C QTES 10
C ..................................................................QTES 20
C QTES 30
C SUBROUTINE QTEST QTES 40
C QTES 50
C PURPOSE QTES 60
C TEST WHETHER THREE OR MORE MATCHED GROUPS OF DICHOTOMOUS QTES 70
C DATA DIFFER SIGNIFICANTLY BY THE COCHRAN Q-TEST QTES 80
C QTES 90
C USAGE QTES 100
C CALL QTEST(A,N,M,Q,NDF) QTES 110
C QTES 120
C DESCRIPTION OF PARAMETERS QTES 130
C A - INPUT MATRIX, N BY M, OF DICHOTOMOUS DATA (0 AND 1) QTES 140
C N - NUMBER OF SETS IN EACH GROUP QTES 150
C M - NUMBER OF GROUPS QTES 160
C Q - COCHRAN Q STATISTIC (OUTPUT) QTES 170
C NDF - NUMBER OF DEGREES OF FREEDOM (OUTPUT) QTES 180
C QTES 190
C REMARKS QTES 200
C M MUST BE THREE OR GREATER QTES 210
C QTES 220
C SUBROUTINES AND FUNCTION SUBPROGRAMS REQUIRED QTES 230
C NONE QTES 240
C QTES 250
C METHOD QTES 260
C DESCRIBED IN S. SIEGEL, 'NONPARAMETRIC STATISTICS FOR THE QTES 270
C BEHAVIORAL SCIENCES', MCGRAW-HILL, NEW YORK, 1956, QTES 280
C CHAPTER 7 QTES 290
C QTES 300
C ..................................................................QTES 310
C QTES 320
SUBROUTINE QTEST(A,N,M,Q,NDF) QTES 330
DIMENSION A(1) QTES 340
C QTES 350
C COMPUTE SUM OF SQUARES OF ROW TOTALS, RSQ, AND GRAND TOTAL OF QTES 360
C ALL ELEMENTS, GD QTES 370
C QTES 380
RSQ=0.0 QTES 390
GD=0.0 QTES 400
DO 20 I=1,N QTES 410
TR=0.0 QTES 420
IJ=I-N QTES 430
DO 10 J=1,M QTES 440
IJ=IJ+N QTES 450
10 TR=TR+A(IJ) QTES 460
GD=GD+TR QTES 470
20 RSQ=RSQ+TR*TR QTES 480
C QTES 490
C COMPUTE SUM OF SQUARES OF COLUMN TOTALS, CSQ QTES 500
C QTES 510
CSQ=0.0 QTES 520
IJ=0 QTES 530
DO 40 J=1,M QTES 540
TC=0.0 QTES 550
DO 30 I=1,N QTES 560
IJ=IJ+1 QTES 570
30 TC=TC+A(IJ) QTES 580
40 CSQ=CSQ+TC*TC QTES 590
C QTES 600
C COMPUTE COCHRAN Q TEST VALUE QTES 610
C QTES 620
FM=M QTES 630
Q=(FM-1.0)*(FM*CSQ-GD*GD)/(FM*GD-RSQ) QTES 640
C QTES 650
C FIND DEGREES OF FREEDOM QTES 660
C QTES 670
NDF=M-1 QTES 680
RETURN QTES 690
END QTES 700