STAT 302 - 502 - Wk 13 - 11/18 explanatory = categorical response = quantitative -we can use one-way ANOVA -one way ANOVA: -analysis of variance -compare several means -do these pops have the same mean? -compare how far apart the sample means are -use variation within samples -ie: we find that 3 sample means are different -does this mean that the population means for these 3 groups are different too? -in test statistic for F: 1. variation among sample means 2. variation among individuals in same sample assumptions: -pop dists of response variable for k groups are NORMAL w/ same SD for each group -groups are independent -ANOVA is robust to deviations from Normality for large-enough samples -given that there are no extreme outliers -given that the largest group's standard deviation is no more than 2 times the smallest group's standard deviation denominator << numerator evidence against null F is larger numerator >> denominator evidence against null F is larger -if the sample sizes are large enough, there's stronger evidence against the null hypothesis -we calculate F-value to find p-value to make
View Full Document