follows: (1) to describe events; (2) to infer future causes, or future events; and (3) to enhance the decision making process.6 Descriptive statistics describe the performance or activity of one group or class of data, without attempting to make general izations about other groups or classes. Inferential statistics permit the findings with respect to one set of relationships to be extended to other relationships, and to generalize findings and conclusions on the basis of statistical inference. Trend analysis is one approach to statistical inference. Decision theory provides managers and administrators with knowledge about events and relationships, which reduce the level of uncertainty in the data upon which decisions for an organiza tion are based. Trend analysis can also be useful for this purpose.
Trend analysis is a form of forecasting. A forecast is an estimate of the future value, or future values, of a variable.7 To be meaningful, forecasts must project values of a variable
6R. I. Levin, Statistics for Management, 4th ed. (Engle wood Cliffs, New Jersey: Prentice Hall, Inc., 1990), 6.
7C. E. Webster, The Executive's Guide to Business and Economic Forecasting (Chicago: Probus Publishing Company, 1986), 3 4. in measurable units of the variable.8 A number of procedures exist, by which forecasts may be developed. Some, such as sub jective opinion forecasts, are not suitable for most sophisti cated applications.9 A somewhat more sophisticated forecasting procedure is the index based forecast. Unfortunately, these forecasts are only as good or as bad as the index that serves as the foundation, and as good or as bad as the degree of cor relation between the actual and projected data.
The best forecasts are those developed through the appli cation of statistical procedures. Among the major statistical forecasting procedures are the following:
1. Trend line projections.
...