R help - Aug 2008 - Major difference in multivariate analyses SPSS and R

Dear colleagues,

 

I know SPSS can not compute linear mixed models. I used 'R' before for
computing multivariate analyses. But, I never encountered such a major
difference in outcome between SPSS and 'R':

 

In SPSS the Pearson correlation between variable 1 and variable 2 is 31%
p<0.001.

 

In SPSS binary logistic regression gives us an Odds Ratio (OR)=4.9 (95%
CI 2.7-9.0), p<0.001, n=338.

                    OR          lower   upper

gender          1,120       0,565   2,221

age               0,985       0,956   1,015

variable 2         4,937   2,698   9,032

 

In R multilevel logistic regression using statistical package 'lmer'
gives us an Odds Ratio=10.2 (95% CI 6.3-14), p=0.24, n=338, groups:
group 1, 98; group 2 84.

                   OR           lower   upper

gender         2,295        -2,840 7,430

age              0,003        -70,047           70,054

variable 2     10,176     6,295   14,056

 

The crosstabs gives us:

     variable A

Var B  0   1

    0 156 108

    1  17  57

 

Would somebody know how it is possible that in SPSS we get p<0.001 and
in R we get p=0.24?

And, in 'R' the 95% CI of the Odds Ratio is 6.2-14.1. Why is the
p-value=0.24?

 

Thanks, 

 

Ronald


	[[alternative HTML version deleted]]

Draga, R. wrote:
> .....
>  
>
> Would somebody know how it is possible that in SPSS we get p<0.001 and
> in R we get p=0.24?
>
> And, in 'R' the 95% CI of the Odds Ratio is 6.2-14.1. Why is the
> p-value=0.24?
>
>  
>
>   
You asked this before and you are still not providing sufficient 
information.

Harold already pointed out that excluding an important variance 
component in SPSS could easily explain a difference in p values. 
However, the discrepancy between CI and p value suggests that you could 
be misreading the output from lmer(), and/or misspecifying the model. 
Now,  you are simply not telling us

  - what you did
  - what the data are like (esp. which variables vary within and between 
levels)
  - what the output was
  - which version of lmer() was used

Also you are confusing us with "variable 1", "variable 2",
"variable A",
and "Var B"

-- 
   O__  ---- Peter Dalgaard             ?ster Farimagsgade 5, Entr.B
  c/ /'_ --- Dept. of Biostatistics     PO Box 2099, 1014 Cph. K
 (*) \(*) -- University of Copenhagen   Denmark      Ph:  (+45) 35327918
~~~~~~~~~~ - (p.dalgaard at biostat.ku.dk)              FAX: (+45) 35327907