Biostatistics

Question 1

Point estimate

Answer 1

Single numbers derived from a sample

- represents our best estimate of the true value of the much larger population parameter

Question 2

What are examples of point estimates derived from samples of populations?

Answer 2

• mean
• proportion
• OR
• relative risk

Question 3

Sampling

Answer 3

Measure only a fraction of individuals in the population of diverse individuals

• may or may not reflect the population
• sample mean value may not be the exact mean value of the population

Question 4

Sampling leads to ______

Answer 4

Error

• the more you sample, the less your error
• on any one sample, you do not know how far away from the true mean you are

Question 5

Confidence interval

Answer 5

Range of values that the point estimate could reasonably take
- always accompanied by a probability statement

Question 6

95% confidence interval

Answer 6

If we repeated sampling of the population many times, and for each sample constructed a 95% confidence interval, then 95% of these intervals would contain the true population parameter

Question 7

You assume that the samples are ______

Answer 7

Independent, randomly selected (no bias), and sufficiently large

Question 8

Common CI inference

Answer 8

The 95% CI contains the population parameter 95% of the time
- applies for means, proportions, RR, OR

Question 9

P value

Answer 9

Probability of obtaining the observed value (or more extreme value) when the null hypothesis (Ho) is true

Question 10

Ho

Answer 10

States there is no difference between the groups being compared
- the groups being compared came from one population, and the difference you see is only sampling error (aka: random error)

Question 11

If the p-value is small, then ______

Answer 11

The probability of obtaining the observed (or more extreme) value is small, under the assumption of Ho

Question 12

What is “small”

Answer 12

Usually 0.05 or below

• if the p value is small, we can reject Ho
• if the p value is not small, we cannot reject Ho

Question 13

If we reject Ho ______

Answer 13

You can accept the alternative hypothesis (Ha)

• states that there is some difference between groups or with a fixed value
• say that the 2 groups are probably not from one population but from distinct populations
• statistically different

Question 14

With the null hypothesis, the 2 groups really are ______

Answer 14

From one population

- the difference that is seen is due to sampling error

Question 15

If the p value is small, then the probability of getting the observed (or more extreme) value, due to sampling error, is _____

Answer 15

Small

- reject Ho, accept that there is some difference between groups (they come from separate populations)

Question 16

If the p value is not small then we cannot reject Ho because ______

Answer 16

There is not sufficient evidence to reject it