How to Report Pearson Correlation in APA from SPSS (2026 Example + Template)
Confused by your SPSS Correlations table? Here is exactly how to interpret Pearson correlation from SPSS — r, p, N, df — and report it in APA 7th style with copy-paste templates, an APA table, and significant vs. non-significant examples.
If you are writing a psychology, nursing, business or education dissertation, you will almost certainly need to answer two questions: how to interpret Pearson correlation from SPSS, and how to report Pearson correlation in APA style without losing marks. Examiners look for four things: correct r(df) format, exact p value, direction and strength language, and a clean APA table when you have three or more variables.
This 2026 guide gives you all of it: SPSS clicks, how to read the output, Cohen (1988) effect sizes, a copy-paste template sentence, before-and-after table conversion, and significant, non-significant and multiple-correlation examples. If you already know how to report SPSS results in APA 7th generally, this is the correlation-specific companion.
- Quick-answer APA template (copy-paste)
- Step 1: Run Pearson correlation in SPSS
- Step 2: Interpret r, p, and N from SPSS output
- Step 3: Effect size, scatterplot and assumptions
- Step 4: SPSS to APA correlation table
- Step 5: In-text write-up examples
- Common mistakes that cost marks
- Checklist + FAQs
Standard APA 7th format for Pearson r: r(df) = .xx, p = .xxx. Round r to two decimals, p to three decimals, drop leading zeros, italicise r and p. Degrees of freedom = N – 2.
Example: A Pearson correlation revealed a significant large positive relationship between perceived stress and sleep disturbance, r(100) = .52, p < .001.
df explained: If SPSS shows N = 102 pairs, df = 102 − 2 = 100, so you write r(100). If SPSS shows Sig. (2-tailed) = .000, write p < .001 — never p = .000.
What Does Pearson Correlation Actually Tell You?
Pearson's r measures the strength and direction of a linear relationship between two continuous variables (interval or ratio, e.g., Likert-scale totals, GPA, hours studied, burnout score). It ranges from −1 to +1:
- +1 = perfect positive line: as X rises, Y rises.
- 0 = no linear relationship.
- −1 = perfect negative line: as X rises, Y falls.
Crucially, correlation is not causation. A significant r = .62 between social media use and anxiety does not prove social media causes anxiety — a third variable, reverse causation, or sampling bias could explain it. Your dissertation discussion must use associative language: "associated with," "related to," "predicted by" (in a statistical sense only), never "caused" or "led to" from correlation alone.
Before running Pearson, confirm your data are suitable: two continuous variables, roughly linear scatterplot, no extreme outliers, and ideally bivariate normality for small samples. If your variables are ordinal or badly skewed, consider Spearman's rho instead — but that is a separate write-up.
Step 1: How to Run Pearson Correlation in SPSS
Use SPSS 27, 28, 29 or 30 — menus are identical. The path is Analyze > Correlate > Bivariate.
1.1 Click-by-click
- Open your cleaned dataset. If you have not yet screened missing values and reverse-scored items, see how to clean survey data in SPSS first.
- Go to Analyze > Correlate > Bivariate...
- Move your two (or more) scale variables from the left box into Variables on the right. Example: Stress_Total and SleepDisturb_Total.
- Under Correlation Coefficients, tick Pearson. Untick Kendall's tau-b and Spearman unless your supervisor asked for them.
- Under Test of Significance, select Two-tailed (default for dissertations — use one-tailed only if you pre-registered a directional hypothesis).
- Tick Flag significant correlations so SPSS adds * and **.
- Click Options... > tick Means and standard deviations (needed for your descriptives and APA table) > Missing Values: leave Exclude cases pairwise unless you have a reason to use listwise. Click Continue.
- Optional but recommended for APA 7th: click Bootstrap... or, in SPSS 27+, use Correlations > Confidence Intervals to request 95% confidence intervals for r. Many examiners now expect CIs.
- Click OK. Your output appears in the Output Viewer as a Correlations table.
1.2 SPSS syntax (for your appendix)
Keep this in your methods appendix to show reproducibility:
CORRELATIONS /VARIABLES=Stress_Total SleepDisturb_Total /PRINT=TWOTAIL NOSIG FULL /MISSING=PAIRWISE.
For confidence intervals in newer versions, add /CI CILEVEL=95 via the dialog paste button.
Also produce a scatterplot to verify linearity: Graphs > Legacy Dialogs > Scatter/Dot > Simple Scatter, put one variable on X and one on Y, then double-click the chart > Add Fit Line. For full descriptives reporting, pair this with how to report descriptive statistics from SPSS.
Stuck on Correlations? Send Your Output for 24h Interpretation
Paste your SPSS Correlations table — we return a clean APA paragraph + table, effect-size interpretation, and assumption check within 24 hours.
Step 2: How to Interpret Pearson Correlation from SPSS Output
Here is a typical SPSS output for N = 102 nursing students (fictional teaching data):
| SPSS row | Stress_Total | SleepDisturb_Total | What it means |
|---|---|---|---|
| Pearson Correlation | 1 | .524** | Your r = .52. Positive, large effect. |
| Sig. (2-tailed) | — | <.001 (shown as .000) | Your p value. < .05 = significant. |
| N | 102 | 102 | Pairs with valid data. df = 102 − 2 = 100. |
Read three numbers only — everything else is duplication because the matrix mirrors itself:
2.1 r — direction and strength
The Pearson Correlation cell is r. Sign = direction (+ means both rise together; − means one rises as the other falls). Absolute value = strength. The diagonal of 1.00 (a variable with itself) is ignored — never report it.
2.2 Sig. (2-tailed) — the p value
Sig. (2-tailed) is your p value for the two-tailed test. If p < .05, the correlation is statistically significant (unlikely under the null of zero correlation). SPSS flags * for p < .05 and ** for p < .01. If SPSS prints .000, it means p < .0005 rounded — report as p < .001.
2.3 N — sample size and df
N is the number of complete pairs for that cell. APA does not report N inside the r() parentheses — it reports degrees of freedom, df = N − 2. So N = 102 → r(100). If you used pairwise deletion and different cells have different Ns, calculate df separately for each correlation you report in text.
Worked reading: r = .524, p < .001, N = 102 → In APA: r(100) = .52, p < .001. Interpretation: significant large positive relationship — higher stress scores co-occur with greater sleep disturbance.
Step 3: Effect Size, Scatterplot and Assumptions
Examiners award marks for interpretation, not just significance. Always add an effect-size label and a sentence about assumptions.
3.1 Cohen (1988) thresholds for r
| |r| (absolute value) | Cohen (1988) label | Plain-English example | r² (shared variance) |
|---|---|---|---|
| .10 – .29 | Small / weak | r = .18: slight tendency, limited practical value alone | 1–8% |
| .30 – .49 | Medium / moderate | r = .38: noticeable relationship worth discussing | 9–24% |
| .50 – 1.0 | Large / strong | r = .62: substantial co-movement | 25%+ |
These are conventions, not laws. In clinical nursing or education research, even r = .25 can be practically important with large samples or costly outcomes. Report r itself as the effect size (APA 7th §3.6) and optionally add r²: "r = .52, indicating 27% shared variance." Some fields cite Gignac & Szodorai (2016) (.10 small, .20 moderate, .30 large for individual differences) — if you do, name the source explicitly so the examiner knows why your labels differ.
3.2 What to say about assumptions
You do not need a full page, but one sentence earns trust: "Preliminary checks confirmed linearity via scatterplot, with no extreme outliers, and approximately normal distributions (Shapiro-Wilk p > .05 / skewness < |1|)." If you found a curvilinear pattern or an influential outlier, say what you did (e.g., retained with sensitivity analysis, or switched to Spearman) rather than hiding it. For survey totals built from Likert scales, briefly note reliability (Cronbach's α) from your earlier analysis — it reassures the reader the variables are meaningful.
3.3 Report the 95% CI when you can
APA 7th strongly encourages confidence intervals. If SPSS gave you 95% CI [.36, .66] for r = .52, append it: r(100) = .52, p < .001, 95% CI [.36, .66]. The CI shows precision: a wide interval (e.g., [.05, .70] with N = 25) warns the reader the estimate is unstable even if p < .05.
Step 4: SPSS Correlation Table in APA Format (Before → After)
SPSS output is not submission-ready. It repeats every correlation twice, includes the useless 1.00 diagonal, and uses "Sig." instead of p. Convert it:
What to change
- Remove the upper triangle and diagonal — keep only the lower triangle.
- Replace "Pearson Correlation / Sig. / N" stacking with a single r value per cell plus asterisks.
- Add numbered variables, M and SD columns, and a note defining asterisks and sample size.
- Use horizontal rules only (APA table style), no vertical gridlines.
SPSS-style (do NOT paste this into your thesis):
| Stress | Sleep | Burnout | |
|---|---|---|---|
| Stress: Pearson / Sig. / N | 1 / — / 102 | .524 / .000 / 102 | .471 / .000 / 101 |
| Sleep: Pearson / Sig. / N | .524 / .000 / 102 | 1 / — / 102 | .385 / .000 / 101 |
| Burnout: Pearson / Sig. / N | .471 / .000 / 101 | .385 / .000 / 101 | 1 / — / 101 |
APA 7th version (use this):
| Variable | M | SD | 1 | 2 | 3 |
|---|---|---|---|---|---|
| 1. Perceived stress | 22.40 | 5.10 | — | ||
| 2. Sleep disturbance | 14.85 | 4.32 | .52** | — | |
| 3. Burnout | 31.20 | 7.64 | .47** | .39** | — |
Note. N = 101–102 due to pairwise deletion. *p < .05. **p < .01 (two-tailed).
Title above the table (APA style): Table 1
Pearson Correlations Among Stress, Sleep Disturbance, and Burnout. In Word, use Insert > Table, apply "no vertical borders," italicise M, SD, p, and r only where they appear in the note. For one or two correlations, skip the table entirely and report in text — APA advises against duplicating numbers in both places.
Step 5: Pearson r APA Write-Up Examples (Copy and Adapt)
Replace bracketed text, keep punctuation and italics exactly. Note APA details: r to two decimals, p to three decimals (or < .001), no leading zero before the decimal for r and p (e.g., .52 not 0.52), space around = sign.
Example A — Significant positive (most common)
Why it works: test named, strength (large per Cohen), direction (positive), full statistic, plain-English interpretation without causal claim.
Example B — Significant negative
Note the minus sign before .37 with no leading zero. Some style guides prefer an en-dash; a hyphen-minus is accepted if consistent.
Example C — Non-significant (equally important)
Never write "no correlation" or "proved no relationship" — a non-significant r means insufficient evidence for a linear relationship, not proof of zero. Report the exact p (.382) and the small effect size.
Example D — Multiple correlations in one paragraph
When df differs across pairs (100 vs. 99), it signals pairwise N differences — explain once in the table note rather than repeating N each time. For full results-chapter flow, see reporting SPSS results in APA 7th.
Example E — Dissertation-ready paragraph (business example)
"To test H1, a Pearson correlation was computed to assess the linear relationship between advertising spend and monthly revenue (N = 15). Preliminary scatterplot inspection supported linearity with no extreme outliers. There was a significant large positive correlation between the two variables, r(13) = .71, p = .003. Higher advertising spend was associated with higher revenue, explaining approximately 50% of shared variance (r² = .50)."
This version adds hypothesis linkage, assumption sentence, and r² — exactly what markers reward at master's and doctoral level.
Stuck on Correlations? Send Your Output for 24h Interpretation
Get an examiner-ready APA paragraph, formatted Table 1, and plain-English meaning of every r and p — returned within 24 hours.
7 Common Mistakes When Reporting Pearson Correlation in APA
| ❌ Mistake | ✅ Fix | Why it matters |
|---|---|---|
| Writing p = .000 | Write p < .001 | p is never exactly zero; .000 is SPSS rounding |
| Missing df: r = .52 | r(100) = .52 | df = N − 2 is required in APA in-text |
| Leading zero: r = 0.52 | r = .52 | APA omits zero when value cannot exceed 1 |
| Claiming causation: "stress causes poor sleep" | "stress was associated with poor sleep" | Correlation alone cannot establish cause |
| Reporting 1.00 diagonal | Omit self-correlations | Trivial and clutters tables |
| Duplicating text + table numbers | Highlight key effects in text, full matrix in table | APA §6.43: do not repeat statistics |
| No effect-size language | Add small / moderate / large + r² | Significance ≠ importance, especially with large N |
Two subtler errors: reporting one-tailed p without pre-registration (stick to two-tailed unless your proposal specified direction), and forgetting to italicise r and p while leaving the Greek ρ or subscript s for Spearman unitalicised in the right places. Examiners notice formatting consistency quickly.
Pre-Submission Checklist
- Scatterplot checked for linearity and outliers before interpreting r?
- Each in-text result has r(df) to two decimals + exact p to three decimals or < .001?
- df = N − 2 calculated per pair (not total dataset N if pairwise deletion)?
- Direction (positive/negative) + Cohen strength + plain-English meaning stated?
- Table (if 3+ variables) has M, SD, lower triangle only, * note, and title?
- No causal verbs, no p = .000, r/p italicised, spaces around =?
- Methods section names Pearson, two-tailed alpha (.05), and software (e.g., IBM SPSS Statistics 29)?
A model methods sentence: "Bivariate Pearson correlations (two-tailed, α = .05) were computed in IBM SPSS Statistics (Version 29) to examine linear associations among stress, sleep disturbance, and burnout." Pair it with clean descriptives from your descriptives write-up and your results chapter will read as one coherent story.
Frequently Asked Questions
Use the form r(df) = .xx, p = .xxx with r to two decimals and p to three decimals. Example: A Pearson correlation revealed a significant moderate positive relationship between sleep quality and GPA, r(98) = .42, p < .001. Italicise r and p, drop leading zeros, and add direction + strength.
Read Pearson Correlation for r (direction + strength), Sig. (2-tailed) for the p value, and N for sample size. If p < .05 the result is significant. Convert N to df (N − 2) for APA. Always check the scatterplot first — Pearson only captures linear patterns and is sensitive to outliers.
df = N − 2, where N is the number of complete pairs for that correlation. N = 102 gives r(100). With pairwise deletion, each pair can have a different N, so calculate df separately per correlation.
Report it as p < .001. SPSS rounds to three decimals, so .000 means the true value is below .0005, not zero. Writing p = .000 is an automatic formatting mark lost.
One or two correlations: text only. Three or more variables: add an APA correlation matrix with means, SDs, numbered variables, lower-triangle r values and a *p < .05, **p < .01 note. Do not repeat every number in both text and table — highlight key findings in text and refer to the table.
Use Cohen (1988): .10 = small, .30 = moderate, .50 = large. Quote Cohen explicitly. Report r itself as the effect size and optionally r² for shared variance. With very large samples tiny r values can be significant — always separate statistical significance from practical importance.
Final Thoughts
Reporting Pearson correlation in APA from SPSS boils down to one habit: run clean, read three numbers, write one precise sentence. Run via Analyze > Correlate > Bivariate (two-tailed Pearson), read r for direction and Cohen strength, Sig. for p, and N for df = N − 2 — then write "A Pearson correlation revealed..." with r(df) = .xx, p = .xxx.
Build a proper APA matrix when you have multiple variables, add 95% CIs where possible, and never let a significant p tempt you into causal language. Do that, and your correlations section will satisfy psychology, nursing, business and education examiners alike. If you want a second pair of eyes before submission, send your output — a 24-hour check is far cheaper than a corrections round.
