Circle the best choice

1. Which of the following situations does Categorical Data Analysis deal with?

A categorical response variable and continuous exposure variable(s)

A categorical response variable and categorical exposure variable(s)

Categorical exposure variable(s)

Both (a) and (b)

2. According to a SAS output, the point estimate of the odds ratio with respect to gender and lung cancer is 1. Which of the following statements is true with regards to the true odds ratio?

The true odds ratio is 1

The true odds ratio is close to 1

The true odds ratio is far from 1

It is hard to say where the true odds ratio is

3. According to a SAS output, (1.3, 1.5) is the 90% confidence interval of the odds ratio with respect to gender and lung cancer. What can one say about the true odds ratio?

The true odds ratio could be outside of (1.3, 1.5), and the chance for this to occuris 10%

One has 90% confidence in believing that the true odds ratio is in (1.3, 1.5)

One can reject the null hypothesis at the significance level, 0.1, that the true oddsratio is 1

All of the above

4. According to a SAS output, 0.049 is the P-value for testing the null hypothesis of no association of gender with lung cancer. Which of the following statements can be made based on this P-value?

The null hypothesis should be rejected

The null hypothesis should be accepted if the significance level is 0.048

The null hypothesis should be rejected if the significance level is 0.05

One fails to reject the null hypothesis

5. Which of the following statements is NOT true with regards to asymptotic tests and exact tests?

It is up to ones preference which of these two types of tests should be used

An asymptotic test can be used only if the sample size is sufficiently large

An exact test should be adopted if the sample size is too small.

Which of these two types of tests should be used depends on the sample size

6. Which of the following tests can be used to test whether or not the proportion of lung cancer is equal to a hypothesized value in Florida?

T test

Binomial test

Pearson Chi-Square test

F test

7. Assume that Dementia is classified into four categories: Dementia free, Alzheimer Disease, Vascular Dementia, and others. Which of the following tests can be used to test whether or not the four Dementia subtype proportions in Florida population are equal to their corresponding hypothesized values ?

Pearson Chi-Squared test

Chi-Squared Goodness Fit test

Z test

All of the above

8. Which of the following are the two measures usually used to measure the strength of the association of a binary exposure variable with a binary response variable?

Odds ratio and difference of proportions

Relative risk and odds ratio

Relative risk and difference of proportions

Odds and difference of proportions

9. Which of the following quantities is approximately equal to odds ratio if the data is collected from a case-control study and the disease is rare?

Relative risk

Odds

Difference of proportions

All of the above

10. The null hypothesis of no association of smoking with liver cancer can be tested using?

Pearson Chi-Squared test

Chi-Squared Goodness-fit test

Cochran-Mantel-Haenszel test

Both (a) and (c)

11. The 95% confidence interval of the odds ratio with respect gender and lung cancer can be used to test?

Whether or not gender is associated with lung cancer

Whether or not gender is independent of lung cancer

Whether or not the odds ratio between gender and lung cancer is equal to one

All of the above

12. The null hypothesis that the odds ratio of gender associated with lung cancer is equal to 2 can be tested using?

Pearson Chi-square test

The 95% confidence interval of odds ratio

Fishers Exact test

None of the above

13. Which of the following variables is a confounding variable?

A variable that is associated with exposure variable but not response variable

A variable that is associated with response variable but not exposure variable

A variable that is associated with neither exposure variable nor response variable

None of the above

14. Suppose that it is known that tea drinking is not associated with Alzheimer Disease (AD), and women are more likely to drink tea, which of the following statements is true?

The marginal odds ratio with respect to gender and AD can be used to measure the effect of gender on AD

The marginal odds ratio with respect to gender and AD overestimates the effect of gender on AD

The marginal odds ratio with respect to gender and AD underestimates the effectof gender on AD

Tea drinking is a confounder

15. Which of the following cases could lead to the situation where marginal odds ratio of an exposure variable, X , associated with a response variable, Y , equals one but the conditional odds ratios on a third variable, Z, are greater than one?

Z is positively associated with both Y and X

Z is positively associated with Y and negatively with X

Z is negatively associated with Y and positively with X

(b) and (c)