Regression and Correlation Analysis-Utility and Implementation. Ned McCarty, controller of Arkansas Distribution Company, is responsible for development and administration of the company’s internal informa¬ tion system as well as the coordination of the com¬ pany’s budget preparation.

At a meeting with Donna Tuma, the vice-president, McCarty proposed that the company employ regression analysis (the least squares method) and correlation analysis as a standard part of its internal information system relating to sales and expenses. He feels that such analyses, including projections, would be signifi¬ cant decision-making aids.

Tuma admitted that she had forgotten the exact mechanics of regression and correlation analysis. However, she made two points.

First, that regression and correlation calculations for weekly or monthly amounts would involve enormous numbers of calculations, because the company’s budget and control system uses weekly amounts for sales and some expenses and monthly amounts for other expenses.

Second, a great deal of caution must be exercised when relying on predictions calculated by regression analysis techniques.

McCarty agreed that a large number of calculations would be required, but feels that this problem might be overcome by computerizing the analysis. The computerized analysis would have to suit the com¬ pany’s budget and control system and cover all sig¬ nificant sales and cost accounts, of which there are about 100.

The company’s microcomputer is not large and operates only with a flexible-diskette data-storage device. No standard computer programs are available for this kind of analysis. Therefore, a program must be specially written, its accompanying data gathered, and the processing problems solved.

To pursue his idea, McCarty decided to obtain sam¬ ple data regarding sales and related selling expenses for the past five years. Using regression analysis, he pre¬ dicts sales of S30,500,000 for the coming year and calcu¬ lates a coefficient of correlation of .4 between sales and the selling expenses.

Required:

(1) Explain both the advantages and the limitations of using regression and correlation analysis according to McCarty.

(2) Identify matters that should be considered before the regression analysis is made, based on the sam¬ ple data collected.