ap stats - unit 1 part 2 quiz vocab

experiment

imposes treatments on experimental unites to measure responses

response variable

outcome that is measured from an experiment or observational study

explanatory variable

explains/predicts changes in experiment/observational study

factors

categorical explanatory variable that divides data into distinct groups, typically has levals of factors

observational study

collects data without interfering, retrospective/prospective

survey

common set of questions given to those who participate in observational study

extraneous variables

variables outside the explanatory variable that may affect the response variable

confounding variable

is an extraneous variable that is associated w/ explanatory and response variables, make sure to explain both parts when justifying whether it is a confounding variable

random sampling

randomly selecting members of population for a sample

investigative question

identify variable of interest, type of inference, type of conclusion that can be made (cause-effect conclusion, population, etc)

simple random sample

all in population have equal chance to be selected without replacement, e.g. numbering from 1-10

stratified random sample

selecting SRS from each stratum and combining into one sample (strata should be homogenous)

cluster random sample

selected by SRS of just CLUSTERS and including every individual in the CLUSTER (clusters shd be heterogeneous)

systematic random sample

selected from ordered arrangement (numbering); selecting an interval k, selecting one of first k items and then choosing every kth item after

multistage sampling

combines 2+ sampling methods

convenience sample

(bad) samples those who are easy to reach - can lead to bias

bias

sample statistic is likely to underestimate/overestimate population parameter because of a flaw in data collection (e.g. convenience, voluntary response, undercoverage, nonresponse)

voluntary response sample

in which participants choose to participate or respond rather than a random sample selection (bad. could → bias)

undercoverage

in which some individuals are less likely to be chosen for a sample for whatever reason, could → bias

sampling frame

a list with every member of population to take a sample from

nonresponse

when individuals chosen for a sample are unable to respond - could lead to undercoverage/bias

response bias

in which responses to survey/questions consistently differ from the truth

control group

used to provide a baseling for comparing effects of treathments, prevents confounding - related to comparison, which is an element of a well designed experiment

placebo

a treatment with no active ingredient → placebo effect

double blind

where neither the subject nor the giver of the treatment knows which treatment is being given

single blind

in which either the subject or the giver knows which treatment it is

random assignment

one element of a well designed experiment, uses a chance process to determine which treatment is given

completely randomized assignment

experimental units assigned to treatments completely at random, allows the researcher to make cause-effect conclusions (no blocks at all)

replication

multiple experimental units are assigned to each treatment (another important element)

direct control

keeping values of extraneous variables the same (another important element)

block/block variable

group of experimental units known to be similar (like strata, but for treatment rather than sampling)

randomized block design

random assignment of experimental units to treatments separately done in each block

matched pairs

blocks of size 2 where the two are given similar or opposite treatments

statistically significant

when an observed difference in responses is so large it cant be explained by change variation in the random assignment

randomized distribution

distribution of statistic generated by repeastedly assigning the response values to the treatment groups