Statistical inference - Method of looking at statistics and drawing conclusions about a population.
Gathering data -
Simple Random Sample (SRS)
- Every observational unit has an equal chance of being included. (In
an ideal world - Doesn’t always work and not the best for all cases)
Stratified Sampling - Divide the population into strata (e.g. men and women) and then perform a random sample of these strata
Systematic Sampling
- Use when you can’t get a random sample, assign a number, then pick
every occurrance of that number. (e.g. Roman legions, killing every 10th
person).
Cluster Sampling
- Creating samples from groups called clusters and randomly selecting
there-in (e.g. zip codes, area codes, floors of a building)
Convenient Sampling - Take sample of those immediately available to you (Poor representation of the population)
Target population - Population we are interested in collecting data on
Bias - Systematic elimination/exclusion of a segment within the target population
Types of Studies:
- Anecdotal Studies - Like an opinion, not scientific, based on limited experience (“useless”)
- Survey Studies - Great source of obtaining opinions, not great for truth/fact
- Observational Studies - “Sit back and see what happens.” No cause and effect relationship
- Experimental Studies - “Impose something” Requires some form of randomization (eliminates bias) and a control group (receives no treatment) which enable us to create a cause and effect relationship.