Imp. Chapter 4 Flashcards
3 pillars
Random-divide into groups
Replication- red. var
Control- eliminate/control out. var
Nonresponse bias
Majority of population didn’t answer
Response bias
Pressure in sampling method
Under coverage
When individuals in population are excluded from sampling
Simple random sample
Number three population, choose sample by random
Use in experiments to divide evenly
Stratified random sample
Split population into groups, then sample from each group
Cluster
Split population into groups. Sample. Then choose ENTIRE groups to sample
Systematic
Number the population, randomly choose a place then choose every _th person
Multistage
Divide groups into equal areas and randomly choose areas
Experiment key words:
Assigned/imposed treatments
Completely randomized experiments
Simple random sample of experiments
Block design
Stratified experiment
Matched pair design
Same experimental unit of very similar through all treatments in random order
Lurking confounding variable
A variable not among explanatory or response that effects outcome. Essentially outside variables
More samples=?
Variability