A business school carried out a survey to identify what combinations of the variables: gender, parent’s education, mosaic (scores in mosaic pattern test) and visualization test scores best distinguishes students who take the subject Algebra 2 from those who do not take the subject Algebra 2.
Codification of data is as follows:
gender (0=male, 1=female)
parent’s education (on a discrete scale of 1 to 10, 1 being illiterate and 10 being Ph.D.)
mosaic is the actual score in mosaic pattern test (between 0 to 50)
visualization test is the actual score in visualization test (between 0 to 20)
An extract of the SPSS output for discriminant analysis is given below:
Functions at Group Centroids
algebra 2 in h.s.
Function
1
not taken
-.595
taken
.680
Unstandardized canonical discriminant functions evaluated at group means
Canonical Discriminant Function Coefficients
Function
1
gender
-.439
parent's education
.332
mosaic, pattern test
-.023
visualization test
.171
(Constant)
-1.485
Unstandardized coefficients
The cut-off value for the discriminant function score that best distinguishes students who take the subject Algebra 2 from those who do not take the subject Algebra 2 is:
0
0.0425
0.02125
0.0850
