One-sided significance test for comparing two population proportions

[Jantonie]

Hi,

In order to compare two proportions, we have to build a 2 x 2 contingency table with Jamovi. The performed test is a chi-square test that corresponds to a two-sided test (H_1: p_1 \neq p_2). Would it be possible to test a unilateral hypothesis as well (H_1: p_1 < p_2 or H_1: p_1>p_2) as with R (option alternative of prop.test: c(“two.sided”, “less”, “greater”)?

Sincerly

[jonathon-love]

oh yup,

In order to compare two proportions, we have to build a 2 x 2 contingency table

this would be a 1 x 2 contingency table, yeah?

but there's not really a way you can do this with more than two proportions (with a hypothesis on one of those proportions)? so wouldn't you just do it with a binomial test?

with thanks

jonathon

[Jantonie]

Currently, with Jamovi, to perform a one-sided significance test for comparing two population proportions, I perform a Chi square test on a 2 x 2 contingency table and I divide the p-value by 2 (see appendix for more detail). Complement_180305.pdf

Sincerely

Jantonie

[jonathon-love]

ah! sorry i misunderstood!

yes, you're right, that would be good to add.

we'll see what we can do.

jonathon

[Jantonie]

Thank you very much for your answer. Sincerely Jantonie