Overdispersion

[florianhartig]

No description provided.

[florianhartig]

@EttnerAndreas - can you have a look if you understand why this branch suddenly has a build error? This was created from the last update in master, but master checks out, and the branch previously also checked out!

[florianhartig]

p.s., Andreas, can you do a review of this? This mostly concerns the code in the functions that are used by users, which are not so major, it's mainly a slight change in the standard overdispersion function, and the introduction of this parametric version of the overdispersion function. I would like to merge this all asap and get 0.3.4 on the way

[EttnerAndreas]

good job
looks like it is fixed, I would be interested if you changed something specific

currently I am reviewing the new fucntions

[florianhartig]

No, I didn't fix anything. Maybe this was just a travis problem, I didn't look at the error message, sometimes one just gets build errors because the tests run too long, and that can happen because the servers are overloaded, if you trigger a re-build, the error is gone!

OK, thanks for the help function check. From my side, the branch could be merged!

[EttnerAndreas]

okay, think the text might disappeared.
alright, well I already merged it, do we need further tests?
Interesting, I read a lot of those internet connection problems causing this error (56)

[florianhartig]

Hi Andreas, thanks, but if you do a review, please don't merge yet, I will always merge, just note what you think should be changed, or add improvements.

Absolutely nothing unclear about the help ;) ? I must be on fire then!

[EttnerAndreas]

It is quite clear, nothing big but

• for me, it would be interesting to get an easy answer on the benefits of 'refit = T' since it is computationally expensive

-The type changed from Pearson to PearsonChisq, Is there a way to implement both? might be useful to have further help with the application. When to switch from type 'DHARMa' to 'PearsonChisq'
: PearsonChisq focusses on any difference, while person focusses on difference in a direction (linear relationship) , right?

• maybe depending on the audience a further explanation of over/under- dispersion, when it is useful

[florianhartig]

OK, thanks, I've made an issue for that!