Abstract
Raouf and Sathé [1] give a computer program for evaluating whether or not a sequence of observations can be considered random, using the number of runs and the sample median as a reference point. In this paper, the basic runs test is reviewed and then it is shown to be useful in a wide variety of situations. The number of runs can be used
to detect patterns in production,
to examine regression models for underfitting or misfitting,
to look for serial correlation in time series or trend analysis,
to evaluate discriminant analysis functions with qualitative or other non-normal preditor variables, and for many other purposes.
Additional information
Notes on contributors
Gary M. Mullet
Dr. Mullet, a member of the American Statistical Association, is Associate Professor of Statistics, College of Industrial Management, Georgia Institute of Technology.