AP statistics unit 3

generating random numbers on calculator

MATH -> PRB -> 5:randInt(minimum, maximum, amount)

simulation

mimic reality using random numbers

elements of simulation

think, show, tell

repeats in random numbers should be skipped if

the probability changes as simulation continues (no replacement)

voluntary bias

respondents decide who is in the sample. Those who voluntarily respond probably differ from the population.

undercoverage bias

The sampling frame doesn’t match the population of interest. Certain people cannot be selected for the sample.

nonresponse bias

A large % of those selected don’t respond. The small group who do probably aren’t representative

response bias

Anything in the survey design that influences responses. This could be the question wording, who is asking, where it’s being asked, etc.

sampling error/variability

not error, just sample to sample variation

simple random sampling (SRS)

every person have the same chance of getting selected

stratified sampling

split population into characterized groups, then SRS in each group (ex: split into male and female, then sample 5 of each randomly)

cluster sampling

split population into random clusters, and pick a few to sample everyone in the cluster.

systematic sampling

non random, sample by pattern (ex: sample every 10th person)

multistage sampling

combing multiple sample methods

retrospective observational studies

looking at past behavior

prospective observational studies

looking at current behavior

principles of experimental design

control make conditions similar, replicate (each treatment has multiple objects), random, block (divide into blocks based on result like matched pairs)

quota sampling bias

no randomization in choosing sample

census

sampling the entire population

experiment

impose treatments on objects, not observational