《Multiple regression - Research Methods (Psychology)》.ppt

* Multiple Regression Overview of the Analysis PSY 4603 Research Methods * Multiple Regression Often, a researcher selects more than one independent variable to be included in a study, assuming that more than one variable affects change in another variable. With the addition of variables, the model becomes a multiple regression model. Its generalized form is Y‘ = a + b1X1 + b2X2+ …+ bkXk+ e where k = (1, 2,. . . , k) and k = number of independent variables. * The objectives of MLR are essentially the same as SLR. The sample data is used to estimate a, b, and the error in the model. Second, we want to know the proportion of variance accounted for in the model. Third, we wish to test the statistical hypotheses concerning the sample estimates. The major difference between the MLR model and the SLR one is the formers determination of the relative contribution of each X in the model in explaining Y. When the actual computations are completed, we refer to values of beta (?) instead of values of b. Beta values (?)are the standardized regression coefficients obtained by standardizing the data before the calculations, or standardizing the b values after the analyses. * The assumptions of the model are essentially the same as those of SLR, with the emphasis on (1) estimating the linearity of the data and (2) the effects of intercorrelation in the independent variables and the residuals. University students often complain that universities reward professors for research but not for teaching, and argue that professors react to this situation by devoting more time and energy to the publication of their findings and less time and energy to classroom activities.? Professors counter that research and teaching go hand in hand; more research makes better teachers. A student organization at Palm Beach Atlantic University decided to investigate the issue. They randomly selected 50 psychology professors (lets pretend) employed by PBA. The students recorded the salar