STAT151 Lecture Notes - Lecture 1: Contingency Table, Five Ws, Scatter Plot

• School: University of Alberta
• Department: Statistics
• Course: STAT151
• Professor: Rosana Fok

Document Summary

Ch 11-13: census: includes everyone and samples the entire population, problems with census: Inferences: population inference: results can be generalized to an entire population. Should only make populations inferences when we have random sampling. Protects against bias, and maintains that the sample is average within the population. When random sampling doesn"t occur, results should stick strictly to the conclusion: causal (cause and effect): difference in treatments causes differences in responses when comparing results in different treatment groups, random selection. No: random sampling methods, simple random samples (srs) Sampling variability: sample-to-sample differences: stratified random sampling, the population is first divided into homogeneous groups, called strata, then take srs within each stratum before the results are combined. Stratifying can also reduce the variability of our: systematic random samples results. Cluster random sampling: splitting the population into similar groups (or clusters), select one or a few clusters at random and perform a census within each of them.