STAT 2050 Study Guide - Midterm Guide: Type I And Type Ii Errors, Pooled Variance, Confidence Interval

• School: University of Guelph
• Department: Statistics
• Course: STAT 2050
• Professor: Zeny Feng

Document Summary

Observational studies do not allow investigators to intervene. Experimental studies allow intervention and to control the subjects to limit bias and confounding factors. Parameters are numbers that summarize data for an entire population ( & 2). Statistics are numbers that summarize data from a sample, i. e. some subset of the entire population (x & s2). A normal distribution is an example of a sampling distribution. Confidence interval (one mean): (x1 x2) for 2 mean ci. Assumption (one sample t test): x1, , xn iid n( , 2) Two sample assumption: two independent samples from two independent populations; x1, , xn iid n( , 2) (for both pops. ) If |tobs| > t*, reject ho (two sided)