The alternate hypothesis choices for the paired and independent groups situations in the t test module seem to imply a null hypothesis of 0. Some studies may hypothesize that a pre-post difference of at least 5 or more is hypothesized, for example. I see no way to specify this a priori null hypothesis mu
Purpose
To permit the user to specify a non-zero mu (test value) for the null hypothesis in the paired and independent groups t tests, which is readily available in R code.
Use-case
t.test(MyData$After, MyData$Before, mu=5, alternative="greater", paired=T) for the paired t test. A similar specification for mu in the independent groups situation would also be expected.
Is your feature request related to a problem?
No
Describe the solution you would like
Offer the user a choice to add a non-zero null hypothesis value (test value) for the paired and independent groups t test.
Describe alternatives that you have considered
I have looked through the possible specifications in the t test module. I can do the pre-post analysis or test a hypothesized group mean difference using the one-sample t test on the difference scores, but adding a test value option in the paired and independent groups would be helpful. It would not immediately be obvious to most users to approach the analysis using a one-sample test.
Description
The alternate hypothesis choices for the paired and independent groups situations in the t test module seem to imply a null hypothesis of 0. Some studies may hypothesize that a pre-post difference of at least 5 or more is hypothesized, for example. I see no way to specify this a priori null hypothesis mu
Purpose
To permit the user to specify a non-zero mu (test value) for the null hypothesis in the paired and independent groups t tests, which is readily available in R code.
Use-case
t.test(MyData$After, MyData$Before, mu=5, alternative="greater", paired=T) for the paired t test. A similar specification for mu in the independent groups situation would also be expected.
Is your feature request related to a problem?
No
Describe the solution you would like
Offer the user a choice to add a non-zero null hypothesis value (test value) for the paired and independent groups t test.
Describe alternatives that you have considered
I have looked through the possible specifications in the t test module. I can do the pre-post analysis or test a hypothesized group mean difference using the one-sample t test on the difference scores, but adding a test value option in the paired and independent groups would be helpful. It would not immediately be obvious to most users to approach the analysis using a one-sample test.
Additional context
No response