TOSTER update (backwards compatibility)

Everything related to the development of modules in jamovi

by arcaldwell49 » Tue Apr 20, 2021 5:49 pm

Hi everybody, I have been working on updating Daniel Lakens TOSTER package (with his permission), and part of that is updating the TOSTER jamovi module.

Part of the idea I had was adding more text output that provides more context/interpretation of the results and more visualizations. For this to occur I had to limit the paired samples TOST and the independent samples TOST to only have 1 dependent variable/pair.

It just occurred to me that this may cause backwards compatibility issues for users opening old jamovi files that included analyses from the old TOSTER module. Would this be an issue? If so, what would you recommend as the best approach for incorporating the new additions? I can send the work-in-progress version of the new TOSTER module if you want. Thanks jamovi team!

Cheers,

Aaron
arcaldwell49
 
Posts: 9
Joined: Wed Oct 11, 2017 7:03 pm

by jonathon » Tue Apr 20, 2021 11:50 pm

hi aaron,

yeah, that will break compatibility with opening old jamovi files. they will still open, the user will be able to view the results, but if they click on analysis to look at its options, i expect something not quite right will happen -- exactly what i'm not sure (it will depend on a few things, whether you kept the option name the same, etc.) it's an empirical question.

the user might need to specify the dependent and independent variables afresh.

in the case of TOSTER, there's not a lot of ambiguity about how the analysis was specified, so i wouldn't be too concerned about breaking the small amount of reproducibility here.

where are you guys at with superpower? maybe after toster you could look into adding that to jamovi :)

cheers

jonathon
User avatar
jonathon
 
Posts: 1729
Joined: Fri Jan 27, 2017 10:04 am

by arcaldwell49 » Wed Apr 21, 2021 12:57 pm

Phew, okay, that is a huge relief. I had a sudden "oh shit" moment when I was working on TOSTER over the weekend.

Also, yes, I am still thinking about the Superpower's module for jamovi. The main issue I keep coming back to is how to make the functions flexible enough as to be usable... The main issue I noticed was getting a flexible way to enter the means for each particular design. However, I thought about it over the weekend, and I think we could just have the means entered via the data window (so people would just enter the means and select the columns).

Even if we don't get a working parallel for the ANOVA_exact/ANOVA_power functions I am thinking that getting the morey_plots function (power as a function of the effect size and degrees of freedom) would be very useful. :slightly_smiling_face:

Cheers!
arcaldwell49
 
Posts: 9
Joined: Wed Oct 11, 2017 7:03 pm


Return to Module development