[R] Seeking pointers to various regression techniques with R?

[Bert Gunter]

OY!

Have you had any courses in linear models/regression? If so, it's the
so-called "effects hierarchy" principle": generally one does not
expect interactions -- 2nd order effects -- when main effects -- first
order effects -- are absent.

If not, it's difficult to explain, but it certainly has nothing to do
with R and you should post on a statistical list, take a course,
consult a statistician, search the web, take it offlist with Peter ...

-- Bert

On Tue, Jun 5, 2012 at 4:47 PM, Michael <comtech.usa at gmail.com> wrote:
> Thank you!
>
> But could you please explain why the 2nd formula is a model to avoid?
> Thanks again!
> On Tue, Jun 5, 2012 at 6:18 PM, Peter Ehlers <ehlers at ucalgary.ca> wrote:
>
>> You mention 3 models. In all of them, the '-1' simply removes
>> the intercept term; help('formula') explains the use of '-'
>> in general.
>>
>> 1. lm(y~ x * w - 1)  is clearly explained in help('formula');
>> 2. lm(y~ x:w - 1)   ditto (and this is a model to avoid);
>> 3. lm(y~ x/w - 1)   this is equivalent to lm(y~ x + w %in% x - 1)
>> where the %in% operator is explained in help('formula'). A good
>> and simple example is found in the MASS book (chapter 6) the code
>> for which is available in the 'Scripts' subfolder of library/MASS
>> of your R installation.
>>
>> Peter Ehlers
>>
>>
>>
>> On 2012-06-05 13:58, Michael wrote:
>>
>>>  I read your website but still don't know the difference between the
>>> three
>>> formulas...
>>>
>>> Thank you!
>>>
>>> On Mon, Jun 4, 2012 at 11:14 PM, Joshua Wiley<jwiley.psych at gmail.com>**
>>> wrote:
>>>
>>>  Hi Michael,
>>>>
>>>> This is far from exhaustive (I wrote it as an introduction some years
>>>> ago) but you may find it useful to start:
>>>> https://joshuawiley.com/R/**formulae_in_R.aspx<https://joshuawiley.com/R/formulae_in_R.aspx>
>>>>
>>>> Cheers,
>>>>
>>>> Josh
>>>>
>>>> On Mon, Jun 4, 2012 at 9:06 PM, Michael<comtech.usa at gmail.com>  wrote:
>>>>
>>>>> Hi all,
>>>>>
>>>>> Could you please point me to good materials on various
>>>>> tricks/intuitions/techniques of regression, and hopefully in R?
>>>>>
>>>>> For example, what does lm(y~ x * w - 1) mean vs. lm(y ~ x/w -1 ) vs. lm
>>>>>
>>>> (y
>>>>
>>>>> ~ x:w-1), etc...
>>>>>
>>>>> I just found that even simple linear regression is not that simple and
>>>>> there are a lot of tricks/techniques in using them...
>>>>>
>>>>> Hopefully I can find good materials on these!
>>>>>
>>>>> Thank you!
>>>>>
>>>>>        [[alternative HTML version deleted]]
>>>>>
>>>>> ______________________________**________________
>>>>> R-help at r-project.org mailing list
>>>>> https://stat.ethz.ch/mailman/**listinfo/r-help<https://stat.ethz.ch/mailman/listinfo/r-help>
>>>>> PLEASE do read the posting guide
>>>>>
>>>> http://www.R-project.org/**posting-guide.html<http://www.r-project.org/posting-guide.html>
>>>> <http://www.**r-project.org/posting-guide.**html<http://www.r-project.org/posting-guide.html>>
>>>>
>>>>
>>>> and provide commented, minimal, self-contained, reproducible code.
>>>>>
>>>>
>>>>
>>>>
>>>> --
>>>> Joshua Wiley
>>>> Ph.D. Student, Health Psychology
>>>> Programmer Analyst II, Statistical Consulting Group
>>>> University of California, Los Angeles
>>>> https://joshuawiley.com/
>>>>
>>>>
>>>        [[alternative HTML version deleted]]
>>>
>>> ______________________________**________________
>>> R-help at r-project.org mailing list
>>> https://stat.ethz.ch/mailman/**listinfo/r-help<https://stat.ethz.ch/mailman/listinfo/r-help>
>>> PLEASE do read the posting guide http://www.R-project.org/**
>>> posting-guide.html <http://www.r-project.org/posting-guide.html>
>>> and provide commented, minimal, self-contained, reproducible code.
>>>
>>
>>
>
>        [[alternative HTML version deleted]]
>
> ______________________________________________
> 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.



-- 

Bert Gunter
Genentech Nonclinical Biostatistics

Internal Contact Info:
Phone: 467-7374
Website:
http://pharmadevelopment.roche.com/index/pdb/pdb-functional-groups/pdb-biostatistics/pdb-ncb-home.htm