Biostatistics Final Exam Review updated
already passed
What are the two principles of random sampling? - correct answer ✔✔Equal chance and independence
What are the three ways to reduce bias in a study? - correct answer ✔✔Control group, randomization,
blinding
What are the three ways to reduce sampling error? - correct answer ✔✔Replication, balancing, blocking
What is blocking/stratification? - correct answer ✔✔Grouping subjects by one shared property and then
randomizing within each group what subject gets what treatment
Sensitivity - correct answer ✔✔Chances that you test positive given that you have the disease
Specificity - correct answer ✔✔Chances that you test negative given that you don't have the disease
PPV - correct answer ✔✔Chances that you have the disease given that you tested positive
NPV - correct answer ✔✔Chances that you do not have the disease given that you tested negative
What measure of spread do you report along with the mean? - correct answer ✔✔Standard deviation
What measure of spread do you report along with the median? - correct answer ✔✔Interquartile range
(IQR)
When would you report the median instead of the mean? - correct answer ✔✔When there are outliers
or a skew in the distribution
, What is a Type I error? - correct answer ✔✔When you reject the null but the null is actually true
What is a Type II error? - correct answer ✔✔When you fail to reject the null but the null is actually false
What is the probability of a Type I error? - correct answer ✔✔Probability equals the significance level,
"alpha"
What is the probability of a Type II error? - correct answer ✔✔Probability represented by "beta"
What is power? - correct answer ✔✔Power is the probability that the null will be correctly rejected (also
1 - "beta")
Why do you not want to use non-parametric tests unless you have to? - correct answer ✔✔Non-
parametric tests have a high value for beta, and therefore low power, so you are less likely to see
significance
What are the assumptions for a paired design 2-sample t-test? - correct answer ✔✔Random sample and
that the differences between measurements are distributed normally
What are the hypotheses for a paired design 2-sample t-test? - correct answer ✔✔Null: average
difference between treatments is 0
Alternative: average difference between treatments is NOT 0
What are the assumptions for an independent 2-sample t-test? - correct answer ✔✔Both are random
samples, variable of interst is normally distributed in both, and that each has the same variance
What are the hypotheses for an independent 2-sample t-test? - correct answer ✔✔Null: mean of pop. 1
equals the mean of pop. 2
Alternative: mean of pop. 1 does NOT equal mean of pop. 2
How do you check for equal variances to fit all assumptions for independent 2-sample t-test? - correct
answer ✔✔Use an F test
already passed
What are the two principles of random sampling? - correct answer ✔✔Equal chance and independence
What are the three ways to reduce bias in a study? - correct answer ✔✔Control group, randomization,
blinding
What are the three ways to reduce sampling error? - correct answer ✔✔Replication, balancing, blocking
What is blocking/stratification? - correct answer ✔✔Grouping subjects by one shared property and then
randomizing within each group what subject gets what treatment
Sensitivity - correct answer ✔✔Chances that you test positive given that you have the disease
Specificity - correct answer ✔✔Chances that you test negative given that you don't have the disease
PPV - correct answer ✔✔Chances that you have the disease given that you tested positive
NPV - correct answer ✔✔Chances that you do not have the disease given that you tested negative
What measure of spread do you report along with the mean? - correct answer ✔✔Standard deviation
What measure of spread do you report along with the median? - correct answer ✔✔Interquartile range
(IQR)
When would you report the median instead of the mean? - correct answer ✔✔When there are outliers
or a skew in the distribution
, What is a Type I error? - correct answer ✔✔When you reject the null but the null is actually true
What is a Type II error? - correct answer ✔✔When you fail to reject the null but the null is actually false
What is the probability of a Type I error? - correct answer ✔✔Probability equals the significance level,
"alpha"
What is the probability of a Type II error? - correct answer ✔✔Probability represented by "beta"
What is power? - correct answer ✔✔Power is the probability that the null will be correctly rejected (also
1 - "beta")
Why do you not want to use non-parametric tests unless you have to? - correct answer ✔✔Non-
parametric tests have a high value for beta, and therefore low power, so you are less likely to see
significance
What are the assumptions for a paired design 2-sample t-test? - correct answer ✔✔Random sample and
that the differences between measurements are distributed normally
What are the hypotheses for a paired design 2-sample t-test? - correct answer ✔✔Null: average
difference between treatments is 0
Alternative: average difference between treatments is NOT 0
What are the assumptions for an independent 2-sample t-test? - correct answer ✔✔Both are random
samples, variable of interst is normally distributed in both, and that each has the same variance
What are the hypotheses for an independent 2-sample t-test? - correct answer ✔✔Null: mean of pop. 1
equals the mean of pop. 2
Alternative: mean of pop. 1 does NOT equal mean of pop. 2
How do you check for equal variances to fit all assumptions for independent 2-sample t-test? - correct
answer ✔✔Use an F test
