Academic Support

How to Read SPSS Output and Report It in APA 7: A Table-by-Table Guide (2026)

SPSS prints several tables for every test, and most of those numbers never belong in your results chapter. This reference shows which table to read, which row and column matter, and how to write the result in APA 7 style, test by test.

Before you start

Five rules that apply to every SPSS table

“Sig.” is the p-value. SPSS labels it Sig., Sig. (2-tailed) or Asymptotic Significance. Unless your hypothesis was registered as one-tailed, report the two-tailed value.
Never write p = .000. SPSS rounds to three decimals. When it shows .000, report p < .001.
Report exact p-values to two or three decimals (p = .016), not just p < .05.
No leading zero for statistics that cannot exceed 1, such as p, r and Cronbach’s α (write .84, not 0.84). Italicise statistical symbols: M, SD, t, F, r, p.
Always add an effect size. A significant p-value says an effect is unlikely to be zero, not that it is large. APA 7 expects an effect size and, where possible, a confidence interval.

Choosing the test comes first. If you are unsure which one fits your design, read how to choose the right statistical test before running anything. The example figures in this guide are illustrative, not real data.

Assumptions

Tests of Normality and Descriptive Statistics

Descriptives table: read N, Mean and Std. Deviation for each variable or group. Minimum and maximum are useful for spotting data-entry errors (a 7 on a 5-point scale), but they rarely go in the write-up.

Tests of Normality table: SPSS shows Kolmogorov-Smirnov and Shapiro-Wilk. Shapiro-Wilk is generally preferred, particularly for smaller samples. A Sig. value below .05 suggests the distribution departs from normal. With large samples, even trivial departures become significant, so check the Q-Q plot and the skewness and kurtosis values as well before switching to a non-parametric test.

APA 7 example: Job-satisfaction scores (M = 3.62, SD = 0.71) were approximately normally distributed, W(150) = 0.99, p = .214.

Scale reliability

Reliability Statistics (Cronbach’s alpha)

Read the single value under Cronbach’s Alpha in the Reliability Statistics table, with N of Items. Values of .70 or above are the common convention for acceptable internal consistency, though the threshold depends on your field and the purpose of the scale.

In the Item-Total Statistics table, the column Cronbach’s Alpha if Item Deleted shows whether one item is pulling reliability down. Do not delete items just to raise alpha. Remove an item only if there is also a content reason, and report it if you do.

APA 7 example: The six-item job-satisfaction scale showed good internal consistency (α = .84).

Two groups

Independent Samples t-Test: read Levene’s test first

The Independent Samples Test table has two rows. Which one you report depends on Levene’s Test for Equality of Variances on the left:

Levene’s Sig. above .05: read the Equal variances assumed row.
Levene’s Sig. at or below .05: read the Equal variances not assumed row (Welch’s t-test; the df will have decimals).

From the correct row, take t, df and Sig. (2-tailed), and the mean difference with its 95% confidence interval. Recent SPSS versions (27 and later) also print an Independent Samples Effect Sizes table; report Cohen’s d from its Point Estimate column.

APA 7 example: Employees who completed the training (M = 3.81, SD = 0.66) reported higher job satisfaction than those who did not (M = 3.47, SD = 0.72), t(98) = 2.45, p = .016, d = 0.49.

If the data are clearly non-normal with small groups, the non-parametric alternative is the Mann-Whitney U test; report U, z and p.

Three or more groups

One-Way ANOVA and post hoc tests

Test of Homogeneity of Variances: check Levene’s statistic first. If variances are unequal (Sig. at or below .05), report Welch’s ANOVA from the Robust Tests of Equality of Means table and use Games-Howell post hoc tests.

ANOVA table: the Between Groups row gives F, its df and Sig.; the Within Groups row gives the second df. Eta squared is Between Groups Sum of Squares divided by Total Sum of Squares, or request it under Options (SPSS 27 and later print an ANOVA Effect Sizes table).

Multiple Comparisons table: only if the overall F is significant, read which pairs differ, usually Tukey HSD when variances are equal. The asterisk marks pairs with p below .05.

APA 7 example: Job satisfaction differed by sector, F(2, 147) = 5.21, p = .007, η² = .07. Tukey post hoc tests showed that private-sector employees scored lower than government employees (p = .005); no other pairs differed.

Relationships

Correlations (Pearson’s r)

The Correlations table is a matrix, so every value appears twice. For each pair, read Pearson Correlation (r), Sig. (2-tailed) and N. Degrees of freedom are N − 2. Ignore the diagonal of 1s.

A common guide treats r of about .10 as small, .30 as medium and .50 as large, but interpret strength in the context of your field. Correlation does not show causation, so avoid words such as “affects” or “leads to” in the discussion.

APA 7 example: Perceived supervisor support was positively correlated with job satisfaction, r(148) = .42, p < .001.

Chi-Square Test of Independence

In the Chi-Square Tests table, read the Pearson Chi-Square row: Value, df and Asymptotic Significance (2-sided). Then read the footnote under the table: it states how many cells have an expected count below 5. If more than 20% do, the chi-square result is unreliable; for a 2 × 2 table report Fisher’s Exact Test instead.

For effect size, request Phi and Cramér’s V under Statistics and read the Symmetric Measures table.

APA 7 example: Gender was associated with preference for remote work, χ²(1, N = 200) = 6.35, p = .012, Cramér’s V = .18.

Prediction

Multiple Linear Regression: three tables, in order

Model Summary: R² is the proportion of variance in the outcome explained by all predictors together. Adjusted R² corrects for the number of predictors; report it alongside R², especially with several predictors or a small sample.
ANOVA table: tests whether the model as a whole predicts better than no model. Report F, both df (Regression and Residual) and Sig.
Coefficients table: for each predictor, B (unstandardised) is the change in the outcome for a one-unit change in the predictor, holding the others constant; Beta (β) is the standardised value for comparing predictors; t and Sig. test each predictor. If you requested collinearity diagnostics, VIF values well above 5 (some texts use 10) signal multicollinearity.

APA 7 example: The three predictors explained 31% of the variance in job satisfaction, R² = .31, adjusted R² = .30, F(3, 196) = 29.35, p < .001. Supervisor support was the strongest predictor (β = .35, t = 5.12, p < .001).

Check the assumptions too: residual plots for linearity and homoscedasticity, a histogram or P-P plot of residuals for normality, and the Durbin-Watson value in the Model Summary for independence of errors.

Quick reference

Which number goes in your write-up

Test	SPSS table to read	Report
Descriptives	Descriptive Statistics	M, SD, N
Normality	Tests of Normality (Shapiro-Wilk)	W(df), p
Reliability	Reliability Statistics	α, number of items
Independent t-test	Independent Samples Test (row set by Levene’s)	t(df), p, d
One-way ANOVA	ANOVA, then Multiple Comparisons	F(df1, df2), p, η²
Pearson correlation	Correlations	r(df), p
Chi-square	Chi-Square Tests + Symmetric Measures	χ²(df, N), p, V
Multiple regression	Model Summary, ANOVA, Coefficients	R², F(df1, df2), p, β, t

Reporting the numbers correctly is only the first half of a results chapter. Explaining what they mean for your research questions is where most revisions come from: see how to interpret data results in a dissertation. For a broader walkthrough of running the analysis itself, read our SPSS data analysis guide for UAE students.

Questions

SPSS output questions

What does Sig. mean in SPSS output?+

Sig. is the p-value: the probability of a result at least as extreme as yours if the null hypothesis were true. Report it as p, to two or three decimals, and write p < .001 when SPSS shows .000.

Which row do I report in an independent samples t-test?+

Check Levene’s test first. If its Sig. value is above .05, report the Equal variances assumed row; if it is .05 or below, report the Equal variances not assumed row.

Should I report R squared or adjusted R squared?+

Report both. R squared is the variance explained by the model; adjusted R squared corrects for the number of predictors and is the fairer figure when you have several predictors or a small sample.

What is an acceptable Cronbach’s alpha?+

.70 or above is the common convention for acceptable internal consistency, but the right threshold depends on your field and the purpose of the scale. Very high values, above about .95, can indicate redundant items.

Next step

Want your analysis checked before your supervisor sees it?

We run or review SPSS, AMOS, SmartPLS, NVivo and R analyses, check assumptions, and explain every table so you can defend the results yourself. You remain the author of your chapter.

See data analysis support Ask Labeeb AI Contact Labeeb

Related guides