Statistics on Product Quality

the branch of economic statistics that deals with the product quality level achieved and with variation in this level, analyzes the factors that determine the product quality level, and makes known the reserves available for raising the product quality level to the economic optimum. For such purposes, the discipline of product quality statistics is developing a system of indexes that provide a quantitative description of quality level and quality dynamics with respect to individual kinds of products, as well as a system of composite indexes of quality level and quality dynamics with respect to the production of both individual enterprises and branches of the national economy. The quality level indexes of a given kind of product may be based on characteristics that can be determined in the production process, or producer quality; or it may be based on indexes that can be determined during use, or user quality.

The level of producer quality is determined by comparison of a product’s technical characteristics (parameters) with all-Union state standards, technical specifications, and other documents that set forth minimum requirements with respect to such parameters. In producer quality control, the checking of many parameters simultaneously is typical; thus, a composite qualitative evaluation of level of quality is necessitated. In such cases it is common practice to evaluate quality according to a point scale. One variety of this method is product grading, a procedure that is widely used. Here, a general description of the level of quality of similar goods may be provided by an average grade or, if prices are proportionate to grades, by an average price. A general quality dynamics index for dissimilar goods classified by grade is provided by a grade index, the data base for which is statistical reporting by the enterprises on the grading of industrial production.

The level of user quality is determined in simple fashion if all or most users are interested in the same parameter of quality. The level of quality of extracted ore, for example, can be characterized by metal content, and the level of quality of machines and instruments by reliability, service life, and other indexes. Data on the level of quality of many kinds of goods, and the raw materials used in their manufacture, are found in the technical-production reports of enterprises in several branches of the extractive and manufacturing industries.

The question of user quality is much more complicated if the users of one and the same product are interested in different parameters of quality or if they specify requirements with respect to several mutually independent parameters. The methodology for obtaining a composite evaluation of level of quality for such products is still in the stage of theoretical development and discussion. For computation of a composite index of quality dynamics using many parameters, V. A. Trapeznikov has proposed a “coefficient of quality”; this coefficient is calculated as the product of the relative magnitudes of change in each control parameter.

If the levels of quality of a given kind of product have been established for two comparable time periods (K₁ and K₀), then the quality dynamics index for this kind of product is defined as i_K = K₁/K₀. If cost data are available for each kind of product actually produced in the reporting period (Q_i), a composite quality index for dissimilar products can be calculated according to the formula

Product quality statistics also deal with the separate question of composite evaluation of the technological level of industrial production. With the introduction of state certification of product quality, in which three product categories are delineated (seeQUALITY OF PRODUCTS), statistical practice has been to determine each category’s share in the overall cost of production output.

In several branches of industry, it is possible, in order to characterize product quality, to use the data provided by the enterprises’ statistical reports on the quality of industrial production in terms of claims as to quality.

Several branches of industry have gained positive experience in the use of mathematical statistics for determination of the level of user quality from the manufacturer’s control parameters of producer quality. Here, it is possible not only to predict the level of user quality at the moment the production process is complete but also to improve the requirements set forth in all-Union state standards and technical specifications and to adjust the level of quality for specific purposes, in so doing taking account of the dependences—obtained by regression analysis—between the parameters of producer quality and the level of user quality. The methods of mathematical statistics, especially sampling, are widely used for statistical-warning quality control during the production process. Such methods make quality control cheaper and faster and ensure prevention of production rejects.

A very important, though still unresolved, problem in product quality statistics is the determination of savings or losses within the national economy that result from changes in the level of product quality. The solution to this problem requires calculation of the total economic effect, as well as the additional producer’s outlays and user’s savings, resulting from the use of products of higher quality.