Sampling Methods
People who choose themselves to be in the sample (usually people with strong opinions)
Voluntary Response
Inaccurate responses (lying or confusing questions)
Response bias
Contains comparison, random assignment, a control group, and able to be replicated
Well-designed experiment
Individuals assigned treatments (humans often called subjects)
Experimental units
Observational study
Choosing individuals easy to reach
Convenience
When members of the population are less likely to be chosen for a sample
Undercoverage bias
A group receiving an inactive treatment
Control group
Variable whose levels are manipulated intentionally
Explanatory variable
Examine existing data
Retrospective
Split the population into groups of similarity (strata), then take an SRS of each group. Those chosen from each SRS will be selected for the sample
Stratified Random Sample
Individuals selected to be in a sample but can't be contacted or refuse
Nonresponse bias
Neither the subjects nor the experimenters who interact with the subjects know which treatment a subject is receiving
Double-blind experiment
Outcome from the treatments administered
Response variable
Track individuals into the future
Prospective
Split population into groups based on location (clusters), then randomly select clusters (1 SRS). All individuals within the cluster are selected for the sample.
Cluster
A teacher wants to survey students on how often they cheat on tests
Response bias
When experiment units have a response to a placebo (fake treatment works)
Placebo effect
Variables that influence the response variable
Confounding variables
A study collects data on the grades of all high school students at a given school. In addition to this, they ask whether each student attended Pre-K. They attempt to see whether the average grades for those who did attend Pre-K are higher than those who did not attend.
Randomly select a starting point and every nth individual thereafter will be selected for the sample
Systematic Random Sample
A mayor wants to find out the average age of people in his city. After obtaining a list of all landline telephones in his city, he conducts a simple random sample of 300 people.
A survey with undercoverage bias
A pair of experimental units that are matched based off similarity then randomly assigned to each treatment
Matched pairs design
When results from a study are too unusual to have occurred purely by chance (=< 5% of results as extreme)
Statistically significant
A pharmaceutical company develops a new drug to treat seizures. In their clinical trial, they have 100 people who are administered the treatment. After a year, they see whether they have fewer seizures than before.
Example of a prospective study