Google
 

Trailing-Edge - PDP-10 Archives - decus_20tap2_198111 - decus/20-0026/biser.doc
There are 2 other files named biser.doc in the archive. Click here to see a list.
SUBROUTINE BISER

PURPOSE
   TO COMPUTE THE BISERIAL CORRELATION COEFFICIENT BETWEEN TWO
   CONTINUOUS VARIABLES WHEN ONE OF THEM HAS BEEN ARTIFICIALLY
   DICHOTOMIZED.

USAGE
   CALL BISER (N,A,B,HI,ANS,IER)

DESCRIPTION OF PARAMETERS
   N   - NUMBER OF OBSERVATIONS
   A   - INPUT VECTOR OF LENGTH N CONTAINING THE CONTINUOUS
	 VARIABLE
   B   - INPUT VECTOR OF LENGTH N CONTAINING THE DICHOTOMIZED
	 VARIABLE
   HI  - INPUT - NUMERICAL CODE TO INDICATE THE HIGHER CATEGORY
	 OF THE DICHOTOMIZED VARIABLE.	ANY VALUE IN VECTOR B
	 EQUAL TO OR GREATER THAN HI WILL BE CLASSIFIED INTO
	 THE HIGHER CATEGORY.
   ANS - OUTPUT VECTOR OF LENGTH 8 CONTAINING THE FOLLOWING
	 ANS(1) - MEAN OF VARIABLE A
	 ANS(2) - STANDARD DEVIATION OF VARIABLE A
	 ANS(3) - PROPORTION OF THE CASES IN THE HIGHER
		  CATEGORY OF VARIABLE B
	 ANS(4) - PROPORTION OF THE CASES IN THE LOWER
		  CATEGORY OF VARIABLE B
	 ANS(5) - MEAN OF VARIABLE A FOR THOSE CASES FALLING
		  INTO THE HIGHER CATEGORY OF VARIABLE B
	 ANS(6) - MEAN OF VARIABLE A FOR THOSE CASES FALLING
		  INTO THE LOWER CATEGORY OF VARIABLE B
	 ANS(7) - BISERIAL CORRELATION COEFFICIENT
	 ANS(8) - STANDARD ERROR OF BISERIAL CORRELATION
		  COEFFICIENT
   IER -  1, IF NO CASES ARE IN THE LOWER CATEGORY OF VARIABLE
	 B.
	 -1, IF ALL CASES ARE IN THE LOWER CATEGORY OF
	 VARIABLE B.
	 0, OTHERWISE.
	 IF IER IS NON-ZERO, ANS(I)=10**75,I=5,...,8.

REMARKS
   THE VALUES OF THE DICHOTOMIZED VARIABLE, B, MUST BE IN
   NUMERIC FORM.  THEY CANNOR BE SPECIFIED BY MEANS OF
   ALPHABETIC OR SPECIAL CHARACTERS.

SUBROUTINES AND FUNCTION SUBPROGRAMS REQUIRED
   NDTRI

METHOD
   REFER TO P. HORST, 'PSYCHOLOGICAL MEASUREMENT AND
   PREDICTION', P.95-96 (WADSWORTH, 1966).