Reporting Pearson Correlation in SPSS
From Output to an APA 7 Matrix
Reporting Pearson correlation in SPSS turns a dense Correlations table into one sentence for a single pair and an APA 7 matrix for several. Each coefficient needs its own N and df, an exact p, an interval and means and SDs that give it scale.
IN SHORT
What One Coefficient Needs Beside It
Pearson’s r describes how strongly, and in which direction, two quantitative variables were linearly associated in the sample, and a report gives it with its degrees of freedom, an exact p-value, an interval and the means and standard deviations that set its scale. For reporting Pearson correlation in SPSS, take the values from the Correlations table (Analyze, Correlate, Bivariate; menu labels vary slightly by release) and place several coefficients in one matrix. Building or checking the matrix from the saved output is work done at OnlineSPSS, a specialist statistical consultancy for academic, clinical and professional research, by an in-house PhD-level statistician, and what an association means for the field is the researcher’s to say.
In This Guide
ANALYSIS SUPPORT
Analysis Support for This Method
Pearson's r answers a narrow question, the strength of a straight-line association. If the data call for ranks, or one variable is better treated as an outcome to be predicted, the analysis is decided first and the report follows.
Correlation Analysis (Pearson & Spearman)
The association between two variables analyzed for your data, from the choice of coefficient to a result sentence that matches the output.
Simple Linear Regression
One predictor modeled against an outcome in its own units, for questions about how much the outcome changes and not only how closely two variables move together.
BEFORE YOU REPORT IT
What the Coefficient Sentence Rests On
Four facts settle which values are quoted and what the sentence may claim, and each one is fixed before any number is copied.
Which Coefficient Pearson r suits two quantitative variables with a roughly straight-line pattern. A curved but steadily rising pattern, or ordinal scores, point to Spearman, and the report names the one used.
One Pair or a Matrix A single pair fits in a sentence, while three or more variables read more clearly as a matrix. Either way, the report states how many coefficients were examined.
Pairwise or Listwise Pairwise exclusion lets N differ from cell to cell, whereas listwise exclusion uses one set of cases throughout, so the report names the rule and gives the N behind each coefficient.
Direction of Scoring A negative r can come from how a scale is scored and not from a lack of association, so the sentence describes direction in the words of the construct and not as a sign alone.
IN SPSS
Where Each Number in the Correlation Report Comes From
Linearity and influential cases are settled first, in Pearson Correlation in SPSS: Checks and Output, and the six values below then come from one run of IBM SPSS.
Take r From the Correlations Table
From the Correlations table, take the Pearson Correlation value for the pair, and note its sign before quoting its size.
Check: Read the sign against how each scale is scored. A negative r between a symptom score and a wellbeing score is coherent, and the sentence should say why.
Work Out df From the Pair's N
Take N for that pair from the same table and subtract 2 to get the degrees of freedom reported beside r.
Watch: With pairwise exclusion, N differs between cells. Use the N printed in the cell of the coefficient being reported and not the size of the dataset.
Take the Exact p-Value
Take the significance value for the pair (Sig. (2-tailed); wording varies by release) and report it to three decimals.
Tip: Stars may flag coefficients in the matrix, but the sentence still gives the exact p for the pair it discusses, since stars belong to the table and not to the text.
Add an Interval for r
Use the interval option of the Bivariate Correlations dialog where your release has one, the Bootstrap option where the license includes it, or a Fisher z interval calculated from r and N.
Check: Say which route gave the interval. A bootstrap interval and a Fisher z interval can differ a little, and a reader who knows only one will otherwise question the other.
Take Means and Standard Deviations
From the Descriptive Statistics table, requested under Options in the dialog, take each variable’s mean and standard deviation.
Tip: Request the descriptives in the same run as the correlations, so the M and SD columns and the coefficients come from one saved output file.
Enter the Values in Sentence and Matrix
Enter each value in the sentence frame and the matrix shell below, then compare both against the saved output.
Check: Compare the sentence with the matrix coefficient by coefficient. A value in both places should carry the same rounding, with r to two decimals and p to three.
READING THE OUTPUT
Each Printed Correlation Value and Its Place in the Report
Each pair appears twice in the printed matrix, with its r, p and N stacked in one cell, so the rows below separate what is reported from what is left out.
| Output value | Where SPSS prints it | Where it goes and what to check |
|---|---|---|
| Pearson Correlation | Correlations table, above and below the diagonal | r in the sentence and the matrix; each pair appears twice, so only one triangle is reported |
| Sig. (2-tailed) | Correlations table, under each r | Exact p to three decimals; a printed .000 becomes p < .001, and the p-value is not read as the size of r |
| N | Correlations table, one per pair | The df as N minus 2, using the N of the pair and not the N of the whole file |
| Asterisks and footnote | Beside r, with a note under the table | A table note giving the alpha levels, if stars are used at all, and the number of tests |
| Mean and Std. Deviation | Descriptive Statistics table | The M and SD columns of the matrix, to the same decimals as the data |
| Confidence interval | The interval output of releases that offer the option (layout varies by release), or the Bootstrap rows | The interval beside r, with the method that produced it named in the text |
| Cross-product Deviations and Covariance | Correlations table, if requested | Left out of the report, since r is the statistic reported |
REPORTING
An APA 7 Sentence, a Matrix Shell and the Size of r
The sentence carries one pair and the matrix carries the rest, and both are filled from the same output so that they cannot disagree.
In Pearson correlation APA format, r is reported with its degrees of freedom in parentheses, r(___) = ___, followed by the interval and the exact p-value, with no zero before the decimal point because r and p cannot exceed 1. The sentence names both variables in the words of the study and gives the direction. For three or more variables, the correlation table APA 7 asks for shows means, standard deviations and the lower triangle only.
A Pearson correlation was computed between [Variable 1] and [Variable 2] for the ___ cases with both scores. The two were [positively / negatively] correlated, r(___) = ___, 95% CI [___, ___], p = ___, so that higher scores on [Variable 1] went with [higher / lower] scores on [Variable 2]. The coefficient indicates a [weak / moderate / strong] linear association by the benchmarks stated in the methods, and no direction of influence is implied.
| Table 1. Means, Standard Deviations and Correlations Among [Study Variables] | |||
|---|---|---|---|
| Variable | M (SD) | 1 | 2 |
| 1. [Variable 1] | ___ (___) | ||
| 2. [Variable 2] | ___ (___) | ___ | |
| 3. [Variable 3] | ___ (___) | ___ | ___ |
| Note. N = ___ (pairwise N ranges from ___ to ___). * p < .05. ** p < .01. The diagonal and upper triangle are left blank. Each coefficient carries the N printed in its own cell of the output. | |||
The coefficient is itself the effect size, so no second measure is needed, but it is named as a Pearson r in the text. Squaring it gives the proportion of variance the two variables share in a linear fit, which is often the easier number to discuss. Verbal labels such as small, medium and large are rough conventions, and a value that is large in one field can be modest in another, so the interval and the values typical of the field carry more weight than a label. In small samples the interval for r is wide, which is why it is reported beside the coefficient.
COMMON MISTAKES
Correlation Statements That Reviewers Query
A reviewer finds these by comparing the manuscript with the output, so that comparison is worth making first.
A Zero That Is Not Zero SPSS prints .000 for any p below .0005, which is a small probability and never zero. The report gives p < .001, and exact values stay exact from .001 upward.
Leading Zeros Kept A coefficient and a p-value cannot exceed 1, so APA 7 drops the zero before the decimal point: .42 and not 0.42. The same applies to every limit of the interval for r.
df Taken From the File The degrees of freedom beside r are the N of the pair minus 2. Quoting the size of the dataset overstates the information behind a coefficient whenever cases are missing on either variable.
Association Described as Effect Verbs such as predicts, influences or leads to claim more than a correlation shows. Observational data support wording about association, and a direction of influence needs a design that can carry it.
Stars Without a Count Asterisks across a large matrix flag chance findings in proportion to the number of pairs. The report states how many coefficients were examined and treats unplanned ones as exploratory.
Size Labels by Rote Calling every r near .30 moderate ignores the field, the measurement error in each scale and the width of the interval, so a label is used only with a benchmark cited in the methods.
Does each cell of the matrix carry the N of its own pair? Independent Statistical Review confirms the N, df and rounding behind every coefficient.
FREQUENTLY ASKED QUESTIONS
Presenting Correlations From SPSS in a Manuscript
A large matrix forces three decisions into the open: how it is laid out, how its p-values are worded and whether r squared accompanies r.
How do I present correlations among many variables?
In one matrix: the lower triangle only, with means and standard deviations in columns beside it and a note that explains any asterisks. Intervals for the key coefficients go in the text or a supplement, because a matrix filled with intervals is hard to read and the sentence can carry the few that matter.
Should the report give r or r squared?
r is the statistic reported, and r squared is a supplementary way of describing its size. The squared value gives the share of variance in one variable associated with the other in a linear fit, which suits a discussion of magnitude. Reporting only r squared hides the direction of the association.
How should p-values in a large matrix be reported?
Exact and unadjusted, with a statement of how many coefficients were examined, unless a correction for multiple tests was planned. If one is applied, the report names it and says whether the p-values shown are adjusted. Coefficients not specified in advance are labeled exploratory.
Can a correlation matrix be rebuilt from my saved output?
The saved Correlations output is enough to rebuild the matrix and the sentence that reports it. An in-house PhD-level statistician at OnlineSPSS drafts both from it, or checks a draft you already have cell by cell, as part of Results Reporting & Interpretation. The meaning of each association in your field is yours to argue.
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
Where in the Output
Does Each r Come From?
A correlation matrix holds many numbers, and each traces to one cell of the Correlations table. A matrix built that way holds up when a reviewer opens the output beside it.
No Commitment Required
Strictly Confidential
Fixed Project Fee
In-House PhD-Level Statisticians
Same-Day Response
Serving Researchers Since 2014