- CFA Exams
- CFA Level I Exam
- Topic 1. Quantitative Methods
- Learning Module 8. Hypothesis Testing
- Subject 2. Null Hypothesis and Alternative Hypothesis
CFA Practice Question
If the underlying theory suggests that the value of an estimated variable is greater than a particular number, it is appropriate to use ______.
A. a two-tailed test
B. a right-tailed test
C. a left-tailed test
Explanation: When an observation is made and we see a value greater than the threshold suggested by the theory, we would like to known the probability of getting the larger value due simply to chance rather than to the event being driven by the theory. For this, we need to known the right-tail probabilities under the null hypothesis.
User Contributed Comments 6
| User | Comment |
|---|---|
| jpducros | Can someone give an explanation? |
| JonClark | Here you go jp Explanation: When an observation is made and we see a value greater than the threshold suggested by the theory, we would like to known the probability of getting the larger value due simply to chance rather than to the event being driven by the theory. For this, we need to known the right-tail probabilities under the null hypothesis. ...nah just kidding |
| geofin | H(0) usually comes from theory. We want to question (test) the theory and so H(a) will be: "an estimated value is less than or equal a particular value. This requires the left-tailed test. So the correct answer should be C. Any other thoughts? |
| Hungerford | seems like the answer should be C to me as well. Hnull is what is currently accepted in the world. Seems like Hnull is "X > 10" Halternative: what you would "accept" if the null is rejected. Halt "X <= 10" So you want to reject the null. That is, we would say, "our observed value of X was so far below 10 that we reject the null in favor of the alternative" Why is this not called a left-tailed test. You're seeing if the observed value of X falls inside the rejection region, which is to the left of the mean (if we call X the mean). I obviously don't see my error. |
| MIKAELAH | The null hypothesis always includes the "equal to" condition. Therefore, in this case, Hnull: X<= 'number' and Halternative X>'number' |
| GBolt93 | Hungerford you've got it backwards. You can't prove hnull, only reject it, so if you wanted to provide evidence that x is larger than a certain number, hnull would be that it's less than that number. I.e. H0 x<=10, H1 x>10 |