ppt课件-g89.2229multipleregressioninpsychology.pptVIP

G89.2229 Lect 6W G89.2229 Lect 1W G89.2229 Multiple Regression Week 6 (Wednesday) Polynomial example Orthogonal polynomials Statistical power for regression Constructing polynomial fits Two approaches for constructing polynomial fits Simply create squared, cubed versions of X Center first: Create squared, cubed versions of (X-C) Xc=(X-`X) Xc and Xc2 will have little or no correlation Both approach yield identical fits Centered polynomials are easier to interpret. Example from Cohen Interest in minor subject as a function of credits in minor Interpreting polynomial regression Suppose we have the model Y=b0+b1X1+b2X2+e b1 is interpreted as the effect of X1 when X2 is adjusted Suppose X1=W, X2=W2 What does it mean to hold constant X2 in this context? When the zero point is interpretable Linear term is slope at point 0 Quadratic is acceleration at point 0 Cubic is change in acceleration at point 0 Orthogonal Polynomials In experiments, one might have three or four levels of treatment with equal spacing. 0, 1, 2 0, 1, 2, 3 These levels can be used with polynomial models to fit Linear, quadratic or cubic trends We would simply construct squared and cubic forms. Making polynomials orthogonal The linear, quadratic and cubic trends are all going up in the same way. The curve for the quadratic is like the one for cubic. Orthogonal polynomials eliminate this redundancy hierarchically. The constant is removed from the linear trend The const and linear are removed from quadratic The const, lin and quad are removed from the cubic. Analysis with Orthogonal polynomials If we substitute orthogonal polynomials for the usual linear, squared, and cubic terms, we Recover the same polynomial fit Obtain effects that are useful in determining the polynomial order Even when cubic effects are included, with orthogonal effects The linear is the average effect The quadratic is adjusted for the linear, but not adjusted for cubic The cubic is adjusted for all before The regression coefficient