search

elmacartney / EE_stress_MA

MIT License
0 stars 0 forks source link

Dplyr mutate affecting VCV #9

Closed elmacartney closed 3 years ago

elmacartney commented 3 years ago

I have tried to run the models after using the club sandwich code but the VCVs don't seem to be working.

models say: Error in rma.mv(yi = lnRR_Sa, V = VCV_S, random = list(~1 | Study_ID, : Processing terminated since k <= 1

itchyshin commented 3 years ago

@elmacartney - I can have a look at this and also the original one - there should be some easy solutions for this. It is often useful to look at whether the matrix looks correct. Please remember to push once you finish working on this. I have a series of meetings now but will look at it after that

elmacartney commented 3 years ago

Thanks @itchyshin. The matrix doesn't look correct when checking if it is positive definite or symmetric.

On Fri, Jul 23, 2021 at 3:58 PM Shinichi Nakagawa @.***> wrote:

@elmacartney https://github.com/elmacartney - I can have a look at this and also the original one - there should be some easy solutions for this. It is often useful to look at whether the matrix looks correct. Please remember to push once you finish working on this. I have a series of meetings now but will look at it after that

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/elmacartney/EE_stress_MA/issues/9#issuecomment-885418013, or unsubscribe https://github.com/notifications/unsubscribe-auth/ATGFCZQZMGAOHEZRP2D7643TZEAG3ANCNFSM5A3LFA5Q .

itchyshin commented 3 years ago

@elmacartney - fixed

I realised that you do not really do "TESTING" in R as you write code (sometimes known as defensive coding) - it is a really fundamental technique to learn and you should code with a lot of testing especially when you are still a beginner to an intermediate user (I consider myself as an intermediate user and I usually do testing as I code or at least do so when stop working).

Learning to test will enhance your coding ability and solve R problems by 10 fold so let's discuss this at our next meeting. This really is an important one and if you learn this, I can really trust your coding and I will probably need to check less (it took a while for Rose to learn this and also I was telling the importance to Sammy the other day).

Have a good break! - Shinichi