tab next to Table Design).
- Indent the answers under each category name: select those stub cells and click Increase Indent on the Home tab once.
- Then give the table its APA finish: lines only at the top, under the headings and spanners and at the bottom, no vertical lines, centred columns, a decimal tab for the numbers and a caption Word numbers for you. The APA table format guide has each step for Windows and Mac.
Twenty SPSS tables to rebuild?Send me the chapter and I'll turn them into APA tables
06What to write about the table
Mention the table by number before it appears and tell the reader what to notice. Don't read the table back: the numbers are already there.
Table 1 shows the descriptive statistics for the five scales. All means were above the scale midpoint of 4, with satisfaction highest and word of mouth lowest. Skewness and kurtosis were between −1 and 1 for every scale, so the scores were treated as approximately normal.
- In the text, use the word, not the symbol, unless a value follows: "the means were above 4", but "satisfaction was highest (M = 5.6, SD = 1.0)". That parenthetical is for a value you want to stress, or for a result that has no table.
- The −1 to 1 cut-off is a convention of your method, not an APA rule. Name the criterion you used and cite where it comes from in your method chapter.
- One or two variables don't need a table. Report them in the sentence. A table pays off from about three variables or more.
- Too many variables for one page? Keep the main scales in the chapter and move item-level statistics to an appendix table (Table A1). If a pattern matters more than the exact values, a chart may say it better: see APA figure format.
07Problems I find in results chapters
The n is different for every variable
Some respondents skipped some scales, so SPSS's "Valid N (listwise)" is smaller than the N of each variable. Add an n column, or, if the differences are small, say in the note "ns ranged from 398 to 412 because of missing responses." State the analysis sample in the method chapter so the numbers in all tables can be traced back to it.
A percentage column adds up to 99 or 101
That is rounding, and it is correct. Add "Percentages may not total 100 because of rounding." to the note. If a column is far from 100, the cause is usually missing answers mixed into some rows and not others: recheck Percent against Valid Percent.
Means with three or four decimals
Usually copied straight from SPSS or Excel. Round survey means and standard deviations to one decimal and everything else to two, and do it before you build the table, because Word can't round the numbers you type.
Alpha written as 0.89 or .890
Alpha can't exceed 1, so it takes no leading zero, and two decimals are enough: .89. The same goes for correlations and proportions in the same table.
Variable names like BT_M or Q12_rev in the stub
They mean nothing to your reader. Use the construct name from your text ("Brand trust"), the same wording in the table, the text and the figures. For reverse-scored items in an appendix table, add "(R)" and explain it in the note.
Two tables give different values for the same variable
This happens when one table was made before a data-cleaning step and the other after. Rebuild both from the final output, and check the N in each note against your analysis sample.
The table follows the APA rules but your university wants something else
Some thesis guides ask for chapter numbers (Table 4.1), a caption below the table, or a different font in tables. Your university's guide wins over APA, and the APA 7 format guide lists the usual places where a thesis departs from APA. Whatever you choose, make every table the same; if you build them in a template, the APA format template in Word keeps the styles in one place.
08Questions people ask
Can I write M (SD) in one column?
Many journals and theses do, as "5.3 (1.1)", to save width. Separate columns are easier to scan and to align on the decimal point. If your table is short of space, one combined column is acceptable; label it M (SD) and use it the same way in every table.
Do I have to report skewness and kurtosis?
APA doesn't require them. Report them when your method depends on them, for example when you checked normality before a parametric test. If you report them, round to two decimals with a leading zero.
Should I report the standard error instead of the standard deviation?
To describe your sample, use the standard deviation: it shows how spread out the scores are. The standard error describes how precise a mean is, which belongs with inferential results and confidence intervals, not in a descriptive table.
How do I make a descriptive statistics table in Excel?
Load the Analysis ToolPak (Windows: File › Options › Add-ins, then Excel Add-ins and Go; Mac: Tools › Excel Add-ins), tick Analysis ToolPak, then click Data Analysis on the Data tab and choose Descriptive Statistics. It produces a report, not an APA table: copy the values you need, round them, and build the table in Word as in section 05.
Where does the descriptive statistics table go in a thesis?
Usually at the start of the results chapter, before the tests of your hypotheses, with the sample characteristics table in the method chapter. Long item-level tables go in an appendix. Embed each table after the paragraph that first mentions it.
The rules for decimals, leading zeros, italics, minus signs and which symbols need no definition come from APA's 7th Edition Numbers and Statistics Guide (last updated September 2024); the stub, spanners, borders and notes from APA's page on tables. The SPSS menus and output columns were checked against IBM's DESCRIPTIVES reference and two university SPSS guides, and the Word and Excel steps against Microsoft Support pages for Microsoft 365, Windows and Mac. The tables use made-up data from my example thesis, with every total and rounding checked.
Sources
- 7th Edition Numbers and Statistics Guide (last updated September 11, 2024) APA Style
- Tables (stub, column spanners, borders, notes) APA Style
- DESCRIPTIVES, command syntax reference (default and optional statistics) IBM SPSS Statistics
- Descriptive analysis in SPSS (Descriptives and Frequencies menus) Barnard College
- Frequency tables for categorical variables (Frequency, Percent and Valid Percent) Kent State University Libraries
- Convert text to a table or a table to text and Merge or split cells in a table Microsoft Support
- Keyboard shortcuts in Word (Windows: Unicode code and Alt+X) and Insert a symbol (Word for Mac: Advanced Symbol) Microsoft Support
- Use the Analysis ToolPak to perform complex data analysis Microsoft Support