Question
Quiz Questions ECON301 Fall 2017 1. Fisher was interested in testing the sharp null hypothesis. The sharp null refers to the hypothesis that a- The treatment has no effect on the majority of people b- The treatment has no effect on any person c- The treatment has no effect, on average, across all people d- The treatment has a non-zero effect on every person are tested while keeping 2. In Fisher exact test, all possible fixed. This is called a- Observations; treatment assignments; bootstrapping b- Treatment assignments; outcomes; bootstrapping c- Observations; treatment assignments; a permutation test d- Treatment assignments ; the observations ; a permutation test 3. Suppose you are interested in estimating Ts, the percentage of voters who support Donald Trump in your Facebook network. To this goal, you administer a survey to poll some randomly selected sample of friends. You do some work in R, then you obtain an estimate Tx" of the fraction of Trump voters from your data. By default. Stata returns a (1-) level confidence interval for your estimand Ts (default is a-5%). Which of the following is true? (Select all that apply) a-You are 95% sure that your estimate of the fraction of Trump voters in Facebook network is correct. b- If you compute the interval in repeated samples, then (1-a) percent of the time, the interval will bracket the true fraction Ts. c- is the significance level of the test. and 4. In hypothesis testing, type I error refers to the possibility of type II error refers to the possibility of a. Not finding a treatment effect when one exists (failing to reject the null when it is in fact false); finding a treatment effect that actually does not exist (rejecting the null when it is in fact true) b. Finding a treatment effect that does not actually exist (rejecting the null when it is in fact true); not finding a treatment effect when in fact there is one (failing to reject the null when it is in fact false) 5. One way to proceed with power calculations is to use a,,t,o, and y, and using these inputs, calculate the sam ple size needed to detect a significant treatment effect. refers to the refers to significance level of the test, and 1- is power. refers to and refers to a- The actual measured average treatment effect ; the actual measured standard deviation b- The target average treatment effect; an estimate of standard deviation of the outcome c- The target average treatment effect; the desired level of significance of the test; the d- The target average treatment effect ; the desired level of significance of the test ; a ; the fraction of the sample in the treatment group ; the fraction of the sample in the treatment group fraction of the sample in the treatment group representative estimate of the variance
Explanation / Answer
1.In Fisher’s example, the null hypoth-
esis is given by H0 : Yi(1) Yi(0) = 0 for all i and an alternative is H1 : Yi(1) Yi(0) 6= 0 for at
least some i. This null hypothesis is said to be sharp because the hypothesis is specified for each
unit. A sharp null hypothesis is strong in that it assumes zero effect for every unit i.
Answer is option B
