AP Stats Unit 3

Simple random sample

Every individual has an equal chance at being selected, least variation, every combo is equally likely, RNG

Stratified random sample

Divide pop into groups based on similarity, homogenous, each strata is random, least variation, everyone is represented

Cluster sample

Heterogenous cluster, group-cluster, pop divided into clusters and survey everyone in that group, more variation

Systemic random sample

Put pop in random orders, choose starting point, select every kth individual, most variation

Convenience random sample

Pick who is easily available, more variation

Voluntary sample

An easy, cheap, and biased survey where individuals choose to participate, more variation

Under coverage

Some individuals are less likely to be chosen ( left out / no chance )

Nonresponse

Individual can't be reached or won't respond

Response bias

Self-reporting, wording of the question

Population

Entire group of interest, parameters

Sample

Subset of population, statistics

Groups

Strata

Causation

Needs an experiment

Experiment

Treatment is being imposed, randomly assigned treatment, causation

Observational study

observe and measure variables, no treatment, only association, no cause and effect can be concluded

Treatment

What is done ( or not ) to the experimental units

Experimental units

Who / what gets the treatment

Confounding variables

more than one variable that could have caused the observed effect in an experiment

Blinding

Unit doesn't know

random sample

generalize to a population, association

random assignment

treatment causes response, causation

sample conclusion

For the [population], there is an association between [explanatory] and [response]. 

assignment conclusion

For the [population], [explanatory] causes [response]. 

block

group of experimental units that are similar

random block design

separate subjects into blocks then randomly assign treatments within each block

comparative experiment

the effects of two or more treatments are compared

bias

systematic favoring of certain outcomes, identify who is over or under represented, describe how that affects results

factor

variable being tested or measured

levels

specific value of the factor

principles of a good experiment

replication, randomization, control, and comparison

replication

use enough subjects to reduce the role of chance variation

randomization

randomly assign treatments to balance confounding variables

control

keep all other variables constant

comparison

compare results between groups

control group

receives no treatment or placebo to provide baseline

placebo

fake treatment that looks real

completely randomized design

all experimental units are randomly assigned to treatments

matched pairs design

each subject receives both treatments in random order, or subjects are paired by similarity and randomly assigned treatments within each pair