Sampling and Distributions

Question 1

Representative Sample

Answer 1

Attempt to have the sample look just like the population; every member of the population has an equal chance of being selected
(ex. simple random sampling -> lists every member of population and randomly select from list; stratified random sampling -> list every member of pop., identify subgroups, randomly select proportionally from subgroups)

Question 2

Non-Representative Sample

Answer 2

Systematically different from the population (ex. voluntary response sample -> individuals choose to respond to a general appeal for information; convenience sample -> includes the people that are easiest to contact and measure)

Question 3

Quota Sample

Answer 3

Select same number of people regardless of their prevalence in pop. (50 democrats and 50 republicans)

Question 4

Sampling Error

Answer 4

Differences between sample and population due to random chance; increasing sample size decreases sampling error; results from collecting data from some (not all) members of population

Question 5

Sampling Distribution

Answer 5

Distribution of sample central tendency from a large number of sample (take low samples from pop., then calculate mean for each sample and make distribution of those means); builds a sampling dis. centered at pop. mean; normally distributed

Question 6

Standard Normal Distribution

Answer 6

Every individual/observation is under curve; area=100%; the proportion of population in a range of values is based on M and SD of normal distribution; tells us that 68% of pop. is +/- 1 SD of M

Question 7

Error

Answer 7

Difference between a sample and population

Question 8

Making Comparisons

Answer 8

You don’t just compare the actual sample average (proportion), also compare underlying population distribution

Question 9

Margin of Error

Answer 9

How much the sample mean (or proportion) differs from the value in population; each sample includes some random error so we need M.O.E.

Question 10

Error Bars

Answer 10

If they overlap: there is no evidence of a significant difference; if we had a different random sample, there might be no difference between the groups
If no overlap: there is likely evidence of a significant difference

Question 11

MOE Calculations

Answer 11

For categorical variables: MOE is based on sample size so if sample size increases, MOE decreases
For continuous variables: MOE is based on sample size and variation; can be represented using standard error of mean and confidence interval

Question 12

Standard Error of Mean (SEM)

Answer 12

Based on both sample size and variation

Question 13

Confidence Intervals (95%)

Answer 13

Confidence interval: the interval within which the population parameter is believed to be

Question 14

Confidence Level

Answer 14

The probability that the confidence interval contains the parameter; multiply the SEM to reach C.I.

Question 15

Calculate 95% CI

Answer 15

Calculate SEM, then MOE, then CI (mean+/- MOE)

Question 16

Categorical Variables

Answer 16

Focus on sample size

Question 17

Continuous Variables

Answer 17

Consider variability