Re: I am complete new to PSPP and have some questions

[Jason Stover]

On Fri, Sep 30, 2011 at 11:09:05AM +0200, Stefan Layer wrote:
> Questions: 
> 1. I need help in interpreting the statistical result number. 
> This is most due to a basic lack in memory to my statistical classes more
> than 25 years ago. 
> But also - ich couldnt find an explanation....
> a) For "my eyes" there is the strange result, that people form Northamerica
> are more likely to be able to twinkle only on the right side, 
> versus people from europe more on the left. 
> I understand the first 3 lines of percentages - but not the last line - the
> fourth one.
> b) Can anybody give me an mathematical explaniation

This could be due to random chance: You may have happened to sample
subjects who gave this table. To know more, you could try a
permutation test, but the test statistics at the lowest table suggests
the apparent difference is due to randomness.

> c) I also do not know how to interpretate correctly the last information
> package 
> about "Statistics". Is there an anybody to tell me more about that?

The results in the "Statistic" table state there is no significance
relaitonship between "Twinkle" and "Kontinent". That is, knowing the
value of one variable doesn't tell you anything about the other,
outside your own sample. You can see this result in the far right
column, which shows p-values much larger than 0. But these statistics
may not be appropriate for your data, since several cells in your
contigency table have zero frequency.

You can read more about this by looking up the Pearson chi-square
statistic and likelihood ratio chi-square statistic. They both test
for dependence among variables, but have problems when the cell counts
are zero (as in your case).