ap stats unit 1

response variable

measures an outcome of a statistical study

explanatory variable

may help explain or predict changes in a response variable

factors

explanatory variables that are manipulated and may cause a change in the response variable

levels

different values of a factor

treatment

a specific condition applied to the items or individuals in an experiment

experimental unit

the item or individual to which a treatment is assigned (often called subjects when human beings)

retrospective

past or current data

prospective

data into future

confounding variable

impacts results of an experiment or observational study

experiment

suggests cause-and-effects relationships

observational study

suggests an association between variablesanr

random sample

randomly selected portion of a population that allows us to generalize our results to the entire population

simple random sample (SRS)

• every individual is equally likely to be selected

• every possible group is equally likely to be selected

systematic random sample

• choose a random k^th starting point from a list of entire population

• select every n^th student until your sample is filled

• only can be done with an organized list

stratified random sample

• separate population into smaller subgroups

• select some from ALL

• guaranteed to have subgroups represented

cluster random sample

• select ALL from some

• easier to accomplish since people are “closer together”

convenience sample (poor sampling method)

surveying people that are all close by

voluntary response sample (poor sampling method)

no one is selected, all responses come from volunteers

bias

consistent over/underestimation of true value

undercoverage

people in a population are not able to be selected for some reason

response bias

people feel pressure to respond a certain way (i.e. anonymity)

nonresponse bias

people who were selected don’t respond or cannot be contacted

random assignment use

determine cause & effect

random selection use

generalize

experimenting well

1. comparison: at least two groups

2. random assignment: create similar groups by chance to make it easier to establish cause & effect (limit confounding variables)

3. replication: larger sample sizes or gather more data

4. control: directly control other possible influences

random assignment

1. assign/label numbers

2. use a RNG to generate different numbers

3. assign treatments in context

placebo effect

subjects think a fake treatment works

blinding

single or double; subjects and/or researchers unaware of groups

completely randomized design

assign treatments using RA using splitting subjects beforehand

block design

split subjects into blocks before RA treatments, eliminating confounding variables

matched pairs

(form of block design) pair up before RA treatments within pairs

• similar to themself

• pair w/ yourself

  • must randomize the order of the treatments