jamovi

Hi all

Loving Jamovi. Thanks.

One thing that annoys me in SPSS is that a test of the difference between means is labelled as Mean difference. Unfortunately this error is repeated in Jamovi (see attached screen shot).

This is clearly incorrect and an error. When we do a paired t-test, we work with the differences (say, Before minus After), and hence are basing a test on the mean difference. But with a independent samples t-test, we are not interested in a mean difference (indeed, it makes no sense), but a difference between the means of the two groups.

Can this simple textual error (and it is an error) be fixed? That is, change the text Mean difference in the attached to Diff. between means or similar?

Thanks.
P.

hi,

i'm comfortable with 'difference between means' being described as 'mean difference'. for example, i'd be comfortable with someone describing the difference between the heights of men and women as 'height difference'.

jonathon

Thanks jonathon.

You say:

i'd be comfortable with someone describing the difference between the heights of men and women as 'height difference'.

Perhaps... but it is the location of the word "mean" in the output that I find annoying and misleading, and the word "mean" does not appear in your example. But in any case, surely that is not really the issue; the issue is accuracy and clarity.

I just see it as a very simple thing to change, that would make teaching so much easier.

Students already can find it hard to distinguish between the two-sample t-test ("a difference between the two means"), and a paired t-test ("a mean of the differences"). The latter is about the mean of the differences, so when the software seems to contradict all you say to help then distinguish the tests, it just seems unnecessarily frustrating.

Perhaps we agree to disagree. But I think such a simple change makes for far greater clarity and much easier and consistent teaching (from the jamovi's Features page: "jamovi's ease of use makes it ideal for introducing people to statistics"). To me, therefore, there seems to be no good reasons not to change it.

But thanks.

P.

yup, i'm sympathetic to what you suggest. the only reason i'm reluctant is because "Diff. between means" is long, and stretches the table out a lot (we owe part to our success to being anal when it comes to how the program looks and behaves). even 'mean difference' is a bit too long for my comfort. in fact, i'm tempted to simply call it mu!

but i appreciate you raising the issue.

cheers

jonathon

So the main issue is the length of the text?

What about "Diff. means" as a shorthand for "Difference between means"? That's even shorter than what is there at present. Or "Diff. b/t means"?

P.

One thought on this: What about replacing "Mean difference" with "Difference: means".

The length of the text is essentially the same (objection no longer relevant), and probably that bit clearer.

Thoughts?

P.

PS: Sorry; this still bugs me!

P.

so what if we made this an addon? you install it from the jamovi library, and it changes the text in the IS t-test?

jonathon

Well, that's certainly an option, and better than nothing! I imagine it would be reasonable simple to do too.

Thanks. (However, I still, I cannot see any disadvantage to changing "Mean difference" to "Difference: means"...)

P.