Statistics and Research Design
Statistics and Research Design PSYC 6430
Popular in Course
Popular in Psychlogy
verified elite notetaker
This 1 page Class Notes was uploaded by Lane Schuster on Sunday October 11, 2015. The Class Notes belongs to PSYC 6430 at East Carolina University taught by Karl Wuensch in Fall. Since its upload, it has received 42 views. For similar materials see /class/221341/psyc-6430-east-carolina-university in Psychlogy at East Carolina University.
Reviews for Statistics and Research Design
Report this Material
What is Karma?
Karma is the currency of StudySoup.
Date Created: 10/11/15
Multiple Regression with SAS You have already done a trivariate two predictors one criterion multiple regression with SAS back in PSYC 6430 let us now try a multiple regression with four predictors Please refer back to the handout Presenting the Results ofa Multiple Regression Analysis in which I presented results from research that resulted in a model for predicting graduate students Grade Point Average from four predictor variables GPA GREQ GREV MAT and AR The data from this research are in the le quotMultregdatquot on my StatData page and the program is the file quotMRegsas39 on my SAS programs page Download both and run the program Look at the output Remember that the slopes are in raw scores units so it is dangerous to compare slopes ofvariables with different metrics For example when you look at the slope for GREQ 004 you might conclude that it has only a tiny effect but that slope is in GPA points increase per one point increase in GREQ A one point increase in GPA is an enormous increase a one point increase in GREQ is a tiny increase When you want to compare slopes do it with the standardized slopes where the units are standard deviation increase in criterion per one standard deviation increase in predictor When you look at the standardized slopes you see that GREQ has the greatest unique contribution 3 324 ofall predictors Note that the single best predictor is AR but that AR has the smallest unique contribution in the context of the other predictors lfyou look at the intercorrelation matrix you will see why this is so AR is highly correlated with each of the other predictors so when the other predictors are included in the model AR becomes redundant The output statement is used to create a data set called quothatsquot This data set will include all ofthe variables in the model statement plus the one variable we created GPAhat which is GPA scores predicted from our model quotHatquot refers to the caret symbol quot which we use to designate an estimator and which looks like a hat Note that when we correlated GPA with GPAhat we obtained the multiple correlation coef cient R More Lessons on Multiple Regression Return to My SAS Lessons Page Karl L Wuensch Dept of Psychology East Carolina University Greenville NC January 2006 Copyright 2006 Karl L Wuensch All rights reserved MRegSASdoc