To read any SPSS output table, find the value named after your test (t, F, r, or Chi-Square) as the test statistic, then read the p-value in the Sig. column, called Two-Sided p in SPSS version 27 and later. If that p-value is below your alpha, usually .05, the result is significant. This page shows exactly where to look in every common table, with worked examples you can match to your own output.
SPSS gives you tidy tables, but a table is not an answer. The marks come from knowing which single value in the table decides your result, and how to write it. Almost every SPSS table follows the same logic: a test statistic that measures the size of the pattern, a degrees of freedom value that reflects the sample, and a p-value that says whether the pattern is likely to be real. Learn to spot those three, and you can read output from a test you have never seen before. This guide is part of our SPSS help and the wider statistics homework help hub.
One version note that trips people up. In SPSS version 27 and later the old Sig. (2-tailed) column is split into One-Sided p and Two-Sided p. For an ordinary non-directional hypothesis you read Two-Sided p, which holds the same meaning as the old Sig. (2-tailed).
Four questions open up almost every output table SPSS produces. Ask them in order.
| Ask this | Where to look | Why it matters |
|---|---|---|
| Which value is the test statistic? | The column named after the test: t, F, r, or Value in the Pearson Chi-Square row. In regression, B and t on each predictor row. | It measures the size of the effect in test units. On its own it does not tell you significance. |
| What are the degrees of freedom? | The df column. A t-test has one, ANOVA and regression report two, chi-square shows rows minus one times columns minus one. | They go in brackets after the statistic and reflect the sample behind the test. |
| Which value is the p-value? | The Sig. column, or Sig. (2-tailed), or Two-Sided p in version 27 and later. | Below your alpha (usually .05) means statistically significant. This is the decision value. |
| How big is the effect? | An effect size such as Cohen's d, eta squared, Cramer's V, r, or R Square. | Significance says an effect exists; effect size says whether it is large enough to matter. |
The rest of this page applies these four questions to the five tables you meet most often.
The single most common reading error is treating the Sig. value as the result and forgetting the statistic, or the reverse. They answer different questions. The test statistic (t, F, r, or Chi-Square) measures how far your data sit from what the null hypothesis predicts, in the test's own units. The p-value translates that distance into a probability, so you can judge whether the pattern is likely to be chance. A complete result reports both, plus the degrees of freedom, for example t(38) = 2.43, p = .020. Report the statistic without the p-value and the reader cannot judge significance. Report the p-value alone and the result is not reproducible.
Significance is not the same as importance. A p-value below .05 tells you an effect is unlikely to be zero, not that it is large or useful. In a big sample even a trivial difference turns out significant. That is why every result below pairs the p-value with an effect size, which measures how large the difference or relationship actually is.
Each recreated table highlights the value that decides the result, with a one-line read and a link to the full walkthrough. Example values follow standard teaching datasets.
Compares one continuous outcome between two unrelated groups. The result lives in the Independent Samples Test table.
Read it. Let Levene's Test choose the row first: if Levene's Sig. is above .05 read Equal variances assumed, otherwise read Equal variances not assumed. On that row, t is the statistic, df is the degrees of freedom, and Two-Sided p is the p-value.
| Levene F | Levene Sig. | t | df | Two-Sided p | |
|---|---|---|---|---|---|
| Equal variances assumed | 1.57 | .217 | 2.428 | 38 | .020 |
| Equal variances not assumed | 2.428 | 34.9 | .021 |
Full menu path, Levene's rule, and APA write-up: see the t-test walkthrough.
Compares one continuous outcome across three or more groups. The omnibus result lives in the ANOVA table.
Read it. Use the Between Groups row. F is the statistic, its two df values are the between-groups df and the within-groups df, and Sig. is the p-value. A significant F says at least one group differs, so you then read the Multiple Comparisons table to see which pairs.
| Sum of Squares | df | Mean Square | F | Sig. | |
|---|---|---|---|---|---|
| Between Groups | 85.5 | 2 | 42.7 | 4.467 | .021 |
| Within Groups | 258.3 | 27 | 9.6 | ||
| Total | 343.8 | 29 |
Post-hoc reading, Levene's, and effect size: see the ANOVA walkthrough.
Predicts a continuous outcome from one or more predictors. Regression prints several tables, and you read three of them together.
Read it. The Model Summary gives R Square, the share of variance in the outcome the model explains. The ANOVA table Sig. tests whether the model as a whole predicts the outcome. The Coefficients table is where the work happens: B is the unstandardized slope, Beta is the standardized slope for comparing predictors, t is each predictor's statistic, and its Sig. tests that predictor. Read the model Sig. first, then each predictor's Sig.
| Model | R | R Square | Adjusted R Square | Std. Error of the Estimate |
|---|---|---|---|---|
| 1 | .699 | .489 | .470 | 3.35 |
| Model | B | Std. Error | Beta | t | Sig. |
|---|---|---|---|---|---|
| (Constant) | 2.86 | 1.44 | 1.99 | .055 | |
| Hours studied | 1.53 | 0.31 | .699 | 4.94 | <.001 |
Assumptions, the ANOVA table, and full reporting: see the regression walkthrough.
Tests whether two categorical variables are associated. The result lives in the Chi-Square Tests table.
Read it. Use the Pearson Chi-Square row and ignore the Likelihood Ratio and Linear-by-Linear rows. Value is the statistic, df equals rows minus one times columns minus one, and Asymptotic Sig. (2-sided) is the p-value. Check the footnote for expected counts below five.
| Value | df | Asymptotic Sig. (2-sided) | |
|---|---|---|---|
| Pearson Chi-Square | 7.864 | 2 | .020 |
| Likelihood Ratio | 7.912 | 2 | .019 |
| N of Valid Cases | 120 |
Menu path, expected-count rule, and Cramer's V: see the chi-square section of the SPSS hub.
Measures the strength and direction of the linear relationship between two continuous variables. The result lives in the Correlations matrix.
Read it. Read an off-diagonal cell, never the diagonal ones, which are always 1. Pearson Correlation is r, from minus one to plus one, Sig. (2-tailed) is the p-value, and N is the number of complete pairs. Here r doubles as the effect size.
| Height | Jump distance | ||
|---|---|---|---|
| Height | Pearson Correlation | 1 | .706 |
| Sig. (2-tailed) | .005 | ||
| N | 14 | 14 |
Scatterplot check, Spearman fallback, and reporting: see the correlation section of the SPSS hub.
A p-value answers whether an effect exists. An effect size answers how big it is, in a unit that does not depend on sample size. Report one next to every p-value. SPSS now prints many effect sizes directly, and each test has its usual measure with rough benchmarks from Jacob Cohen.
| Test | Usual effect size | Small | Medium | Large |
|---|---|---|---|---|
| Independent or paired t-test | Cohen's d | 0.20 | 0.50 | 0.80 |
| One-way ANOVA | Eta squared | 0.01 | 0.06 | 0.14 |
| Correlation and regression | r, and R Square for the model | 0.10 | 0.30 | 0.50 |
| Chi-square | Phi (two by two) or Cramer's V | 0.10 | 0.30 | 0.50 |
These cutoffs are conventions, not laws, and some fields expect larger effects than others. The point stands: a significant result with a tiny effect size is often less useful than a non-significant result in a study that was simply too small, so read the two together.
Output layouts on this page were checked against the IBM SPSS Statistics documentation, the UCLA annotated SPSS output, and the Kent State University SPSS tutorials.
A statistician can run and explain it on your data, confirm the test fits your design, read the correct rows, and write the result so you can defend it. This is done-with-you help, not a black box.
Upload your SPSS file or output, the brief, and your deadline, then get a free quote from support.
Approve the price and a statistician who works in SPSS every day starts right away.
Receive the read output, effect sizes, and write-up in your account, with free revisions if you need them.
"This company is wonderful. I was able to depend on them all semester for my Statistics assignments. Dependable with competitive prices. Would definitely use again!"
"I was a little skeptical at first but this is definitely the best decision I could have made when it comes to getting help with my homework. This is a legit business and the staff are very helpful and professional."
Sig. is the p-value, the probability of a result at least this extreme if the null hypothesis were true. If it is below your alpha, usually 0.05, the result is statistically significant. In a t-test or correlation the column is Sig. (2-tailed), renamed Two-Sided p in version 27 and later. Sig. is never the test statistic itself.
SPSS rounds any p-value smaller than .0005 down to .000, but a probability is never exactly zero. Report it as p < .001, not p = .000.
The value named after the test: t for a t-test, F for ANOVA and regression, Pearson Correlation for correlation, and Value in the Pearson Chi-Square row. In regression each predictor has its own B and t. The p-value sits in a separate Sig. or Two-Sided p column beside it.
Effect size measures how large a difference or relationship is, apart from significance. Common Cohen benchmarks: Cohen's d of 0.2 small, 0.5 medium, 0.8 large; r of 0.1, 0.3, 0.5; eta squared of 0.01, 0.06, 0.14. Report one beside every p-value, because a tiny effect can still be significant in a large sample.
Regression gives two. The Sig. in the ANOVA table tests the whole model, and the Sig. on each predictor's row in the Coefficients table tests that predictor. Read the model Sig. first, then the per-predictor values.
The df column describes the sample behind the test and goes in brackets after the statistic. A t-test uses one df, ANOVA and regression report two, and chi-square df equals rows minus one times columns minus one. After a Greenhouse-Geisser correction the df become fractional.
Yes. Send your output file or data set and a statistician reads the correct rows, confirms the test was appropriate, checks the assumptions, and writes the interpretation in your required style before your deadline, with free revisions.
Hand it to a statistician and get the correct rows, effect sizes, and a clear interpretation before your deadline.
Get my free quote →