PSY 313 Lecture Notes - Lecture 4: Stratified Sampling, Observational Error, Scatter Plot
Document Summary
Assignment how we place that sample into conditions of the experiment. Sampling how we select people from the population. Bias: systematic difference between sample and population: sample differs from population on important decisions, avoid bias by developing a sampling plan. Stability: the spread/variance of the sample (how much noise in the data: high stability = low spread! Unbiased and stable is the best: the margin of sampling error in a poll is based on stability, avoid an unstable sample by using a sufficiently large sample size. Positively skewed tail points toward + end || negatively skewed tail points negative end. Non-probability sampling not drawing from entire population: quota: selectively take what is available according to a plan, convenience sampling: take whatever you can get. Stratified: break population into sub-samples (5 different brackets of diff. incomes) and choose randomly from those sub-samples: stratified random sample: an equal number from each strata, ex.
