Data analysis with ordinal and interval dependent variables: examples from a study of real estate salespeople

This paper re-examines the problems of estimating the parameters of an underlying linear model using survey response data in which the dependent variables are in discrete categories of ascending order (ordinal, as distinct from numerical) or, where they are observed to fall into certain groups on a continuous scale (interval), where the actual values remain unobserved. An ordered probit model is discussed as an appropriate framework for statistical analysis for ordinal dependent variables. Next, a maximum likelihood estimator (MLE) derived from grouped data regression for interval dependent variable is discussed. Using LIMDEP, a packaged statistical program, survey data from an earlier manuscript are analyzed and the findings presented.