APSTAT U1 (4.1)

experiment

deliberately imposes some treatment on individuals to observe their responses

observational study

observes individuals and measures variables of interest but doest not attempt to influence the responses

population

group of individuals we are drawing conclusions about from a study.

census

studying every individual from the population (possible sampling method, but not ideal)

sample

investigating a smaller group of individual from the population

biased

sampling methods that systematically favor certain outcomes

design

method used to choose the sample from population

Biased sampling methods:

voluntary response sample, convenience sample, (maybe systematic sampling, when a hidden pattern or repeating cycle in the population list matches the chosen fixed sampling interval)

voluntary response sample

people who choose themselves to respond
(often elicits strong opinion, middle opinions not received) could be both pos/neg

convenience sample

chooses individuals easiest to reach
(only listens to one opinions bc its not randomized)
(no full picture, representation, of the whole population)

Unbiased (random) sampling methods

simple random sample (SRS), stratified random samples, cluster samples, systematic sampling

Simple random sample (SRS)

n individual has an equal chance of being selected
1) gives each individual an equal chance of being chosen

2) gives every possible sample of size n (group) an equal chance of being chosen

assign each individual a number,
(*each nomber should have an equal amount of digits, no repeated numbers)

01-33, ignore 00, 33-99, ignore repeated numbers (=only count once even repeated many times)

strata

groups of similar individuals,

stratifying by race, income, age and other characteristics can be important

samples taken within statum vary less, so estimates can be more precise, and therefore reduce sampling variability

Stratified random sample

divided into strata, then chose within stratum and combined to make full sample
(could be different size of each strata, just % of them, unless said otherwise)

Custer samples

split populations into groups such that each group is representative of the population
randomly selected group and conduct census (everyone) within that group

(lunches a,b,c,d)

systematic sampling

random individual is chosen from population then every nth individual is selected (following a system)

multistage samples

sampling schemes that combine several methods

undercoverage

portion of the population is left out of the process of choosing the sample

nonresponse biase

individual chosen for sample cant be contacted or does not cooperate

response bias

design that influences the reponses (respondent may lie, especially if asked about illegal or unpopular behavior)

question wording bias

questions on surveys are confusing, misleading, or one sided
(type of response bias)