PSYC 2040 Lecture Notes - Lecture 3: Null Hypothesis, Ibuprofen, Analysis Of Variance
Document Summary
One independent variable with three levels (ex. whether three pain medicaions (tylenol, advil, bayer) are equally efecive, iv = medicaion) There is a chance you are making a type one error when using a t-test (5% each t-test) Running a single analysis saves us from the 15% chance of making a type one error when using three t-tests. Even if the populaion means are idenical we would not expect the three sample means to be the same. Random sampling will result in diferent sample means. With signiicant tesing we compare two things: variability in sample means to, variability we would expect in sample means due to random sampling (assuming the three populaions means and variances are idenical) F = variance of sample means / expected variance of sample means when populaion means are the same. If the populaions have idenical means the f-value will be close to 1.