Practice with Multiple Regression



Practice with Multiple Regression

Here’s the output from modeling average ratings on recommendation letters) from gpa, greq, grev and gender

-----------------------

Does the model “work”?

Yes – p < .001

How well ???

The model accounts for about 34% of the variance among GRE scores

Which predictors contribute to the regression model? Interpret the b of each.

Write out the regression model

Averate’ = (.344 * GPA) + (.119 * Gender) + (.003 * GREQ) + (.002 * GREV) + .730

Undergraduate grades and GREQ contribute (both with p < .05)

• With each 1-point increase in GPA, letter rating is expected to increase by .344

• With each 1-point increase in GREQ, letter rating is expected to increase by about .003

Coefficients

a

.730

.762

.958

.341

.344

.084

.401

4.106

.000

.119

.146

.074

.819

.415

2.996E-03

.001

.304

2.630

.010

-2.42E-03

.001

-.187

-1.690

.094

(Constant)

Undergraduate grade

point average on 1-9

scale

GENDER

Quantitative subscore of

GRE

Verbal subscore of GRE

Model

1

B

Std. Error

Unstandardized

Coefficients

Beta

Standardized

Coefficients

t

Sig.

Dependent Variable: rating derived from letters of recommendaton

a.

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download