STAT 20 Lecture Notes - Lecture 20: Summary Statistics, Observational Error

• School: Berkeley
• Department: Statistics
• Course: STAT 20
• Professor: Fletcher Ibser

Document Summary

First lets jump into an overview of concepts in chapters 17,18, 20. Chapter 19, 21 has similar concepts, but different. Statistic = parameter + chance error + bias. Box model of known data for unknown draws. Box model of unknown values with a known number of draws. In other words earlier we knew what the mean, median, and range of values before. Now we are trying to find these values. Sample : subgroup of the population that we know something about. For example the tickets randomly drawn from box. This is usually unknown and what we try to estimate using our statistic. Statistic : some math fact about a sample (usually known). Bias: is a non-random, systematic error, any difference between statistic and parameter that is not random. Selection bias: bias due to differences between those selected and those not selected. Examples of this: a sample of landlines will have older people in their sample.