How to Interpret SPSS Output for Your Thesis

Introduction

Most Master’s and PhD scholars can click their way to an SPSS result table. Far fewer can read that table confidently. The columns of numbers, the rows labelled “Sig. (2-tailed),” the cryptic footnotes about Levene’s test — SPSS output is dense, and without a structured way to read it, the typical response is to copy the table into Word, glance nervously at the p-value, and hope for the best. That hope is not a strategy. Examiners can tell.

The good news is that every SPSS procedure produces output in the same predictable structure: a descriptive summary, an assumption check, the test statistic, the p-value, and an effect size. Once you learn to interpret SPSS output systematically — knowing which numbers matter, which to ignore, and how to convert them into an APA-style sentence — reading your results becomes a five-minute job rather than an afternoon of anxiety.

This guide walks through the most common SPSS procedures, shows you the exact tables to read, and gives you an APA-style reporting template for each. For the broader workflow, see our SPSS data analysis guide; for the test-selection logic, see our statistical tests guide. If you would like a specialist to handle the analysis and write-up, our data analysis service covers both.

Anatomy of the SPSS Output Viewer

SPSS sends every procedure to the Output Viewer (the .spv file). The left pane is a navigable tree; the right pane is the actual tables and charts. Learn to scan the tree first — it tells you at a glance which procedures you have run and in what order, which is invaluable when you revisit an analysis weeks later. Most scholars who struggle to interpret SPSS output do so because they read the test tables without first reading the descriptive and assumption tables above them — the context that gives the test statistic its meaning.

Every procedure produces at least two blocks. The first block is the descriptive summary (frequencies, means, sample sizes). The second block is the inferential test (t-statistic, F-ratio, chi-square). Many procedures also produce an assumption-check block in between. Read them in order: descriptive → assumption → test → effect size. Skipping the assumption block is the most common reason scholars report tests they cannot defend.

Descriptive Statistics Output

Run Analyze → Descriptive Statistics → Descriptives for the mean/SD block and Frequencies for the count/percentage block. The Descriptives table reports N, Minimum, Maximum, Mean, and Std. Deviation for each variable. The Frequencies table adds the cumulative percentage, which is useful for ordinal variables.

What to check:

• N: Does the sample size match what you collected? A drop in N across variables signals missing data — investigate before proceeding.
• Minimum and Maximum: Are they within the possible range? An age of 999 means you forgot to declare 99 as a missing-value code.
• Mean and Std. Deviation: For Likert scales, a mean near the midpoint with a moderate SD suggests reasonable distribution; a mean near a floor or ceiling with a small SD suggests skew.

Report descriptives in an APA-style summary table: variable, M, SD, n, possible range, observed range. SPSS produces the raw numbers; format the table yourself rather than pasting the SPSS table verbatim. Examiners notice the difference.

Independent-Samples t-Test Output

Run Analyze → Compare Means → Independent-Samples T Test. The output produces two key tables. The first is Group Statistics — sample size, mean, standard deviation, and standard error of the mean for each group. The second is the Independent Samples Test table, which is where most scholars get lost.

The Independent Samples Test table has two rows: “Equal variances assumed” and “Equal variances not assumed.” Which row you read depends on Levene’s Test for Equality of Variances, which appears in the first two columns of the same table. If Levene’s p-value is greater than 0.05, variances are equal — read the top row. If Levene’s p is 0.05 or below, variances are unequal — read the bottom row (which uses Welch’s correction).

From the correct row, extract four numbers:

• t: the test statistic.
• df: degrees of freedom. Note that the bottom row produces a fractional df — report it to one decimal place.
• Sig. (2-tailed): the p-value. Report the exact value (p = .032, not p < .05).
• Mean Difference: the raw difference between group means, used to compute Cohen’s d.

SPSS does not compute Cohen’s d automatically. Calculate it as d = Mean Difference / pooled SD, or use SPSS’s ESM (Effect Size Mean) syntax, or the free online calculator at Psychometrica. Report in APA style: “An independent-samples t-test showed that Group A (M = 78.4, SD = 6.2) scored significantly higher than Group B (M = 72.1, SD = 7.5), t(118) = 4.82, p < .001, d = 0.88.”

One-Way ANOVA Output

Run Analyze → Compare Means → One-Way ANOVA. The output produces the ANOVA table with columns: Sum of Squares, df, Mean Square, F, and Sig. The rows are “Between Groups,” “Within Groups,” and “Total.” The F-statistic is the ratio of the Between-Groups Mean Square to the Within-Groups Mean Square; a large F indicates that group means differ more than would be expected by chance.

Read the Sig. column for the F-test. If p < 0.05, at least one group mean differs from the others — but ANOVA does not tell you which. To identify the specific group differences, run a post-hoc test. In the One-Way ANOVA dialog, click Post Hoc and select Tukey (for equal variances) or Games-Howell (for unequal variances). The Multiple Comparisons table reports the mean difference, standard error, p-value, and 95% confidence interval for every pair of groups.

SPSS does not report eta-squared by default in One-Way ANOVA. To get it, click Options and check “Descriptive” and “Homogeneity of variance test” — the latter runs Levene’s test, which you need to interpret the F-test correctly. Calculate eta-squared manually as η² = SSbetween / SStotal. Report in APA style: “A one-way ANOVA showed a significant effect of teaching method on exam score, F(2, 117) = 8.43, p < .001, η² = 0.126. Tukey post-hoc tests indicated that the flipped condition significantly outperformed the traditional condition (p = .004).”

Chi-Square Output

Run Analyze → Descriptive Statistics → Crosstabs, then click Statistics and check “Chi-square.” The output produces the Crosstabulation table (observed and expected frequencies for each cell) and the Chi-Square Tests table.

Read the “Pearson Chi-Square” row. Extract three numbers: the chi-square value, df, and Asymp. Sig. (2-sided). The degrees of freedom are (rows − 1) × (columns − 1). A p-value below 0.05 indicates a significant association between the two variables.

Before reporting, check the footnote under the Chi-Square Tests table: “a. X cells (Y%) have expected count less than 5.” If more than 20% of cells have expected count below 5, the chi-square approximation is unreliable — report Fisher’s Exact Test instead (SPSS provides it automatically for 2×2 tables). For effect size, run Crosstabs → Statistics → Phi and Cramér’s V. Report in APA style: “A chi-square test of independence showed a significant association between gender and voting preference, χ²(1, N = 412) = 9.87, p = .002, φ = 0.155.”

Linear Regression Output

Run Analyze → Regression → Linear. The output produces four tables. The Model Summary reports R, R², adjusted R², and the standard error of the estimate. The ANOVA table reports the F-test of overall model significance. The Coefficients table reports the B, standard error, beta, t, and Sig. for each predictor. The optional Collinearity Diagnostics table reports tolerance and VIF.

Read in this order:

1. Model Summary — R² and adjusted R²: How much variance in the outcome is explained by the predictors? Report adjusted R² when there are multiple predictors, since it penalises for predictors that add no explanatory power.
2. ANOVA — F and Sig.: Is the overall model significant? If Sig. > 0.05, the model is no better than predicting the mean; stop.
3. Coefficients — B, beta, t, Sig.: For each predictor, is the contribution significant? B is the unstandardised coefficient (used for prediction); beta is the standardised coefficient (used to compare predictor strength).
4. Collinearity Diagnostics — Tolerance and VIF: Tolerance below 0.20 or VIF above 5 indicates multicollinearity — predictors are too highly correlated.

Check assumptions on residual plots: linearity (residuals vs predicted plot should show no curve), homoscedasticity (residuals evenly spread), normality (P-P plot of residuals should follow the diagonal), and independence (Durbin-Watson statistic should be close to 2). Report in APA style: “A multiple linear regression predicted exam score from study hours and prior GPA, F(2, 197) = 64.2, p < .001, adjusted R² = .39. Both predictors were significant: study hours (B = 2.14, β = .48, p < .001) and prior GPA (B = 8.92, β = .31, p < .001).”

Cronbach’s Alpha (Reliability) Output

Run Analyze → Scale → Reliability Analysis. Move the items of a scale into the box, set Model to Alpha, and click Statistics → Item, Scale, and “Scale if item deleted.” The output produces three tables.

Read the Reliability Statistics table for the overall Cronbach’s alpha. Conventional thresholds: 0.70 acceptable, 0.80 good, 0.90 excellent. Below 0.60, the scale is questionable; below 0.50, unacceptable.

Read the Item-Total Statistics table — specifically the column “Cronbach’s Alpha if Item Deleted.” If deleting an item would raise the overall alpha, the item is weakly related to the others and should be dropped. Use this table to iteratively prune a scale to its best-performing items.

Writing Results in APA Style

SPSS output is not writing. Converting output into prose is the final and most important step. For every test, report four things in a single sentence: what you did, what you found, the statistical evidence, and the effect size. Templates:

• t-test: “An independent-samples t-test showed that [Group A] (M = X, SD = Y) scored significantly higher than [Group B] (M = X, SD = Y), t(df) = X.XX, p = .XXX, d = X.XX.”
• ANOVA: “A one-way ANOVA showed a significant effect of [factor] on [outcome], F(between df, within df) = X.XX, p = .XXX, η² = .XXX. [Post-hoc test] indicated [specific comparison].”
• Chi-square: “A chi-square test of independence showed a significant association between [Var 1] and [Var 2], χ²(df, N = XXX) = X.XX, p = .XXX, φ = .XX.”
• Regression: “A multiple linear regression predicted [outcome] from [predictors], F(df1, df2) = X.XX, p < .001, adjusted R² = .XX. [Predictor 1] was significant (B = X.XX, β = .XX, p = .XXX) and [Predictor 2] was significant (B = X.XX, β = .XX, p = .XXX).”

SPSS output is a starting point, not a destination. The numbers it produces are only as trustworthy as the assumptions you checked before clicking OK — and only as meaningful as the effect sizes and confidence intervals you report alongside the p-values. A thesis chapter that pastes raw SPSS tables without converting them into APA-style prose is not finished, no matter how many numbers are on the page.

Conclusion

Learning to interpret SPSS output is a five-step skill: scan the Output Viewer tree, read descriptives first, check assumptions (especially Levene’s and the expected cell counts), extract the test statistic, df, p-value, and effect size, then convert them into a single APA-style sentence. The procedures differ in their tables but share this structure. Run assumption checks before reporting parametric tests; switch to non-parametric equivalents when assumptions fail; report effect sizes alongside every p-value; never paste raw SPSS tables into your thesis. If you would like a specialist to run the analysis and write up the results, our data analysis service handles both. Pair this guide with our SPSS workflow guide, statistical tests overview, and data analysis chapter template, and reach us through our contact page for a free consultation.

Frequently Asked Questions