# quant: sample exam: two tail test and 95% significance

the question about " exhibit 2 and two tail t-test to deterine if the coefficients are equal to zero, at the 0.05 significance level, … is it is at 0.05 significance level should we also accept Hull hyp for intercept=0? as its p-value >0.025? Thanks.

i was actually a little bit confused about the solution too. I would use p-value < 0.025 for 95% CI 2-tail, but CFAI used p < 0.05

p-value is to be compared to alpha (level of significance)…if it’s a 5% signficance level, p-value needs to <0.05 two tail tests simply have alpha/2 on each tail…but p-value is still compared to alpha

so for a two tailed test, if it is 5% significance level, and p-value is 0.04, than reject Ho? or if p-value is 0.015, then we could reject H0? which value p-value should compare against ? 5% or 2.5% in two tail test? Thanks. MFIN— Wrote: ------------------------------------------------------- > p-value is to be compared to alpha (level of > significance)…if it’s a 5% signficance level, > p-value needs to <0.05 > > two tail tests simply have alpha/2 on each > tail…but p-value is still compared to alpha

If you look at the table, a one sided at 95 gives the same t value as a two sided at 90. For this reason, the p-value for a one sided 5% sig will never be the same as a two-sided at 5% sig.

good question… i guess we apply p less than <0.025 … can anyone confirm?

the cfai sample used p<0.05 for a 95% CI, 2 tail. is this wrong?

no, obviously not wrong. I mentioned this already…compare p-value to the significance level 5% significance level, regardless of one or two tail, is compared to the p-value. if p-value is <0.05, reject the null

The p value for an intercept is like a t-value, each one of them will be unique and you compare this to the critical p (or t) value. As MFIN mentioned above, 95% significance means a 0.05 p-value is the p-critical and if you are given a p-value in your table for the coresponding intercept that is less than 0.05 than can safely reject at 95% significance.

except you likely wouldn’t be testing the intercept? not sure if I agree with your post… but i think you get the basic idea

double checked with 2 economics major students, we look at p < 0.05 even if its 2 tail.

thanks everyone!

come on Bilal…you don’t trust a Master of Finance graduate? my feelings are hurt I’ve taken several stats courses along with econometrics courses…this is literally 1st year Intro to Stats…get a grip