I realise there are better forums to ask, but does anyone have experience of multilevel imputation using mice in R

[StealthPolarBear]

And be willing to answer some idiotic questions?

I thought this was going to be some sort of mad scientist thread about testing on mice in a laboratory. Disappointed.

I used to find debugging R a right pain in the ass. My job was to convert stuff in R into a language we could integrate with our products (C#) but stepping through R was very fiddly IIRC.
Happy days, that was a great job

I thought this was going to be about lab experiments on mice. I am relieved it’s not. At least I think it’s not.

No worries, Stealth! It's always easier to proof read someone else's code than your own... I'm a biologist rather than a statistician and you're doing a lot better as a beginner in R than I did.

This reminds me of pirate jokes we used to tell about R... What's a pirates favourite computer language? Aarrgh!

It works! Clustering and all. Thank you for all the help and humour on the way.

Well done 👍
Coding is frustrating when you are stuck but very satisfying when you crack it!

Also proving Mumsnet probably is the best forum for pretty much most things.

Hi @StealthPolarBear, am I right to think from previous posts you’re in the NE? If this is research at the uni then send me a PM as I might be able to point you in the direction of help there, for future reference if not for this issue.

hey gals, can I ask a super stupid R question?

If I have a script that only has something like Print "hello world" and I am at the R prompt that looks like a Chevron (running R on unix)

How do I run that script so that it prints

hello world

I imagine the syntax would look like

[command to run scripts] nameofscript.txt
hello world

But I'll be damned if I know what [command to run scripts] is.

I can use Rscript from the unix shell prompt to run the script, I just think there MuST be a way to just run a script from the live > prompt. I know this is so simple that no one ever explains how to do it online. I am so embarrassed that I can read R code but I can't do the most basic thing like run a macro from the >.

I have too many problems trying to make Rstudio work, so please don't suggest Rstudio.

TIA, if you can help. stata is so much nicer ime but someone sent some R code to do clever things

Lljkk I'm afraid I don't know. But that's the problem I seem to have, my queries are too easy to Google!

Norrie I am on the NE but this is not university related. Appreciate the offer though

Wouldn't you just navigate to the R directory and then just type in the name of the script file

Eg (oncw you're in the right directory)

helloworld.R

i think i tried that...

Google run script

Something Like On command line

R scriptname

I read this whole thread despite not understanding a single word of it.

JoshLyman...

It's oddly compelling

Or even Rscript scriptname

It's late

I am in awe ! I am way too old !

I thought you might have done. Worth checking!

My script is taking too long. Is a reduction in the predicotss for each predicted variable likely to help much?

It’s a bit like the shipping forecast

Some stuff can take days on big datasets ( not used MICE )

If you are just checking the functionality of the script, would a smaller dataset help initially and then leave it for a week? How does it / underlying algorithm scale?

@lljkk the command you want is source("myscript.R")
I can't take credit for knowing that - I looked it up in my copy of the R Cookbook! I use Rgui which might be more amenable than Studio.

@StealthPolarBear how long is too long? I have one script I always leave running over lunch...