T test

CFA Level II
#1

I can’t figure these questions out! Consider a sample of 60 observations on variables X and Y in which the correlation is 0.42. If the level of significance is 5%, we: A) cannot test the significance of the correlation with this information. B) conclude that there is statistically significant correlation between X and Y. C) conclude that there is no significant correlation between X and Y. The calculated t is t = 3.5246 and the critical t is approximately 2.000. So I figured that -2 < 3.5246 but 3.5246 is NOT LESS THAN 2. So shouldn’t the answer be C??? (its not)

#2

t calc = 3.5246 and t crit = 2 t calc > t crit - so you conclude that the correlation coefficient is significant… B

#3

I thought that since the t calc is > than the t critical value, then there is no significant correlation since it is “outside” the range, this is killing me!

#4

look at what you are trying to prove with the correlation t-test. your null hyp is rho=0 while alternate is rho != 0. so you end up proving that rho is not = 0… because you are in the “reject null hyp” part of the test.

#5

what cpk said. t calc > tcrit, you reject that correlation = 0, therefore it is significant.

#6

I think I’m getting there, But if you conclude that rho is not equal to zero and you reject the null, aren’t you concluding that there is NO statistically significant correlation???

#7

null is saying rho=0 - so that means no significant correlation. now you reject the null - so it IS SIGNIFICANT CORRELATION.

#8

Thank you so much, I had it backwards! One more question Is the null hypothesis always = 0 and no significant correlation??

#9

correlation is almost always 0 and no significant correlation test. but with b0 and b1 it could be check whether coefficient is diff from 1 or some other value…

#10

if you’re testing if the correlation is significant, then yes.

#11

> now you reject the null - so it IS SIGNIFICANT CORRELATION. cpk, to be more accurate, when you reject the null, you are simply saying it is not true that there is no significant correlation. You cannot say it IS SIGNIFICANT CORRELATION.

#12

dreary please read the answer choices given and then make your statement. I understand what you are saying… but in the CFA world (unlike the world of statistics) there is more black and white rather than the gray you are used to seeing… to clarify A) cannot test the significance of the correlation with this information. A is wrong obviously. B) conclude that there is statistically significant correlation between X and Y. C) conclude that there is no significant correlation between X and Y. C is not true given t calc > t crit. So B is your only CHOICE

#13

Sure. But, if they gave you the choices as below, you will not get the correct answer if you don’t truely understand what you are doing:. Check. A) conclude that there is statistically significant correlation between X and Y. B) conclude that there is no significant correlation between X and Y. C) conclude it is not true that there is no statistically significant correlation between X and Y

#14

lol can you type the entire question before typing the options? I picked A initially because I didn’t see the t-stat information anywhere. The rule of 2!! If ABS(Tstat)>Tcrit, we reject the null (Ho:x=0) because the value is significantly different from zero Think of it this way, if Ho: x=0, means there’s no linear relationship which is bad. The Null (Ho) is what we are testing… HighBall Wrote: ------------------------------------------------------- > I can’t figure these questions out! > > > Consider a sample of 60 observations on variables > X and Y in which the correlation is 0.42. If the > level of significance is 5%, we: > > A) cannot test the significance of the > correlation with this information. > > B) conclude that there is statistically > significant correlation between X and Y. > > C) conclude that there is no significant > correlation between X and Y. > > > The calculated t is t = 3.5246 and the critical t > is approximately 2.000. > > > So I figured that -2 < 3.5246 but 3.5246 is NOT > LESS THAN 2. > > > So shouldn’t the answer be C??? (its not)

#15

@ Dreary: H0: rho = 0 (no significant correlation ) Ha: rho =/= 0 (there is significant correlation.) Since the calculated t statistic is greater than the critical value, you reject the null hypothesis. And so, there is a statistically significant relationship, i.e. my ans is A…

#16

Answer is C.

#17

You got it all backwards friend. If Tstat>Tcrit, reject Ho:x=0. Meaning the value is significantly different from 0. I am also going with B on this one. kenanadu Wrote: ------------------------------------------------------- > lol can you type the entire question before typing > the options? I picked A initially because I > didn’t see the t-stat information anywhere. The > rule of 2!! If ABS(Tstat)>Tcrit, we reject the > null (Ho:x=0) because the value is significantly > different from zero > > Think of it this way, if Ho: x=0, means there’s no > linear relationship which is bad. The Null (Ho) > is what we are testing… > > > HighBall Wrote: > -------------------------------------------------- > ----- > > I can’t figure these questions out! > > > > > > Consider a sample of 60 observations on > variables > > X and Y in which the correlation is 0.42. If > the > > level of significance is 5%, we: > > > > A) cannot test the significance of the > > correlation with this information. > > > > B) conclude that there is statistically > > significant correlation between X and Y. > > > > C) conclude that there is no significant > > correlation between X and Y. > > > > > > The calculated t is t = 3.5246 and the critical > t > > is approximately 2.000. > > > > > > So I figured that -2 < 3.5246 but 3.5246 is > NOT > > LESS THAN 2. > > > > > > So shouldn’t the answer be C??? (its not)