POLI 210 Lecture Notes - Lecture 14: Type I And Type Ii Errors, Statistical Hypothesis Testing, Null Hypothesis
Document Summary
Hypotheses are really statements about population parameters. Making assertion about population but can only use samples to see if its true. For ex: canadian woman vote more for liberals than men do. Think of the hypothesis you are going to test. Set the hypothesis against null hypothesis this states that there is no relationship or your hypothesis is wrong. Calculate the probability of the data that could have resulted randomly. Which level of alpha is best (when do we reject ho?) Null hypothesis: ho for ex: a pollster wants to test whether hispanics support clinton more than non- Ho true: accept ho, correct decision, type ii error. Ho false: reject ho, type 1 error correct decision. For ex: researcher wants to compare two life-saving medicine. Null hypothesis: the rates of people healed of the two groups are the same. Alternative hypothesis: the rates people healed of the two groups are different.