To determine if something is statisically significant do you follow the rule reject if -
-test stat critical
If we reject should we treat as not statistically significant?
In EOC 26, t stat provided for Intercept = 3.0275 and t stat for slope of coefficient = -2.3014, I believe the critical value is 1.96. So I would think that these regressions would not be statisically significant, but alas according to the answers I’m wrong - would greatly appreciate if someone could put me in my place and tell me where I’m going wrong… This question has gotten to me, please, please help. Thanks again guys
Rejecting the null would indicate statistical significance. Large test statistics (F, z, t, chi-squared, etc.), indicate statistical significance (or you can say small p-values). Of course, large or small is determined relative to a critical value or the selected significance level, respectively.
Also, for the rejection region rules (assuming a two-tailed test), it might be easier to think of it in this manner:
the interval (-critical value,+critical value)
—> if the observed/calculated test statistic is outside of this interval, we can reject Ho.
It’s the same thing, but provides another (possibly simpler) way to look at it instead of the “-test stat critical” perspective.
For this question, it works, but I would be careful with this blanket statement (since people who are unfamiliar with statistics might try to apply this logic to situations where it doesn’t apply (in terms of significance implying nonzero)).
Statistically significant means the test results indicate a contradiction to the null hypothesis that is not likely due to chance.
The CFA curriculum tends to use “statistically significant” as a synonym for “not zero.”
I agree that in the broader context we should be careful about equating those phrases, but in the narrow context of the Level II exam, it’s appropriate.
Ah, okay. I haven’t cracked the LII books just yet (so the CFAI scope is unfamiliar to me), although I already have them. I was trying to be more cautious that someone might take it as a golden rule and apply it elsewhere.