BUS 336 Chapter Notes - Chapter 9: Type I And Type Ii Errors, Null Hypothesis, Statistical Inference
21 Nov 2018
School
Department
Course
Professor
Document Summary
Statistical inference = any procedure for extracting information about a probability distribution from an observed sample. 3 types of statistical inference: point estimation = make reasonable guess of unknown value. Plug-in principle = estimate unknown value by computing mean of empirical distribution: hypothesis testing = guess which of 2 possible statements is correct, set estimation. Significance probability = probability that chance will produce coincidence at least as extraordinary as the phenomenon observed. Neyman-pearson formulation of hypothesis testing states the null hypothesis is a privileged status. Ho maintained unless there is compelling evidence against it. Type i errors much worse than type ii. Neyman-pearson imposes upper bound on maximal probability of type i error tolerated. Significance level specific prior to examining data (alpha = 0. 05 most common) Only decision rules which prob. of type i error is not greater than alpha considered. Descriptive stats = toolbox to describe the features/charac of a data set.
