AP STATS TEST 10/19

census

using every member of a population to make an inference or conclusion

sample survey

using part of a population as a representation

representative

a sample thats statistics accurately represent the population parameters

sampling frame

list of individuals from which the sample is drawn

sampling variability

samples drawn at random generally differ from one another

SRS

every individual has the same chance of being selected

stratified sampling

taking a population and grouping like things together, then doing an SRS within each group (strata)

groups in stratified sampling

strata

cluster sampling

breaking date into groups and then randomly selecting a single cluster and doing a census on the cluster

multistage sampling

sampling schemes that combine several methods

systematic samples

drawing a sample by selecting individuals systematically

target sample

the individuals for whom you intend to measure responses

your sample

the people who actually responded or were chosen

convenience sample

a sample chosen primarily based on how easy they are to find

reponse bias

when respondents give answers that don’t actually reflect their perspective

voluntary response sample

when someone can choose to participate or not to participate

nonresponse

when someone was eligible to be chosen but wasn’t able to respond and thus is not counted

undercoverage

a sample in which a portion of the population is not represented well or not eligible to be chosen

when you describe bias, make sure you include the ________ of the bias

direction

observational study

a study where researchers simply observe data rather than create data

retrospective study

a study done by looking at past results

experimental units

the thing the study is done on; when the experimental units are humans they are called subjects or participants

lurking variable

a factor that influences both the explanatory and response variable which we have not considered (summer —> drowning and ice cream ^)

confounding

a factor that influences our response variable which we have not controlled; to the best of our abilities, we try to make sure these confounding variables are equally distributed into each population (exercise more, creatine —> stronger)

experimental study

a study done specifically to prove a cause-and-effect relationship

factor

the explanatory factor that causes the response variable to be predictable (ex: number of calories eaten, amount of exercise)

levels

the amount of the factor that each subject gets (ex: 2000/3000 cal, little/lot of exercise)

treatment

each combination of levels (ex: 2 factors with 2 levels each is 4 total treatments)

an experiment ____ factor levels to create treatments

controls

an experiment subjects are assigned treatment levels _____

randomly

an experiment ____ the responses of treatment levels

measures

with sampling, we used randomness to select participants, _____________

in experiments, we randomly assign treatments

control

we must reduce lurking and confounding as much as possible

randomize

this allows us to reduce the effects of unknown sources of variation (without randomization studies cannot be considered valid)

replicate

our sample size must be sufficiently large to draw a conclusion

blocking

when we have factors, we cannot control which may affect our outcomes, we can first group similar individuals together then use a random assignment (blocking is to experiments as stratifying is to sampling)

matched pair

we are comparing individuals that are linked and get different treatments

examples of matched pairing:

old shoe on right new shoe on left, twins, comparing you to past you

statistically significant

we believe that the results are so conclusive that they could not have randomly occurred by chance

control group

a group who has no treatment applied to it

blinding

when either the participants or the judges in the experiment do not know which treatment was assigned to each group

single blind

when the participants do not know which treatment they recieved

double blind

when neither the participants nor the judges know which treatment was received

placebo

when we have a group who receives a treatment that should not affect the result

the best experiments are ___, ___, ___, and _____

randomized, controlled, double-blind, and have a control group