Reporting a Chi-Square Test in SPSS
From Crosstabs Tables to APA 7 Wording
Reporting a chi-square test in SPSS means taking a handful of values from the Crosstabs tables and placing each in a sentence or a table. The test row, the N and the percentage direction are where a report drifts from the output, and reporting Cramér's V or phi means naming the measure, not only giving its value.
IN SHORT
What a Chi-Square Report Has to Contain
A chi-square report states whether two categorical variables were associated in the sample, how strongly and in which direction. Its core is the statistic with its degrees of freedom and N, an exact p-value, a named effect size and the percentages that carry the finding. For reporting Cramér’s V and the rest of a chi-square result in SPSS, run Analyze, Descriptive Statistics, Crosstabs, tick Chi-square, Phi and Cramér’s V (and Risk for a 2×2 table), and display cell percentages in the direction of the comparison. Menu labels can differ slightly between versions. OnlineSPSS, a specialist statistical consultancy for academic, clinical and professional research, prepares and checks this reporting through an in-house PhD-level statistician.
In This Guide
ANALYSIS SUPPORT
Analysis Support for This Method
Reporting begins once the test is settled. When the choice of test, or the analysis itself, is still open, these two pages set out the support available.
Chi-Square Test
The association between two categorical variables analyzed for your design, from the choice of test to a result sentence that matches the output.
Fisher's Exact Test
The exact p-value that replaces the Pearson line when counts are small, worked through for tables of any size.
BEFORE YOU REPORT IT
What the Sentence Depends On
Four facts are fixed before any number is copied, because each one changes which value is quoted or how it is worded.
Which Test Row Decide from the expected counts, before looking at p, whether the Pearson row or an exact test is the one to quote, and say which in the text.
Table Size A 2×2 table can carry phi or an odds ratio as its effect size, while a larger table calls for Cramér's V, so the shape of the table settles the measure.
Direction of the Percentages Choose the percentages that answer the question, by row or by column, and say which. The sentence about direction is only readable against that choice.
Which N The N beside the statistic is the number of valid cases in the table, which is smaller than the sample recruited when either variable has missing values.
IN SPSS
Which Value to Take From Which Table
IBM SPSS prints everything below from one Crosstabs run, and the setup behind that run, including the expected-count check, is covered in the guide to the chi-square test in SPSS.
Take the Statistic, df and N
From the Chi-Square Tests table, take N of Valid Cases, the Pearson Chi-Square value, df and Asymptotic Significance (2-sided), which older releases label Asymp. Sig. (2-sided).
Check: Compare N of Valid Cases with the sample you recruited. If it is smaller, say in the text how many cases were lost to missing values on either variable.
Check the Expected-Count Footnote
Read the footnote under that table, and quote the Fisher row in a 2×2 table when it shows sparse expected counts.
Check: Note how many cells have an expected count below 5 and what the smallest expected count is. The footnote gives both, and the exact-test sentence in the frame below asks for them.
Choose Phi or Cramér's V
From the Symmetric Measures table, take phi or Cramér’s V, whichever the table size calls for.
Tip: In a 2×2 table phi and Cramér’s V match in absolute value, so naming one measure across all tables keeps a manuscript consistent, with the direction given in words.
Take the Odds Ratio and Its CI
For a 2×2 table, take the Odds Ratio with its 95% confidence interval from the Risk Estimate table.
Watch: An odds ratio is not a ratio of probabilities. When the outcome is common, use the word “odds” in the sentence and avoid saying that one group was that many times more likely.
Copy Counts and Percentages
From the Crosstabulation, take the counts and the row or column percentages that carry the finding.
Tip: Set the percentages under Cells to match the comparison you describe in words, and take counts and percentages from the same run, so the sentence and the table share one source.
Place Each Value in the Report
Place each value in the sentence and the table shell below, and keep the output file beside the draft.
Check: Compare the paragraph with the table shell value by value. A statistic that appears in both places should carry the same rounding in each, for example p to three decimals.
READING THE OUTPUT
Where Each Value in the SPSS Crosstabs Output Goes
Every value has one destination in the report, and this table pairs each printed element with its place in the sentence or the table and the slip to avoid there.
| Output value | Where SPSS prints it | Where it goes and what to check |
|---|---|---|
| Pearson Chi-Square, Value and df | Chi-Square Tests table | The statistic and its degrees of freedom, reported together in the sentence |
| Asymptotic Significance (2-sided), or Asymp. Sig. (2-sided) in older releases | Chi-Square Tests table | The p-value to two or three decimals; a printed .000 is reported as p < .001 |
| N of Valid Cases | Last row of the Chi-Square Tests table | N in the sentence and the table note, even when it is below the sample recruited |
| Exact Sig. (2-sided) | Fisher’s Exact Test row, 2×2 tables | Replaces the Pearson p when the footnote shows sparse counts, and the test is named in the text |
| Footnote on expected counts | Under the Chi-Square Tests table | A table note or a methods sentence whenever it shaped the choice of test |
| Phi and Cramér’s V, Value | Symmetric Measures table | The effect size, named by type; phi has a sign that follows the coding, V does not |
| Odds Ratio, Lower and Upper | Risk Estimate table, 2×2 only | The ratio with its 95% CI, after reading the row label to see which category is the reference |
| Count and % within | Crosstabulation | The n (%) cells of the table shell, in the direction of the comparison |
REPORTING
An APA 7 Paragraph, a Table Shell and a Named Effect Size
The frame shows where each printed value lands, and the table shell mirrors the sentence so that the two cannot drift apart.
In chi-square test APA format, the statistic is reported with its degrees of freedom and N together, χ²(___, N = ___) = ___, and the sentence names the test row used, the effect size and the direction of the difference in words. The frame below is for a 2×2 table. A larger table drops the odds ratio and reports Cramér’s V, as in the note below.
A chi-square test examined whether [variable 1] was associated with [variable 2]. The association was [statistically significant / not statistically significant], χ²(___, N = ___) = ___, p = ___, φ = ___, odds ratio = ___, 95% CI [___, ___]. [Group A] had the higher proportion of [category] (___%) compared with [Group B] (___%). Where an exact test was quoted instead, use: Fisher’s exact test, p = ___, with expected counts below 5 in ___ of ___ cells.
| Table 1. [Outcome] by [Group], n (%) | |||
|---|---|---|---|
| [Outcome category] | [Group A] n (%) | [Group B] n (%) | Total n (%) |
| [Category 1] | ___ (___%) | ___ (___%) | ___ (___%) |
| [Category 2] | ___ (___%) | ___ (___%) | ___ (___%) |
| Total | ___ | ___ | ___ |
| Note. χ²(___, N = ___) = ___, p = ___, φ = ___. Percentages are calculated within [rows / columns]. | |||
The effect size is named in the text and in the table note. Phi suits a 2×2 table, and its sign follows the coding of the two variables, so the direction is also given in words. Reporting Cramér’s V for a larger table follows the same frame without the odds ratio, χ²(___, N = ___) = ___, p = ___, V = ___; V runs from 0 to 1, and benchmarks for small and large depend on the smaller of rows and columns minus one, which is stated beside V when a benchmark is used. The Symmetric Measures table prints no interval for either measure, so an interval comes from the Bootstrap option in Crosstabs where the license includes it, or from another program.
COMMON MISTAKES
Reporting Slips That Change What the Result Says
Each slip is found by laying the printed tables beside the finished sentence and the table, one value at a time.
p = .000 Copied SPSS rounds a very small p-value to .000, which is not a probability of zero. The report gives p < .001 and keeps exact values for anything larger.
Wrong Row Quoted Quoting the Pearson line where the footnote calls for an exact test, or the continuity correction without naming it, reports a different test from the one the data support.
N of the Table Left Out A sentence that gives only the sample recruited hides the cases lost to missing values. The N beside the statistic is the N of Valid Cases for that table.
Percentages Without a Direction Row percentages quoted where the groups form the columns answer a different question, so the report says which direction was used and the table follows it.
Effect Size Left Unnamed A bare value after the p-value leaves a reader unable to tell phi from Cramér's V or to read the sign of phi, so the measure is named in the sentence.
Significance Without Direction Reporting that the association was significant without saying which group had the higher proportion leaves the finding unusable, since the test itself shows no direction.
Paragraph drafted, one row still in doubt? Independent Statistical Review treats the printed tables as the reference and tests every reported value against them.
FREQUENTLY ASKED QUESTIONS
Questions About Reporting a Chi-Square Test in SPSS
A significant larger table raises the question of which cells drive it, an odds ratio raises the question of direction, and two large-sample statistics raise the question of which to quote.
How do I report which cells differ in a larger table?
Report them as follow-up comparisons, labeled as such, using adjusted standardized residuals or column comparisons with an adjusted p-value. Both options sit under Cells in Crosstabs. Residuals beyond about 1.96 turn up by chance when many cells are inspected, so the text states how many were examined and treats the result as exploratory unless planned.
Which way does the odds ratio point in my SPSS output?
The row label in the Risk Estimate table tells you: the two compared categories appear in parentheses, and the odds are those of the first category of the column variable. A value below 1 means lower odds in the first-named group. To reverse the direction, recode the variable or take the reciprocal, which also swaps the two limits of the interval.
Can I report the likelihood ratio instead of Pearson's statistic?
Both test the same hypothesis and agree closely when counts are adequate. Pearson’s is the conventional choice in most fields, and the likelihood ratio is more usual when a table comes from a log-linear model. The statistic is chosen before the p-values are compared and is named in the text, for example the likelihood ratio chi-square with its df and N.
Is a drafted chi-square paragraph checked against the printed tables?
It is. An in-house PhD-level statistician at OnlineSPSS sets the paragraph and the table shell against the Crosstabs output, confirms the row quoted and the effect size named, or prepares the report from the saved output. This stage sits under Results Reporting & Interpretation, and what the association means for your field is for you to say.
How do you protect my research data and confidentiality?
Your data and project details are handled confidentially, and project information is not published. Researcher names and institutions are not used publicly without permission. An NDA is available where required and can be arranged before sensitive materials are shared. Read more about confidentiality and data security.
What happens when I request an assessment, and how long does it take?
You share your research question, design and relevant materials, such as the dataset or the output you have, and a statistically trained Research Coordinator reviews the request and responds the same day. The analysis, deliverables, timeline and one fixed project fee are then confirmed before any work begins, so the timeline follows from the agreed scope.
Still have a question? Find more answers in our full FAQ.
NEXT STEP
Does the Paragraph Match
the Crosstabs Tables?
Each value in a chi-square paragraph has a printed source in the Crosstabs output. Those sources can be matched to a paragraph already drafted, or a report can be built from the saved tables.
No Commitment Required
Strictly Confidential
Fixed Project Fee
In-House PhD-Level Statisticians
Same-Day Response
Serving Researchers Since 2014