SPSS command: SPSS PROPORTIONS
Tests and confidence intervals for proportions in your browser
The PROPORTIONS procedure tests and estimates binomial proportions in three designs: one sample against a test value, two paired measurements on the same cases, and two independent groups.
Free · No account · Files stay on your device · Updated 18 October 2026
What it tells you
Crosstabs runs it on your device and prints every test and interval method side by side, from Wald and score tests to exact binomial, Wilson, Clopper-Pearson, Newcombe and Agresti-Caffo.
- Whether one proportion differs from a target value, with score, Wald, exact binomial and mid-p tests.
- Whether a proportion changed between two measurements on the same cases (McNemar-type tests).
- Whether two independent groups have different proportions, with pooled and unpooled Wald tests and Hauck-Anderson.
- A confidence interval for the proportion or difference by every method the panel offers, so you can see how much the choice matters.
When to use it
Use it when
- A share from a survey or audit compared with a benchmark: 62% satisfied against a 50% target.
- Before-and-after yes/no outcomes on the same respondents, such as awareness before and after a campaign.
- A/B tests and two-group comparisons of conversion, response or defect rates.
Use something else when
- More than two groups or a multi-category outcome: use a chi-square test in Crosstabs.
- You need to adjust for covariates: use binary logistic regression.
- Data come from a weighted, clustered survey: use Complex Samples Crosstabs, since sampling weights are not accepted here.
How to run it in Crosstabs
- 01
Open your data
Open a file with one row per case and switch to Statistics mode.
- 02
Choose Proportions
In the Agreement and diagnostics group, pick Proportions (PROPORTIONS).
- 03
Pick the design
Set Design to One sample, Paired samples or Independent samples, then choose the Value that counts as a success.
- 04
Choose the variables
One sample: tick up to 20 Test variables and set the Test value (proportion), default 0.5. Paired samples: tick the Two measurements. Independent samples: tick one test variable, choose the Grouping variable, then Group 1 and Group 2.
- 05
Set the confidence level and run
Choose the Confidence level (%), declare any weight as a frequency weight, and press Run.
Worked example
Computed by the Crosstabs engine from the synthetic data below when this page was built — not typed in.
Data
A synthetic checkout experiment: 120 visitors saw variant A and 115 saw variant B. Each visitor converts at random with probability .30 (A) or .44 (B), drawn with a fixed seed; "converted" is the success value.
| Group | Successes | Trials | Proportion |
|---|---|---|---|
| A | 30 | 120 | 0.250 |
| B | 47 | 115 | 0.409 |
| Difference | −0.159 |
| Test Type | Z | One-Sided p | Two-Sided p |
|---|---|---|---|
| Hauck-Anderson | −2.539 | .006 | .011 |
| Wald | −2.622 | .004 | .009 |
| Continuity-corrected Wald | −2.481 | .007 | .013 |
| Wald (H0) | −2.591 | .005 | .010 |
| Continuity-corrected Wald (H0) | −2.452 | .007 | .014 |
Two-sided p is twice the one-sided p, capped at 1.
| Interval Type | Point Estimate | 95% CI Lower | 95% CI Upper |
|---|---|---|---|
| Agresti-Caffo | −0.159 | −0.274 | −0.038 |
| Brown-Li Jeffreys | −0.159 | −0.276 | −0.039 |
| Hauck-Anderson | −0.159 | −0.282 | −0.035 |
| Newcombe | −0.159 | −0.273 | −0.039 |
| Continuity-corrected Newcombe | −0.159 | −0.279 | −0.033 |
| Wald | −0.159 | −0.277 | −0.040 |
| Continuity-corrected Wald | −0.159 | −0.286 | −0.032 |
Bounds outside [−1, 1] are truncated.
How to read it
Variant A converts 25.0% of visitors and variant B 40.9%, a difference of −0.159 (Newcombe 95% CI −0.273 – −0.039). The Wald test gives Z = −2.62, two-sided p = .009. Lower bounds of the 7 interval methods range from −0.286 to −0.273 and upper bounds from −0.040 to −0.032, and every one excludes 0. With small groups or proportions near 0 or 1 the methods diverge more, and Newcombe or Agresti-Caffo are safer than the plain Wald interval.
Output tables explained
- Proportions
- The summary for each variable or pair: successes, trials, observed proportion and test value (one sample); both proportions, the difference and the four cells (paired); or successes, trials and proportion per group with the difference (independent).
- Hypothesis Tests
- Every test for the chosen design with Z where applicable, one-sided and two-sided p. Continuity corrections shrink the difference toward 0.
- Confidence Intervals
- The point estimate with lower and upper bounds for each method: nine for one sample (including Wilson, Clopper-Pearson, Jeffreys, Agresti-Coull), five for paired (including Newcombe, Agresti-Min, Bonett-Price) and seven for independent samples (including Newcombe, Agresti-Caffo, Hauck-Anderson).
Assumptions and notes
- Each case contributes one binary outcome (success or not) per variable, independently of other cases.
- Wald-type tests and intervals need reasonably large counts; with small samples or proportions near 0 or 1, prefer exact, Wilson or Newcombe methods.
- Paired design: the two measurements are on the same cases; independent design: the groups do not overlap.
How the results are validated
One-sample score and Wald z match a direct computation to 1e-12; exact and mid-p one-sided p values match SciPy 1.18 binomial probabilities to 1e-10; Agresti-Coull, Clopper-Pearson, Jeffreys, Wilson and Wald intervals match statsmodels 0.15 proportion_confint to 1e-9; Wilson with and without continuity correction match R prop.test to 1e-12. Independent-sample Wald, pooled z, Agresti-Caffo, Newcombe and Wald intervals match statsmodels and R PropCIs 0.3.0 to 1e-9 or better; paired McNemar z, exact p and the Newcombe interval match an independent implementation. Validation evidence
Questions
Which confidence interval should I report?
For one proportion, Wilson score is a good default and Clopper-Pearson if you need guaranteed coverage. For a difference, Newcombe (independent or paired) or Agresti-Caffo (independent) perform well. Avoid the plain Wald interval with small counts.
Is the paired test the same as McNemar's test?
Yes. The paired design's McNemar row is the square root of the McNemar chi-square with the sign of the change, and the exact binomial row is the exact McNemar test on the discordant pairs.
Why do Wald (H0) and Wald give different Z for two groups?
Wald (H0) uses the pooled proportion under the null, which matches the usual two-proportion z test and the chi-square test; Wald uses each group's own proportion.
What is the mid-p test?
The exact binomial p value minus half the probability of the observed count. It is less conservative than the exact test while keeping close to nominal error rates.
Related procedures
- Crosstabs StatisticsCROSSTABS /STATISTICS
- One-Sample Nonparametric TestsNPAR TESTS
- Binary Logistic RegressionLOGISTIC REGRESSION
- Complex Samples CrosstabsCSTABULATE
- Free SPSS alternative: full coverage list
Run Proportions on your own data
Open a .sav, CSV or Excel file. Everything runs in your browser; the file never leaves your device.
SPSS is a trademark of IBM Corp. Crosstabs is not affiliated with or endorsed by IBM.