2
$\begingroup$

Supposing we have 0,1 labeled samples, when I using SPSS to perform a binomial test, and to see if case 0 and case 1 has equal proportion(0.5), the software automatically pick a two-tailed binominal test, but the problem is how can I perform a test to see if the case 1 proportion equals to 0.1 or not.

Since, when I select the testing proportion to 0.1, the software automatically switches to 1-tailed binomial test, and this will give me an alterative hypothesis like how the case 1 propotion likely to be less than 0.1 (and a p value).

How can I switch the test to two-tailed test when the testing proportion is not equals to 0.5?

Anyone can help?

$\endgroup$
2
  • $\begingroup$ In the case that the sample is in the direction of the alternative, e.g. 15/20 and the alternative is greater than 0.5, you can simply divide the 2-tailed p-value SPSS spits out by two. In the obverse case though, e.g. 5/20, and the alternative is greater than 0.5 I'm not sure how to calculate that p-value off-hand. $\endgroup$
    – Andy W
    Commented Jun 13, 2016 at 12:27
  • $\begingroup$ maybe this can help stats.stackexchange.com/questions/203693/… $\endgroup$
    – user83346
    Commented Jun 13, 2016 at 13:07

0