STAT 205 Lecture Notes - Lecture 1: Dependent And Independent Variables, Simple Linear Regression, Linear Regression
Document Summary
When to reject ho with fcalc vs fcrit. Must have the same n mud > or < do. If p > 0. 05: equal variances pooled. Use dof from sp2 for t table. If p < 0. 05: unequal variances not pooled. Ho: the treatments make no effect on the response variable. Fcalc > fcrit: qq plot, levenes, look at anova p, fcalc, tukeys (tells you which groups are statistically different), calculate hsd, use n-k when looking up q. N means total n, so n1 + n2 etc. If ix1-x2i > hsd, then the 2 groups are statistically different. There is no association between the row and column variable. There is an association between the row and column variable. Estimates the straight line relationship between the mean values of 2 variables. Interpretation: represents the predicted value of y when x=0. Not meaningful if the value x=0 is outside the range of sample data.