1.

a)

This is false. We can not conclude that the rate of patients with cardiovascular problems is higher by just looking at the numerators.

b)

True

c)

This is false. Just because the data suggests patients taking rosiglitazone are more likely to have cardiovascular problems in this study does not prove rosiglitazone causes heart damage.

d)

True

2.

a)

In the control group, 30/34 patients died. In the treatment group, 45/69 died.

b)

i)

The claim being tested is whether the treatment causes an decreased rate of death after heart transplants. The null hypothesis is that treatment will not decrease death rate. The alternative hypothesis is that treatment will decrease death rate.

ii)

28, 75, 69, 34, 0, negative

iii)

The simulation results seem to be relatively evenly distributed. This would suggest that we should fail to reject the null hypothesis.

3.

a)

Experiment

b)

Yes this study utilizes blinding. By giving a placebo participants do not know if they are in the treatment or control group.

c)

-0.026

d)

At first glance, placebo seems to be more effective because the placebo group had a higher rate of improvement in symptoms.

e)

Null Hypothesis: Treatment with antibiotic causes no improvement in sinusitis symptoms. Alternative Hypothesis: Treatment with antibiotic causes improvement in sinusitis symptoms.

f)

Rather than using the number of patients who reported improvement in symptoms, the difference of the rate of patients reporting improvement in the treatment group and rate of patients reporting improvement in the placebo group should be used as a summary statistic. This type of test requires proportions.