-
Compute the following
types of factor scores for the two scales in the engineering data:
-
A weighted proper factor
score, using the standardized scoring coefficients as weights (the
built-in method in SAS)
-
A weighted proper factor
score, using the pattern matrix as weights (refer to the syntax for
the chapter on the book website if you have trouble with this syntax)
-
An unweighted improper
factor score
-
Examine the different
factor scores computed above, the distribution of each, and their
respective correlations. How do the methods compare?
-
Use the code presented
below to split the engineering data into two data sets. Each data
set will contain 197 subjects, of which 175 will be in only that data
set, and 22 will be in both data sets. Compute the weighted proper
factor score, using the standardized scoring coefficients as weights
(the built-in method in SAS). Compare the factor scores for the 22
subjects that are in both groups. How do the results compare?
data data1 data2;
set engdata;
recordN = _N_;
if recordN <= 175 then inGrp='Grp 1';
if 175 < recordN <= 350 then inGrp='Grp 2';
if recordN > 350 then inGrp='Both';
if inGrp in ('Grp 1','Both') then output data1;
if inGrp in ('Grp 2','Both') then output data2;
run;