Chapters 12-13 Flashcards
survey
asking questions of
subjects
population
an entire group of
individuals
census
Everyone in the entire population is surveyed - ideal but impossible
sample
subset of a population
proper sample must be:
unbiased, random, with a size that represents the population
simple random sampling
First, establish a sample frame. then select a sample.
sample frame
a list of individuals from which to draw the sample
stratified random sampling
First, the population is divided into homogeneous groups called strata (stratum singular); second, a SRS is drawn from each stratum.
Cluster Sampling
First, divide the population into similar heterogeneous groups in which each
cluster represents the population; second, choose one or few clusters at
random; third, survey the census of each chosen cluster.
Systematic Sampling
A methodical system is assigned in the selection process of subjects. Every _th person
Multistage Sampling
combination of a bunch
Voluntary Response Sampling (bad)
a large group invited to respond
Convenience Sampling (bad)
respondents are chosen because they are easily accessible to the surveyors
Sources of Bias
undercoverage, non-respondent, and response bias
Observational study
Researchers observe subjects in the field and gather data
types of observational study
retrospective and prospective
Experiment
A treatment applied to subjects to test a cause and effect relation
factor in an experiment
explanatory variable
response variable in an experiment
dependent variable
treatments in an experiment
different combinations of factors
samples have to be
unbiased, randomly chosen, and representative of the population
Blocking
The subjects are grouped homogeneously, then the subjects of
each block are randomly assigned a treatment.
control
The subjects do not undergo a treatment and they know that. A
control is counted as a treatment.
Placebo
the subjects are given a treatment (a sugar pill) which has no effect,
but they do not know that.
Matched Pair Design
a special case of a randomized block design
simulation
models a real life scenario by using random numbers that specify the particular event
