- CFA Exams
- CFA Level I Exam
- Topic 1. Quantitative Methods
- Learning Module 8. Hypothesis Testing
- Subject 6. The Decision Rule
CFA Practice Question
A poll of 100 randomly sampled adult Americans found that 53% supported banning handgun sales. A newspaper wants to use this data to conclude that a majority of Americans support banning handgun sales. At the 10% level of significance, the newspaper's conclusion should be that ______
A. a majority, p > 50%, of adult Americans support a ban on handgun sales.
B. there is evidence that 53% of adult Americans support a ban on handgun sales.
C. there is no evidence that a majority, p > 50%, of adult Americans support a ban on handgun sales.
Explanation: The hypothesis test would appear as shown below:
User Contributed Comments 11
User | Comment |
---|---|
danlan | Where comes the value 1.28? |
danlan | 1.28 from z-table |
ahmado | why is 0.5*(1-0.5) the variance? |
Mariecfa | The test statistic, 1.28 is the t-test since the standard deviation is unknown. This is a proportion sample test so we use .5 for the proportion parameter. |
Thecatz | isn't it (0,53-0,5)/(square root of(np(1-p)/n)) so it is 0,4 no ? |
teje | since the sample size is fairly large, despite the fact that population standard deviation is unknown, we can still use t or z test. although the t test would be more conservative. |
nostalgia | the explanation given raises more questions than it answers |
Shaan23 | TheCats its p(1 - p) /n not np(1 - p) / n |
alles | shouldn't variance be ((0.53*0.47)/100)^0.5 ? |
lawlee | I don't understand the p(1-p)/n, can someone kindly explain |
DRamirez | p(1-p) is the variance of a bernoulli random variable. but since the survey is repeated (n = 100), this should be a binomial random variable with variance: np(1-p). the answer is still C though. |