Enter your search terms Submit search form
	Web pdp-10.trailing-edge.com

SUBROUTINE SIGNT

PURPOSE
   TO PERFORM A NON-PARAMETRIC SIGN TEST, GIVEN TWO SETS OF
   MATCHED OBSERVATIONS.  IT TESTS THE NULL HYPOTHESIS THAT THE
   DIFFERENCES BETWEEN EACH PAIR OF MATCHED OBSERVATIONS HAS A
   MEDIAN EQUAL TO ZERO.

USAGE
   CALL SIGNT (N,A,B,K,M,P,IE)

DESCRIPTION OF PARAMETERS
   N - NUMBER OF OBSERVATIONS IN SETS A AND B
   A - INPUT VECTOR OF LENGTH N CONTAINING DATA FROM THE FIRST
       SAMPLE, A
   B - INPUT VECTOR OF LENGTH N CONTAINING DATA FROM THE SECOND
       SAMPLE, B
   K - OUTPUT VARIABLE CONTAINING THE NUMBER OF PAIRS OF
       OBSERVATIONS FROM THE TWO SAMPLES WHOSE DIFFERENCES ARE
       NON-ZERO
   M - OUTPUT VARIABLE CONTAINING THE NUMBER OF PLUS OR MINUS
       DIFFERENCES, WHICHEVER IS FEWER.
   P - COMPUTED PROBABILITY OF AS FEW AS M NUMBER OF PAIRS
       HAVING THE SAME SIGN, ASSUMING THAT THE SAMPLES CAME
       FROM THE SAME POPULATION.
   IE- 0, IF THERE IS NO ERROR.
       1, IF K IS ZERO.  IN THIS CASE, P IS SET TO 1.0 AND
       M TO 0.

REMARKS
   IF K IS LESS THAN OR EQUAL TO 25, THE PROBABILITY WILL BE
   COMPUTED USING THE BINOMIAL DISTRIBUTION.  IF K IS GREATER
   THAN 25, THE PROBABILITY WILL BE COMPUTED USING THE NORMAL
   APPROXIMATION TO THE BINOMIAL DISTRIBUTION.
   P COMPUTED IS THE PROBABILITY FOR A ONE-TAILED TEST.  THUS,
   FOR A TWO TAILED TEST, DOUBLE THE VALUE FOR P.

SUBROUTINES AND FUNCTION SUBPROGRAMS REQUIRED
   NDTR

METHOD
   REFER TO DIXON AND MASSEY, INTRODUCTION TO STATISTICAL
   ANALYSIS (MCGRAW-HILL, 1957).