Studied by 0 people
SRS (Simple Random Sample)
each person gets randomly assigned a number, numbers are picked randomly
Stratified Random Sample
split people into groups and survey a few selected people from each group
Cluster Sample
survey all people in one group
Systematic Sample
start from a certain point and count by a certain amount, survey on each interval
Voluntary Response
bad - people elect to answer, biased
Convenience Sampling
bad - people nearest to you/easiest to survey
undercoverage
not enough perspective from different groups
non-response
people elect not to respond
response bias
providing too much info, not just a question, influencing answers
self-reporting
people can lie
Completely Randomized
no important variables like gender, weight, income, etc., general random testing
Randomized Block
distributed for important variables like gender, weight, and treatments are randomly distributed
Matched pairs
do one treatment first and the second later, only used with two variables
