troubleshooter
Active Member
Hello David:
If this issue has not been raised before, would you have time to look into this for me please?
In Slide No. 102 of the second part of the first Quant video (2009_2.a.i_quantitative_continued), the F-test is used to test whether the variance of the Male and Female Test scores come from the same distribution.
At 99% Confidence and 1% Significance, Critical F value for df1=24 and df2=24 is given as 2.66.
Based on the F Value lookup table, It seems F critical Value 2.66 is based on one tail test, ie, this value corresponds to 1% probability to the right of the critical value. Based on the question, we are actually testing whether the variances are different (and NOT whether Female variance is greater than Male Variance). Therefore, we should actually use the F value of 2.97 which corresponds to 0.5% area to the right of it. As the test actually is two tail, this area 0.5% along with the area of 0.5% at the lower tail makes up 1% of Significance. Look up tables for critical F values typically do not provide the lower tail value, which, of course, is redundant for most of F-test as F-ratio is always calculated with the larger variance being placed at the at the numerator.
Thanks.
If this issue has not been raised before, would you have time to look into this for me please?
In Slide No. 102 of the second part of the first Quant video (2009_2.a.i_quantitative_continued), the F-test is used to test whether the variance of the Male and Female Test scores come from the same distribution.
At 99% Confidence and 1% Significance, Critical F value for df1=24 and df2=24 is given as 2.66.
Based on the F Value lookup table, It seems F critical Value 2.66 is based on one tail test, ie, this value corresponds to 1% probability to the right of the critical value. Based on the question, we are actually testing whether the variances are different (and NOT whether Female variance is greater than Male Variance). Therefore, we should actually use the F value of 2.97 which corresponds to 0.5% area to the right of it. As the test actually is two tail, this area 0.5% along with the area of 0.5% at the lower tail makes up 1% of Significance. Look up tables for critical F values typically do not provide the lower tail value, which, of course, is redundant for most of F-test as F-ratio is always calculated with the larger variance being placed at the at the numerator.
Thanks.
