Hypothesis testing - Jamovi's naming convention

Discuss the jamovi platform, possible improvements, etc.
Post Reply
wasia
Posts: 3
Joined: Tue Dec 29, 2020 5:21 pm

Hypothesis testing - Jamovi's naming convention

Post by wasia »

Hi!
Jamovi is a great piece of software. I just started to use it and I have one simple (I think) question.
Performing for example paired samples t-test we can select "Hypothesis" (see attached image) and here's my problem.
What kind of hypothesis it is? Null or alternative? I think it is the alternative hypothesis but wanted to ask to be sure.
Attachments
Jamovi-H1.png
Jamovi-H1.png (89.03 KiB) Viewed 3471 times
User avatar
jonathon
Posts: 2613
Joined: Fri Jan 27, 2017 10:04 am

Re: Hypothesis testing - Jamovi's naming convention

Post by jonathon »

hi yes, it's the alternative hypothesis.

kind regards

jonathon
wasia
Posts: 3
Joined: Tue Dec 29, 2020 5:21 pm

Re: Hypothesis testing - Jamovi's naming convention

Post by wasia »

OK, thanks for the confirmation. This may look like a small thing but it is important.
Thank you.
User avatar
reason180
Posts: 274
Joined: Mon Jul 24, 2017 4:56 pm

Re: Hypothesis testing - Jamovi's naming convention

Post by reason180 »

However, in fact, the t test is necessarily a test of the null hypothesis (the alternative hypothesis plays no role unless one is computing a Bayes factor). Correct?
wasia
Posts: 3
Joined: Tue Dec 29, 2020 5:21 pm

Re: Hypothesis testing - Jamovi's naming convention

Post by wasia »

Well, as far as I understand the concept of hypothesis testing if you test for null hypothesis you test also for alternative one. Hope you understand what I'm trying to say.
User avatar
reason180
Posts: 274
Joined: Mon Jul 24, 2017 4:56 pm

Re: Hypothesis testing - Jamovi's naming convention

Post by reason180 »

A classical t test is a null-hypothesis test. For a given sample size, the p value is the probability that you would obtain an effect-size at least as large as the one you obtained, if the null hypothesis were true.

If you try to take the result and draw an inference relating to the alternative hypothesis, you're on very thin ice, conceptually. That's why people are turning to Bayesian analysis instead.
User avatar
reason180
Posts: 274
Joined: Mon Jul 24, 2017 4:56 pm

Re: Hypothesis testing - Jamovi's naming convention

Post by reason180 »

Perhaps, technically, it should be:

Measure 1 = Measure 2
Measure 1 not < Measure 2
Measure 1 not > Measure 2
Post Reply