SPSS Z-Test for Independent Proportions Tutorial

A z-test for independent proportions tests if 2 subpopulations
score similarly on a dichotomous variable.
Example: are the proportions (or percentages) of correct answers equal between male and female students?

Z-Test Independent Proportions

Although z-tests are widely used in the social sciences, they were only introduced in SPSS version 27. So let's see how to run them and interpret their output. We'll use exam-questions.sav -partly shown below- throughout this tutorial.

SPSS Exam Questions Variable View

Now, before running the actual z-tests, we first need to make sure we meet their assumptions.

Z-Test - Assumptions

Z-tests for independent proportions require 2 assumptions:

Regarding this second assumption, Agresti and Franklin (2014)2 propose that both outcomes should occur at least 10 times in both samples. That is,

$$p_a n_a \ge 10, (1 - p_a) n_a \ge 10, p_b n_b \ge 10, (1 - p_b) n_b \ge 10$$


Note that some other textbooks3,4 suggest that smaller sample sizes may be sufficient. If you're not sure about meeting the sample sizes assumption, run a minimal CROSSTABS command as in crosstabs v1 to v5 by sex. As shown below, note that all 5 exam questions easily meet the sample sizes assumption.

SPSS Z-Test Sample Sizes Check Output

For insufficient sample sizes, Agresti and Caffo (2000)1 proposed a simple adjustment for computing confidence intervals: simply add one observation for each outcome to each group (4 observations in total) and proceed as usual with these adjusted sample sizes.

SPSS Z-Tests Dialogs

First off, let's navigate to Analyze SPSS Menu Arrow Compare Means SPSS Menu Arrow Independent-Samples Proportions and fill out the dialogs as shown below.

SPSS Z-Test Independent Proportions Dialogs

Clicking “Paste” results in the SPSS syntax below. Let's run it.

*Z-tests for independent proportions (requires SPSS 27+).


SPSS Z-Test Output

SPSS Z-Test Independent Proportions Descriptives Output

The first table shows the observed proportions for male and female students. Note that female students seem to perform somewhat better: a proportion of .768 (or 76.8%) answered correctly as compared to .720 for male students.

SPSS Z-Test Independent Proportions Confidence Intervals Output

The second output table shows that the difference between our sample proportions is -.048.

The “normal” 95% confidence interval for this difference (denoted as Wald) is [-.141, .044]. Note that this CI encloses zero: male and female populations performing equally well is within the range of likely values.

I don't recommend reporting the Agresti-Caffo corrected CI unless your data don't meet the sample sizes assumption.

SPSS Z-Test Independent Proportions Significance Output

The third table shows the z-test results. First note that p(2-tailed) = .309. As a rule of thumb, we reject the null hypothesis if p < 0.05 which is not the case here. Conclusion: we do not reject the null hypothesis that the population difference is zero. That is: the sample difference of -.048 is not statistically significant.

Finally, note that SPSS reports the wrong standard error for this test. The correct standard error is 0.0475 as computed in this Googlesheet (read-only) shown below.

Z-Test And Confidence Interval Independent Proportions In Googlesheets

SPSS Z-Tests - Strengths and Weaknesses

What's good about z-tests in SPSS is that

However, what I really don't like about SPSS z-tests is that

APA Reporting Z-Tests

The APA guidelines don't explicitly mention how to report z-tests. However, it makes sense to report something like the difference between males and females
was not significant, z = -1.02, p(2-tailed) = .309.
You should obviously report the actual proportions and sample sizes as well. If you analyzed multiple dependent variables, you may want to create a table showing


  1. Agresti, A & Caffo, B. (2000). Simple and Effective Confidence Intervals for Proportions and Differences of Proportions. The American Statistician, 54(4), 280-288.
  2. Agresti, A. & Franklin, C. (2014). Statistics. The Art & Science of Learning from Data. Essex: Pearson Education Limited.
  3. Twisk, J.W.R. (2016). Inleiding in de Toegepaste Biostatistiek [Introduction to Applied Biostatistics]. Houten: Bohn Stafleu van Loghum.
  4. Van den Brink, W.P. & Koele, P. (2002). Statistiek, deel 3 [Statistics, part 3]. Amsterdam: Boom.

Tell us what you think!

*Required field. Your comment will show up after approval from a moderator.