The Wald-Wolfowitz test, also known as the Runs test for randomness, is used to test the hypothesis that a series of numbers is random. A run is a set of sequential values that are either all above or below the mean. To simplify computations, the data are first centered about their mean. To carry out the test, the total number of runs is computed along with the number of positive and negative values. A positive run is then a sequence of values greater than zero, and a negative run is a sequence of values less than zero. We can then test if the number of positive and negative runs are distributed equally in time.
The test statistic is asymptotically normally distributed, so this program computes Z, the large sample test statistic, as follows:
Z = (R – E(R)) / sqrt(V(R))
where R is number of runs. The expected value and variance of R are:
E(R) = ( 2nm / (n + m) ) + 1
V(R) = ( 2nm(2nm – n – m )) / ((n + m)2 (n + m – 1))
where n is the number of positive values and m is the number of negative values.
The following statements create an example data set using the random number generator RANNOR. The Wald-Wolfowitz test will be performed on the variable D.
data one; drop i; do i=1 to 75; d=rannor(123); output; end; run;
The MEAN=0 option in the PROC STANDARD step below centers the variable D about its mean.
proc standard data=one out=two mean=0; var d; run;
The following DATA step computes the total number of runs (RUNS), the number of positive values (NUMPOS), and the number of negative values (NUMNEG).
data runcount; set two nobs=nobs; if d=0 then delete; if d>0 then n+1; if d<0 then m+1; retain runs 0 numpos 0 numneg 0; previous=lag(d); if _n_=1 then do; runs=1; prevpos=.; currpos=.; prevneg=.; currneg=.; end; else do; prevpos=( previous > 0 ); currpos=( d > 0 ); prevneg=( previous < 0 ); currneg=( d < 0 ); if _n_=2 and (currpos and prevpos) then numpos+1; else if _n_=2 and (currpos and prevneg) then numneg+1; else if _n_=2 and (currneg and prevpos) then numpos+1; else if _n_=2 and (currneg and prevneg) then numneg+1; if currpos and prevneg then do; runs+1; numpos+1; end; if currneg and prevpos then do; runs+1; numneg+1; end; end; run; data runcount; set runcount end=last; if last; run;
Finally, these steps compute and display the Wald-Wolfowitz (or Runs) test statistic and its p-value.
data waldwolf; label z='Wald-Wolfowitz Z' pvalue='Pr > |Z|'; set runcount; mu = ( (2*n*m) / (n + m) ) + 1; sigmasq = ( (2*n*m) * (2*n*m-(n+m)) ) / ( ((n+m)**2) * (n+m-1) ); sigma=sqrt(sigmasq); drop sigmasq; if N GE 50 then Z = (Runs - mu) / sigma; else if Runs-mu LT 0 then Z = (Runs-mu+0.5)/sigma; else Z = (Runs-mu-0.5)/sigma; pvalue=2*(1-probnorm(abs(Z))); run; title 'Wald-Wolfowitz Test for Randomness'; title2 'H0: The data are random'; proc print data=waldwolf label noobs; var z pvalue; format pvalue pvalue.; run;
As you would expect of values from a random number generator, the test fails to reject the null hypothesis that the data are random (p=0.9988).
____
Mendenhall, Scheaffer, and Wackerly (1986), Mathematical Statistics with Applications, 3rd Ed., Duxbury Press, CA.
Wald, A. and Wolfowitz, J. (1940), "On a test whether two samples are from the same population," Ann. Math Statist. 11, 147-162.
These sample files and code examples are provided by SAS Institute Inc. "as is" without warranty of any kind, either express or implied, including but not limited to the implied warranties of merchantability and fitness for a particular purpose. Recipients acknowledge and agree that SAS Institute shall not be liable for any damages whatsoever arising out of their use of this material. In addition, SAS Institute will provide no support for the materials contained herein.
These sample files and code examples are provided by SAS Institute Inc. "as is" without warranty of any kind, either express or implied, including but not limited to the implied warranties of merchantability and fitness for a particular purpose. Recipients acknowledge and agree that SAS Institute shall not be liable for any damages whatsoever arising out of their use of this material. In addition, SAS Institute will provide no support for the materials contained herein.
Type: | Sample |
Topic: | Analytics ==> Distribution Analysis Analytics ==> Exploratory Data Analysis Analytics ==> Nonparametric Analysis |
Date Modified: | 2014-02-07 16:14:16 |
Date Created: | 2008-08-27 15:46:20 |
Product Family | Product | Host | SAS Release | |
Starting | Ending | |||
SAS System | Base SAS | z/OS | ||
OpenVMS VAX | ||||
Microsoft® Windows® for 64-Bit Itanium-based Systems | ||||
Microsoft Windows Server 2003 Datacenter 64-bit Edition | ||||
Microsoft Windows Server 2003 Enterprise 64-bit Edition | ||||
Microsoft Windows XP 64-bit Edition | ||||
Microsoft® Windows® for x64 | ||||
OS/2 | ||||
Microsoft Windows 95/98 | ||||
Microsoft Windows 2000 Advanced Server | ||||
Microsoft Windows 2000 Datacenter Server | ||||
Microsoft Windows 2000 Server | ||||
Microsoft Windows 2000 Professional | ||||
Microsoft Windows NT Workstation | ||||
Microsoft Windows Server 2003 Datacenter Edition | ||||
Microsoft Windows Server 2003 Enterprise Edition | ||||
Microsoft Windows Server 2003 Standard Edition | ||||
Microsoft Windows XP Professional | ||||
Windows Millennium Edition (Me) | ||||
Windows Vista | ||||
64-bit Enabled AIX | ||||
64-bit Enabled HP-UX | ||||
64-bit Enabled Solaris | ||||
ABI+ for Intel Architecture | ||||
AIX | ||||
HP-UX | ||||
HP-UX IPF | ||||
IRIX | ||||
Linux | ||||
Linux for x64 | ||||
Linux on Itanium | ||||
OpenVMS Alpha | ||||
OpenVMS on HP Integrity | ||||
Solaris | ||||
Solaris for x64 | ||||
Tru64 UNIX |