How to Report SPSS Results in APA 7 (Without Your Results Chapter Getting Rejected Again)
To report SPSS results in APA 7, state the finding in plain language first, then give the test statistic, degrees of freedom, exact p-value, and an effect size, all inside one sentence. If you are googling how to report SPSS results in APA 7 right now, chances are your results chapter just came back from your supervisor marked “reformat per APA” with no further explanation. I have read that exact comment on close to two hundred results chapters over the last twelve years at Statssy, and in almost every case, the statistics underneath were fine. The write-up around them was not.
This article fixes that. I am going to show you the five rules APA 7 actually enforces, then walk you through bad versus good examples for a t-test, ANOVA, regression, and chi-square, straight from SPSS output. At the end, there is a real rewrite I did for a client, so you can see the before and after.
Why SPSS Results Write-Ups Get Rejected in APA 7 (Even When the Stats Are Right)
Most rejections have nothing to do with whether you ran the correct test. They happen because the paragraph around the number is incomplete or wrongly worded. This is not just my own observation. A similar pattern has been documented in published research on results-section writing errors. Streiner (2007), writing in the Canadian Journal of Psychiatry, found authors reporting p-levels without the actual test values behind them, and spending a sentence on which software was used while skipping the statistical choices that actually matter.
I will go a step further than that paper does. Most online APA tutorials stop at “here is the formula for Cohen’s d,” and leave you to figure out on your own that SPSS does not calculate it for you automatically in a standard t-test output. That gap between the rule and what your actual SPSS screen shows is where most students go wrong in dissertation apa statistics reporting, and it is the gap this article is built to close.
Four mistakes account for almost all the rejections I see, in the same order, across every discipline I consult on.
- Reporting p = .000. SPSS rounds to three decimals, so this should always be written as p < .001. Writing p = .000 tells your committee you do not understand what a p-value represents, and I have seen this single line trigger a full resubmission.
- Missing effect size. APA 7 treats effect size as mandatory, not a bonus. A significant result without Cohen’s d, eta-squared, or Cramer’s V next to it is an incomplete sentence as far as your committee is concerned.
- Degrees of freedom placed wrong, or missing entirely. This is a formatting detail, but examiners notice it instantly because it is the easiest thing to check.
- No confidence interval. For t-tests and regression coefficients, APA 7 expects a 95% CI alongside the point estimate.
None of these require you to touch your analysis again. They are writing fixes, which is exactly why they are so frustrating when a committee flags them late.
The 5 Non-Negotiable Rules of APA 7 Statistical Reporting
Before you touch a single test, get these five rules fixed in your head. Every bad-versus-good example later in this article is really just these five rules applied to a different test.
1. Italicise statistical symbols. t, F, p, M, SD, r, and df are always italicised. The numbers next to them are not.
2. Round to two decimal places, except p-values, which get three (or the “< .001” convention once SPSS shows .000).
3. Report the exact statistic, its degrees of freedom, and the p-value together, in that order, inside one sentence.
4. Always include an effect size. Which one depends on the test, covered below for each. If your data is not normally distributed to begin with, your choice of test and effect size can change before you even get to the writing stage, so it is worth confirming that upfront with a Shapiro-Wilk normality check rather than assuming parametric tests are safe by default.
5. State the finding in plain English first, then the statistics in parentheses. Not the other way around.
What does APA 7 require in a results sentence?
APA 7, the seventh edition of the American Psychological Association’s Publication Manual, requires the test statistic, degrees of freedom, exact p-value (or p < .001), and an appropriate effect size, written in italics where applicable, inside one sentence that states the finding in plain language first.
Quick Reference: What to Report for Each Test
| Test | SPSS table to check | What APA 7 requires | Example |
|---|---|---|---|
| T-test | Independent Samples Test | t, df, p, M, SD, 95% CI, Cohen’s d | t(58) = 2.31, p = .025, d = 0.61 |
| ANOVA | ANOVA table + Effect Size | F, both df, p, eta-squared, post-hoc | F(2, 87) = 5.64, p = .005, η² = .11 |
| Regression | Model Summary, ANOVA, Coefficients | R², F, B, β, t, p, 95% CI per predictor | B = 2.34, 95% CI [1.45, 3.23], p < .001 |
| Chi-square | Chi-Square Tests + Symmetric Measures | χ², df, N, p, Cramer’s V | χ²(3, N = 212) = 14.82, p = .002, V = .26 |
How to Report a T-Test in APA 7
Which SPSS Row to Pull
Open your SPSS output and look at the Independent Samples Test table. There are two rows: Equal variances assumed, and Equal variances not assumed. Check the Levene’s Test significance value first. If it is above .05, use the top row. If it is .05 or below, use the bottom row. If you are not sure which test design even applies to your data, this comparison of paired t-test vs repeated measures ANOVA for pre-post data will settle it before you report anything.
Bad vs Good Example
Rejected version:
“There was a difference between the two groups (t=2.3, p=0.03).”
Accepted version:
“An independent samples t-test revealed that participants in the intervention group (M = 24.5, SD = 3.1) scored significantly higher than the control group (M = 21.9, SD = 3.4), t(58) = 2.31, p = .025, 95% CI [0.34, 4.86], d = 0.61.”
Notice the second version names the groups, gives means and standard deviations, includes the confidence interval, and ends with Cohen’s d. That is what a proper apa write up spss looks like: complete enough that your committee can verify every number against the table you pulled it from.
How to Report ANOVA Results in APA 7
What to Read in SPSS
For a one-way ANOVA, your APA sentence comes from the ANOVA table (F, both df values, and Sig.), plus eta-squared from the Effect Size table, plus a post-hoc result if your overall F was significant. If you are running a repeated-measures ANOVA instead, check Mauchly’s Test of Sphericity first. If sphericity is violated, you report the Greenhouse-Geisser corrected F and df, not the uncorrected ones, and most student write-ups I review skip this step entirely. If your Levene’s test for a standard one-way ANOVA is violated, this Tukey HSD vs Games-Howell comparison explains which post-hoc test to switch to.
Bad vs Good Example
Rejected version:
“ANOVA showed a significant effect, F=5.6, p<0.05.”
Accepted version:
“A one-way ANOVA showed a significant effect of teaching method on exam scores, F(2, 87) = 5.64, p = .005, η² = .11. Tukey post-hoc comparisons showed that the flipped-classroom group (M = 78.3) scored significantly higher than the lecture-only group (M = 71.2, p = .004).”
The second version tells the reader which groups actually differed, which is the one thing a bare F-value can never do.
How to Report Regression Results in APA 7
What to Read in SPSS
Regression write-ups have three SPSS tables to pull from: Model Summary (R² and adjusted R²), ANOVA (model-level F), and Coefficients (B, β, t, and p for each predictor). Before you report any of this, check your variance inflation factor (VIF) values if you have more than two predictors, since multicollinearity will quietly distort your B values without SPSS ever flagging it as an error.
Bad vs Good Example
Rejected version:
“Regression was significant, R2=0.34, and age was a predictor (p<0.01).”
Accepted version:
“A multiple regression was conducted to predict job satisfaction from workload and autonomy, R² = .34, F(2, 97) = 24.95, p < .001. Autonomy significantly predicted job satisfaction, B = 2.34, 95% CI [1.45, 3.23], β = .41, t(97) = 5.20, p < .001, while workload did not, B = -0.12, p = .31.”
Report the model-level statistics first, then every individual predictor, significant or not. Dropping the non-significant predictor is one of the most common mistakes I correct in spss results apa style write-ups, because committees want to see what you tested, not just what worked.
How to Report a Chi-Square Test in APA 7
What to Read in SPSS
Pull your chi-square value, df, and significance from the Chi-Square Tests table, and Cramer’s V from the Symmetric Measures table underneath it.
Bad vs Good Example
Rejected version:
“Chi square was significant (p<0.05).”
Accepted version:
“A chi-square test of independence showed a significant association between department and attrition status, χ²(3, N = 212) = 14.82, p = .002, V = .26.”
The chi-square symbol (χ²) is one of the few APA symbols students consistently skip because it is awkward to type. Word’s symbol menu or Alt+967 on Windows solves this in two seconds.
Case Study: The Rejected Paragraph I Rewrote Last Year
A public health master’s student came to me after her second-round feedback flagged her entire results chapter as “not APA compliant,” with no further detail. Her original t-test paragraph read:
“The results showed t=1.98, p=0.04 which means there was a significant difference.”
Three problems jumped out immediately. No group means were reported, no effect size was given, and the p-value was not aligned with a confidence interval anywhere in the chapter. I rewrote it as:
“An independent samples t-test revealed that nurses who completed the training (M = 4.21, SD = 0.52) reported significantly higher confidence scores than those who did not (M = 3.89, SD = 0.61), t(142) = 1.98, p = .041, 95% CI [0.01, 0.63], d = 0.33.”
Same data. Same test. The second version went back to her committee and was approved without a single further comment on formatting. The statistics did not change at all, only the completeness of the sentence around them.
Your 5-Point Self-Check Before You Resubmit
Run through this before you send the chapter back to your supervisor:
- Is every statistic (t, F, p, M, SD) italicised?
- Does every significant result include an effect size?
- Is p = .000 rewritten as p < .001 everywhere?
- Does the sentence name the groups or variables in plain English before the numbers?
- Is a 95% CI included for every t-test and regression coefficient?
Most of the time, running through this list yourself is enough to turn a rejected chapter into an accepted one. The one exception is when you are unsure whether you even picked the right effect size or test design for your data in the first place, because that is a statistical judgment call a self-check cannot fully catch.
If you can tick all five, your results chapter is in far better shape than most of what crosses my desk.
Getting the statistics and the write-up both right is genuinely two separate skills, and I have watched strong researchers get stuck on the second one simply because nobody taught it to them. If that sounds like where you are right now, here is how to tell if you actually need a dissertation expert rather than another rewrite on your own. Once your results section is accepted, the next hurdle is usually explaining what these numbers mean in your discussion chapter, and defending them in front of your committee, which is a different writing problem entirely.
FAQ
Should I report R² or adjusted R² in APA 7?
Report both when you have them. R² describes your sample, adjusted R² corrects for the number of predictors, and including both lets your committee judge model fit properly.
How do I cite SPSS itself in APA 7?
Cite it as software in your reference list: IBM Corp. (Year). IBM SPSS Statistics for Windows (Version X.X) [Computer software]. IBM Corp. In text, you generally do not need a separate in-text citation for the software itself.
How do I report a non-significant result in APA style?
Report it exactly the same way as a significant one, with the test statistic, df, exact p-value, and effect size. Never write “no significant difference was found” without the full statistics behind it.
How do I report Likert-scale or ordinal results in APA 7?
Treat genuinely ordinal data with non-parametric tests such as Mann-Whitney U or Kruskal-Wallis, and report the U or H statistic with its p-value, rather than means and standard deviations alone.
Do I need to report effect size for every single test in APA 7?
Yes, for every inferential test where a standard effect size exists. This is one of the most common gaps I find in otherwise correct results chapters.
What does “Sig. = .000” actually mean in SPSS output?
It means the true p-value is smaller than .0005, which SPSS rounds down to three decimals. It is never actually zero, and should always be reported as p < .001.
About the author: Siddharth Gupta has spent 12 years guiding dissertation and thesis students through statistical analysis and APA reporting, through Statssy. Connect on LinkedIn.
Reference: Streiner, D. L. (2007). A Shortcut to Rejection: How Not to Write the Results Section of a Paper. Canadian Journal of Psychiatry, 52(6), 385-389.