BIOL 373 Lecture Notes - Lecture 17: Ratio Test, Analysis Of Variance, Dependent And Independent Variables
Document Summary
Can plot data in a distribution or each individual data point, and then bring together into a new way to represent. I. e size isn"t just small or large, but continuous: compare two continuous axes to predict a response. Regression: considers a predictive relationship between two variables. Simple = straight line = simple linear regression. Linear regression for a population: yi=a + bxi+ i. Linear regression for a sample: yi=a + bxi + i. I = error = residual a= intercept, b = slope: departure of yi from value at y predicted from regression line at xi, departure of measured from predicted. Similarities between regression and anova due to common mathematical basis. Uses f ratio:variance due to regression/variance due to error. F ratio: variance due to factor/variance due to error. The closer r2 is to 1, the greater the strength of your regression i. e the predictive regression is close to the actual data.