PSYC 305 Lecture Notes - Lecture 18: Coefficient Of Determination, Null Hypothesis, Linear Regression
Document Summary
Note these assumptions" similarities with those of the anova test: basically those of the anova test, plus an assumption for linearity. Y minus y-hat, again, is the residual value, a. k. a. error. Y-hat: it is our prediction of the actual value based on the. Y-hat is the line of best fit, y-hat = + . A residual plot should have no pattern at all, like on the right- hand side. Random scatters => all 3 assumptions are now satisfied. If you have some pattern => violation of at least 1 assumption! Like the left: there"s some curvature, non-linear relation b/w x and y. Again, on left there"s a pattern => violation of homogeneity of variance. Multiple linear regression, with 2 ivs and 1 dv. Y-hat here, is our prediction of the data points using the ivs. Slope of one iv is the effect of that iv on the dv, holding the other ivs constant!