Dear All,
Thanks for all those replied to my previous question.
Here is a small problem I am facing. I am working on a
huge data (consisting of almost 20000 observations and
780 variables). But for my purpose I am focusing on
only 7 variables and 4000 observaitions. But for all
these variables, when I am generating a scatter plot,
I get almost a bimodal distribution. Now my question
is whether I can apply a ttest for comparison of means
for any two variables since for large sample data I
can assume normality or should I go for non parametric
tests since the sample is giving bi modal data or
whether I should try to seperate out the observations
first to remove bimodality?
thanks in advance,
indrajit
____________________________________________________
Start your day with Yahoo! - make it your home page
http://www.yahoo.com/r/hs
|