R help - Apr 2008 - pearson's correlation

Hello,
 
I used the function cor to calculate the pearson correlation coefficient between
variables. However, the resulting values do not correspond to the outcome of my
excel-calculations, for which I used the formula Cor(x,y)=Cov(x,y)/(SD(x)*SD(y))
So my question is: How does the function "cor" compute the pearson
correlation coefficient?
 
Thank you in advance,
 
Ake Nauta
_________________________________________________________________


	[[alternative HTML version deleted]]

> I used the function cor to calculate the pearson correlation coefficient
> between variables. However, the resulting values do not correspond to the
> outcome of my excel-calculations, for which I used the formula
> Cor(x,y)=Cov(x,y)/(SD(x)*SD(y)) So my question is: How does the function
> "cor" compute the pearson correlation coefficient?
It would help if you specified exactly what you enter in R and what R's
output
is. In any case, in R the correlation coefficient (using the Person method) 
corresponds exactly to the formula you give, as a quick test
shows:
> x=rnorm(100)
> y=rnorm(100)+x
> cov(x,y)/(sd(x)*sd(y))
[1] 0.7561354
> cor(x,y)
[1] 0.7561354

If Excel gives you something else, I'd rather doubt Excel than R.

Claus


-- 
Claus Wilke
Section of Integrative Biology 
 and Center for Computational Biology and Bioinformatics 
University of Texas at Austin
1 University Station C0930
Austin, TX 78712
cwilke at mail.utexas.edu
512 471 6028

The difference may be due to different handling of missing values.

If you do cor(x,y) "by hand" in excel, you use all available
information
of x and y to calculate sd(x) and sd(y) seperately. But cov(x,y) in 
excel will use only complete pairs of (x,y), which is likely not the 
same set. So your sd and cov (and mean within cov) will be calculated on 
different data. In R, if you use the option use="complete.obs" in cor 
all intermediate calculations will be done on the same (complete) set.
If that is the case of your problem you should got an error message if 
you tried cor() in R without this option on your dataset. But without an 
explanatory example of what you did, this is just guessing.

hth.
Ake Nauta schrieb:
> Hello,
>  
> I used the function cor to calculate the pearson correlation coefficient
between variables. However, the resulting values do not correspond to the
outcome of my excel-calculations, for which I used the formula
Cor(x,y)=Cov(x,y)/(SD(x)*SD(y))
> So my question is: How does the function "cor" compute the
pearson correlation coefficient?
>  
> Thank you in advance,
>  
> Ake Nauta
> _________________________________________________________________
>
>
> 	[[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.
>   


-- 
Pflichtangaben gem?? Gesetz ?ber elektronische Handelsregister und
Genossenschaftsregister sowie das Unternehmensregister (EHUG):

Universit?tsklinikum Hamburg-Eppendorf
K?rperschaft des ?ffentlichen Rechts
Gerichtsstand: Hamburg

Vorstandsmitglieder:
Prof. Dr. J?rg F. Debatin (Vorsitzender)
Dr. Alexander Kirstein
Ricarda Klein
Prof. Dr. Dr. Uwe Koch-Gromus