Skip to main content

Table 4 Discriminant Function Analysis Results.

From: Reliability, validity and psychometric properties of the Greek translation of the zung depression rating scale

A

Classification Functions

Classification Matrix

SDS item No

normal

depressed

D-C

Rows: Observed classifications

 

comparison

      
 

subjects

      

Constant

-47.84

-63.36

-15.52

Columns: Predicted classifications

Item No 1

4.50

7.15

2.65

 

%

normal

depressed

     

Correct

comparison

 
      

subjects

 

Item No 2

2.32

1.87

-0.45

Comp subjects

100.00

120

0

Item No 3

4.05

6.08

2.03

depressed

90.00

4

36

Item No 4

3.10

4.24

1.14

Total

97.50

124

36

Item No 6

3.00

2.03

-0.97

    

Item No 7

3.19

4.74

1.56

    

Item No 8

3.32

4.14

0.82

    

Item No 9

1.35

2.53

1.18

    

Item No 10

2.04

3.17

1.13

    

Item No 11

4.46

5.48

1.02

    

Item No 12

3.12

0.75

-2.37

    

Item No 15

4.59

3.12

-1.46

    

Item No 16

2.06

2.98

0.91

    

Item No 17

0.02

0.99

0.97

    

B

Classification Functions

Classification Matrix

Factor No

normal

depressed

D-C

Rows: Observed classifications

 

comparison

      
 

subjects

      

Constant

-0.42

-2.63

-2.21

Columns: Predicted classifications

Factor 1

-0.43

1.30

1.72

 

%

normal

depressed

     

Correct

comparison

 
      

subjects

 

Factor 2

0.12

-0.37

-0.49

Comp subj

95.00

114

6

Factor 3

0.32

-0.98

-1.30

depressed

72.50

11

29

Factor 4

0.21

-0.62

-0.83

Total

89.38

125

35

Factor 5

-0.38

1.15

1.53

    
  1. Analysis with individual ZDRS items entering the procedure. When the equation: 2.65*(itl) - 0.45*(it2) + 2.03*(it3) + 1.14*(it4) - 0.97*(it6) + 1.56*(it7) + 0.82*(it8) + 1.18*(it9) + 1.13*(it10) + 1.02*(itll) - 2.37*(it12) - 1.46*(it15) + 0.91*(itl6) + 0.97** itl7) > 15.52 is true, then the subject is a depressed patient. This method correctly classified all normal comparison subjects and 90% of patients. B: Analysis with factor scores entering the procedure. When the equation: 1.72*(f1) - 0.49*(f2) - 1.3*(f3) - 0.83*(f4) + 1.53*(f5) > 2.21 is true, then the subject is a depressed patient. This method correctly classified 95% of normal comparison subjects and 72.5% of patients