Quant :2007 Random Set 1 Q18, Random Set 2 Q9

[liewpw05]

Hi David,

With regards to the above questions in the random set , i am curious why the chi square test statistic used is n*(sample variance)/(population variance) instead of (n-1)*(sample variance)/(population variance)?

Thanks

Peggy

 

[David Harper CFA FRM]

Hi Peggy,

I had to look this up, frankly. Last year the Stat text was Spiegel Statistics (the Gujarati replaced just this year for stat, 2008). Even for his practice questions with *samples* he is using:

n*(sample variance)/(population variance)

So, the short answer is: we were following the assigned Spiegel. But Gujarati (at 5.21) appears to be more precise/thoughtful on this...

David