focus on all years combined?

[teixeirak]

@mcgregorian1 , (and again @ValentineHerr ),

As I work more on the presentation, I'm also returning to this comment from R1: "Following the two previous comments, I suggest to leave in Table 5 just the best model (rather that a multimodel inference as is presented now) to discuss more clearly the covariates included. And I suggest discussing only the model with all years together. This would be more robust and avoid ramble differences between years that mostly look spurious and with a difficult physiological explanation. This would help to ensure a better use of the Resilience indices."

I think it makes sense to focus on all years combined. Okay, @mcgregorian1 ?

Note that this affects #71.

[teixeirak]

That would make Table 1 much cleaner!

[teixeirak]

@mcgregorian1 , can you let me know ASAP what you think of this?

[mcgregorian1]

Hi I'm here. I think this works and it does present it more succinctly

[mcgregorian1]

I'm going to be working on this for the next hour or so before coming back later tonight.

[mcgregorian1]

Did you want to keep Table 5 as only all years together, and then put the individual years as a separate SI table but formatted the same?

[teixeirak]

Okay, great. I'm working on big changes related to this and #112 right now.

[teixeirak]

Did you want to keep Table 5 as only all years together, and then put the individual years as a separate SI table but formatted the same?

I'm not sure if we want Table 5 in the main article... It's pretty well covered in Fig. 4.

[mcgregorian1]

Sorry just to clarify, does this mean we want 4 separate panels (for the traits) for the visreg results, with one line in each one? That's what the output would be if we drop all years from the visualization

[teixeirak]

Yes, I think just show the model for all years combined, although that's a pretty (overly?) simple figure. Could we now make it work with the confidence intervals?

[mcgregorian1]

Unfortunately the confidence interval issue is because we're using linear mixed models to begin with, not because we were showing too many models. The only way to appropriately show confidence intervals is by having a top model that doesn't include a random effect.

I will update the visreg figure with just the top model and show it in #71.

The other option is to include the arimaratio as a different color? Not sure that's ideal though if we're putting arimaratio in SI to begin with.

[ValentineHerr]

Did you check this?

http://www.remkoduursma.com/post/2017-06-15-bootpredictlme4/

Get Outlook for iOShttps://aka.ms/o0ukef

---

From: Ian McGregor notifications@github.com Sent: Thursday, July 16, 2020 4:03:44 PM To: SCBI-ForestGEO/McGregor_climate-sensitivity-variation McGregor_climate-sensitivity-variation@noreply.github.com Cc: Herrmann, Valentine HerrmannV@si.edu; Mention mention@noreply.github.com Subject: Re: [SCBI-ForestGEO/McGregor_climate-sensitivity-variation] focus on all years combined? (#113)

External Email - Exercise Caution

Unfortunately the confidence interval issue is because we're using linear mixed models to begin with, not because we were showing too many models. The only way to appropriately show confidence intervals is by having a top model that doesn't include a random effect.

I will update the visreg figure with just the top model and show it in #71https://nam02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2FSCBI-ForestGEO%2FMcGregor_climate-sensitivity-variation%2Fissues%2F71&data=02%7C01%7Cherrmannv%40si.edu%7C9d5e86e163e24b1b937208d829c3582b%7C989b5e2a14e44efe93b78cdd5fc5d11c%7C0%7C0%7C637305266258913938&sdata=iKKBR%2FZ8Tp2UY0McMisRjo%2FbcU6SgOy6Bn1xpR3JZ3A%3D&reserved=0.

The other option is to include the arimaratio as a different color? Not sure that's ideal though if we're putting arimaratio in SI to begin with.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHubhttps://nam02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2FSCBI-ForestGEO%2FMcGregor_climate-sensitivity-variation%2Fissues%2F113%23issuecomment-659641148&data=02%7C01%7Cherrmannv%40si.edu%7C9d5e86e163e24b1b937208d829c3582b%7C989b5e2a14e44efe93b78cdd5fc5d11c%7C0%7C0%7C637305266258913938&sdata=juLLemI6JdbodlNbmEjuRYO1wEGLQQPDa5hjZcbg14E%3D&reserved=0, or unsubscribehttps://nam02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FAEWDCIOGNNYUQYOYNLWVRX3R35MKBANCNFSM4O4XLHDQ&data=02%7C01%7Cherrmannv%40si.edu%7C9d5e86e163e24b1b937208d829c3582b%7C989b5e2a14e44efe93b78cdd5fc5d11c%7C0%7C0%7C637305266258923928&sdata=tU5EbVFVj%2FzKMq%2FBTeuVU7EhDQxlpPl1NwDwMdgRxZg%3D&reserved=0.

[mcgregorian1]

oh this is perfect! I'll implement this tonight, this will be great for our plots

[teixeirak]

I think we're fine with this now.