[R] Urgent - I really need some help lme4 model avg Estimates

Mitchell Maltenfort mmalten at gmail.com
Wed Mar 28 13:08:26 CEST 2012

You said it right in your first letter.  You want to keep it at a
level that is comprehensible.  Not just to you, but your colleagues,
reviewers, readers...

Remember the sage wisdom that all models are wrong, but some of them
are useful.

Rather than try to forge the Excalibur of statistical models, consider
writing your paper around presenting 2-4 different, relatively simple
statistical models and compare and contrast their interpretations.
How well do they agree with each other and with informed expectations?

Ersatzistician and Chutzpahthologist
I can answer any question.  "I don't know" is an answer. "I don't know
yet" is a better answer.

On Wed, Mar 28, 2012 at 1:21 AM, Dragonwalker
<dragonwalkerart at hotmail.com> wrote:
> I understand where you are coming from, but the issue is that some
> exploration of the data through graphs and the like, showed that patterns
> could be seen. However with only 7 means it is extremely difficult to get
> any kind of statistical evidence and as some mean values are the same some
> of the tests that I wanted to use such as a Mann-Whitney would not even run
> so I had to resort to a one-sample Wilcoxon with a set mu value. (minimum
> p-value that was even possible was p=0.280).
> I asked a couple of forums in December about the issues at hand and they
> suggested that I look into mixed-effect models so I read some chapters on
> them and got very excited, but at the time still thought of them as some
> test that could give me means. However it all clicked and I realise that
> they can be more useful as a tool to illustrate which factors and covariates
> best fit to the response variable.
> I understand the concepts of fitting an intercept and slope somewhat but the
> paperwork on it can be a little confusing, however the way they were used in
> the paper (of which I attached one of the tables) seemed a very
> straightforward method of teasing the intricate factors of habitat, age and
> other factors that could be affecting behaviour such as time feeding and
> foraging rate. Believe me, if i could have survived with Kruskal-Wallis then
> I would have had my thesis written up three months ago with a lot less
> stress. I am not looking for pretty as I don't even want it published, but I
> did hope to be able to give the time that I spent collecting data justice.
> I have come really far, thanks to some great people, but I do not have
> anyone near me who can help and my adviser is 3000 miles away too and is not
> a statistician either.
> All I would like to know is how could Maslo et al. have calculated estimates
> for all categories AND an intercept and is there a method to do this in R.
> I have spent months trying to find these answers and so I would greatly
> appreciate an answer to this question.
> Thank you again.
> --
> View this message in context: http://r.789695.n4.nabble.com/Urgent-I-really-need-some-help-lme4-model-avg-Estimates-tp4511178p4511396.html
> Sent from the R help mailing list archive at Nabble.com.
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.

More information about the R-help mailing list