In large samples, the sampling distribution of the risk difference is approximately ?â€¨
Plus-four confidence interval method for a difference in proportions is accurate in samples as small as â€¨
a. 100 per group
b. 50 per group
c. 25 per group
d. 5 per group
Which of the following is not a requirement for consideration before continuing with the calculation of a sample size in an observational study observing the difference in incidence of disease X based on exposure?â€¨
a. Projected drop-out rate to inflate the sample size estimate based on the projected loss to follow-up in a study.
b. A predetermined power for the study to detect a difference when one actually exists.
c. The correlation coefficient between the two groups being compared.
d. The informed estimation of the incidence in both groups being compared.
Which of the following is not an example of systematic error in an observational study?â€¨
a. A cross-sectional study recruits participants that are willing to sign up outside of a major university and meet the inclusion and exclusion criteria to take part in the survey relating unsafe sex habits to STIs.
b. A researcher is interested in the relationship between coffee drinking and lung cancer, and after careful multivariate linear regression modeling determines that a significant percentage of the relationship is due to another variable, cigarette smoking.
c. An observational study recruits participants for a study looking at Alzheimerâ€™s disease due to exposure to industrial hazards by asking participants to recall their exposure over the past 10 years.
d. A data-entry specialist responsible for adding in fasting glucose levels to a database accidentally skipped an observation during the input phase of data cleaning.
Proportions are tested for a significant _____ difference?â€¨
What is not a method for testing proportions for significance?â€¨
a. z test (large sample)
b. Fisherâ€™s exact procedure (small samples)
c. the chi-square test
d. the Wassermann test
In calculating for tests of proportions for small samples (fewer than 5 successes expected in either group), avoid the z test and useâ€¨
a. the exact Fisher or Mid-P procedure
b. the asymptotic procedure
c. the random event generator
d. the special theory of relativity
Before conducting Fisherâ€™s test, data are rearranged to form a â€¨
a. 1-by-2 table
b. 2-by-1 table
c. 2-by-2 table
d. A single row table
The best way to calculate a p value for a Fisherâ€™s Exact is toâ€¨
a. Use an adding machine
b. Use computer program
c. Call a statistician
d. Use pencil and paper
The RR (risk ratio or prevalence ratios) is a risk â€¨
a. multiplier, e.g., an RR of 2 suggests that the exposure doubles risk
b. adder, e.g., an RR of 2 suggests that the exposure sums risk
c. divider, e.g., an RR of 2 suggests that the exposure divides risk
d. neutralizer, e.g., an RR of 2 suggests that the exposure neutralizes risk
When p1 = p2 , RR = 1. â€¨
a. This is the â€œmultiplier RR,â€ indicating increased association.
b. This is the â€œbaseline RR,â€ indicating no association.
c. This is the â€œsmall effect RR,â€ indicating slightly increased association.
d. This is the â€œpopulation RR,â€ indicating population prevalent association.
An RR of 1.15 indicates â€¨
a. No association;
b. a negative association; specifically, 15% lower risk (in relative terms) with exposure.
c. a positive association; specifically, 15% higher risk (in relative terms) with exposure.
d. a positive association; specifically, 115% higher risk (in relative terms) with exposure.
In observational studies, systematic errors are often _____important than random sampling error?â€¨
d. None of the above
The three types of systematic error considered do not includeâ€¨
b. Information bias
c. Selection bias
d. Random sampling error
Nondifferential misclassification tends to bias results â€¨
a. Away from the null and show a positive effect.
b. can bias results in either direction.
c. toward the null (or have no effect).
d. Away from the null and show a negative effect.