R help - Nov 2011 - SPSS F-test on change in R square between hierarchical models

Hi,

I am wondering if anyone knows how to perform an F-test on the change in R
square between hierarchical models in R? SPSS provides this information and
a researcher that I am working with is interested in getting this
information. Alternatively, if someone knows how I can calculate the test
statistic (SPSS calls it F-change?) and dfs that would be helpful as well.

The output and the test I am looking for is available here ->
http://dl.dropbox.com/u/1501309/spss_rsquared.pdf

Chris

On Wed, Nov 23, 2011 at 6:20 AM, Smart Guy <smartguy3k@gmail.com> wrote:
> Hi All,
>           I was adding a new row of data to my data frame using rbind(). I
> was surprised to see that after adding new row, I lost my data frame level
> attibute as well as  col level attribute. Please help me to insert a new
> row at frist or middle position so that my custom attribute is not lost.
>
>  Here is what I did.
>
> age<-c(15,20,18)
> weight<-c(40,42,30)
>
> ### creating my data frame ####
> mydata <- data.frame(age,weight)
>
> ### creating data frame level attribute ####
> attr(mydata,"myattr")<-c("myinfo")
>
> ### creating col level attribute for 'age' column  ###
> attr(mydata$age,"mycolattr")<-c("mycolinfo")
>
> #### Checking attributes  ###
> attributes(mydata)
> attributes(mydata$age)
>
> ### creating new row  #####
> newrow <- data.frame(age=16, weight= 42)
>
> #### Inserting newrow as first row to my data frame ####
> mydata<- rbind(newrow, mydata)
>
> #### Checking attributes again ###  I lost my custom attributes
> attributes(mydata$age)
> attributes(mydata)
>
>
> Thanks in advance,
>
>
> --
> SG
>
>        [[alternative HTML version deleted]]
>
> ______________________________________________
> R-help@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.
>
	[[alternative HTML version deleted]]

Probably

?anova
?add1
?drop1

are what you're looking for.

-- Bert


On Wed, Nov 23, 2011 at 6:07 AM, Christopher Desjardins
<desja004 at umn.edu> wrote:
> Hi,
>
> I am wondering if anyone knows how to perform an F-test on the change in R
> square between hierarchical models in R? SPSS provides this information and
> a researcher that I am working with is interested in getting this
> information. Alternatively, if someone knows how I can calculate the test
> statistic (SPSS calls it F-change?) and dfs that would be helpful as well.
>
> The output and the test I am looking for is available here ->
> http://dl.dropbox.com/u/1501309/spss_rsquared.pdf
>
> Chris
>
> On Wed, Nov 23, 2011 at 6:20 AM, Smart Guy <smartguy3k at gmail.com>
wrote:
>
>> Hi All,
>> ? ? ? ? ? I was adding a new row of data to my data frame using
rbind(). I
>> was surprised to see that after adding new row, I lost my data frame
level
>> attibute as well as ?col level attribute. Please help me to insert a
new
>> row at frist or middle position so that my custom attribute is not
lost.
>>
>> ?Here is what I did.
>>
>> age<-c(15,20,18)
>> weight<-c(40,42,30)
>>
>> ### creating my data frame ####
>> mydata <- data.frame(age,weight)
>>
>> ### creating data frame level attribute ####
>> attr(mydata,"myattr")<-c("myinfo")
>>
>> ### creating col level attribute for 'age' column ?###
>> attr(mydata$age,"mycolattr")<-c("mycolinfo")
>>
>> #### Checking attributes ?###
>> attributes(mydata)
>> attributes(mydata$age)
>>
>> ### creating new row ?#####
>> newrow <- data.frame(age=16, weight= 42)
>>
>> #### Inserting newrow as first row to my data frame ####
>> mydata<- rbind(newrow, mydata)
>>
>> #### Checking attributes again ### ?I lost my custom attributes
>> attributes(mydata$age)
>> attributes(mydata)
>>
>>
>> Thanks in advance,
>>
>>
>> --
>> SG
>>
>> ? ? ? ?[[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.
>>
>
> ? ? ? ?[[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

Christopher,
> I am wondering if anyone knows how to perform an F-test on the change in R
> square between hierarchical models in R? SPSS provides this information and
> a researcher that I am working with is interested in getting this
> information. Alternatively, if someone knows how I can calculate the test
> statistic (SPSS calls it F-change?) and dfs that would be helpful as well.
What you describe is just the standard F test for comparing two models, 
or testing deviance reduction between tow *nested* models (I suspect 
this is what you mean by "hierarchical"). The anova() function will do
that. The R2STATS GUI will also give you these tests, along with the 
R-squared, in the same table.

A common misconception about an F-test is that it is the test on a 
variable effect, when strictly speaking it is a test on the deviance 
reduction between two models that include or not that particular 
variable (and there may be several ways to do that, each leading to 
possibly different F-values).

Yvonnick Noel
University of Brittany
Department of Psychology
Rennes, France