Decision-making requires having access to various increasingly effective tools serving preparation of such processes. Econometrics provides economists with efficient decision-making instruments.¹ The world economy crisis between 1929 and 1933 has triggered necessity of continual analysis and prediction of the business cycle. In answer to the demand, first econometric tools, the so-called economic barometers, emerged.

Econometrics is a branch of economic studies. As Oskar Lange² wrote, “econometrics is a science concerned with assessment of specific quantitative regularities occurring in the economy, through the use of statistical methods. (…) it combines economic theory with statistics, and seeks to use mathematical-statistical methods for assigning a specific quantitative expression to general diagrammatic regularities determined by economic theories.”

On one side, econometrics creates research tools for economics and management, called the models, on the other, it uses those instruments to study economic phenomena and processes. As such, two branches can be distinguished: the theory of econometrics, which provides such models along with the methods for their construction and operation, and applied econometrics, which uses theoretical models for specific applications in a given economy.

Affiliation of econometrics usage with macroeconomic research falls within the realm of macroeconometrics. In contrast, the use of economic models for a single business entity, called microeconometrics, is a subject matter within a part of applied econometrics.³ Microeconometrics, therefore, deals with creation and customization of the tools dedicated to mathematical statistics, which are applicable to business entities, including companies of all sizes, as well as with construction of empirical models designed to support decision-making processes of those economic entities.

Efficiency of microeconometrics is conditioned by a respectively proper background knowledge of microeconomics and of enterprise theory as well as by knowledge of basic economic entity management on the part of the constructor of empirical microeconometric models. One important prerequisite, on which the design of such models is based, is correct measurement of the phenomena and processes occurring in an enterprise and within its environment. Mass processes, during which statistical regularities are being observed, can be the subject to such modeling.

The purpose of this book is to present econometric tools, which are practical for management of variably sized enterprises. Familiarity with business processes constitutes the basis for such structures. It is necessary to have knowledge about the possibilities of measuring the characteristics, which are expressed directly using numbers on a relativity scale, as well as about the quality characteristics, also called the descriptive ones.

Economists often divide statistical characteristics into quantitative (measurable) and qualitative (immeasurable) ones. In the eyes of measurement theory, all mass processes are measurable; they can, therefore, be reflected by numbers. The numbers, however, have various contents, depending on their affiliation with a corresponding measurement scale, often also called a measurement level.⁴

The magnitudes of measurement scales stem from the sense and meaning of the numbers, which result as a consequence of a given measurement. The following measurement scales can be distinguished⁵:

• nominal
• ordinal (ranked)
• range (interval)
• ratio (quotient).

On the nominal scale, numbers are used for labeling, identification, or classification of disjoint categories. Resultant numbers play the role of symbols, which usually substitute names or verbal descriptions. On this scale, the only acceptable relations between the numbers are (i) equality of elements within distinguished category frames or (ii) variety of disjoint categories. Summing the numbers up is the only acceptable arithmetic procedure. Out of the available statistical techniques, only those based on counting are allowed. Some examples of nominal scale measurements can be, for instance, citizen’s Social Security Numbers, tax identification numbers, postal codes, phone numbers, and so on.

Within the nominal scale, attention is drawn onto a special case – the dichotomous scale, which is commonly used in statistics to extract disjoint category pairs. A simultaneous definition of an A variant of a given phenomenon allows classification of events in a variant form: A or Ā (not A). Assignment of the number 1 to each observation A, while the number 0 is assigned to observation Ā, forms the so-called dummy variable.

On the ordinal scale, numbers are the ranks indicating the order of elements or characteristics of a given phenomenon. Not only the ranks reflect the elements’ irregularities, but also their arrangement in terms of a considered ownership. The categories of the phenomenon being considered, in this case, are disjoint. The numbers on this scale are comparable on a modular basis. However, they are of a relative (not absolute) importance, since the distances between the ranks are not known. What is more, the distances between the adjacent ranks are not equal. As such, comparison of the ranks can be done by finding equality and majority relationships, and thus, minority relationships as well. Determining the distances between the ranks, that is, determining how they differ, is not possible.

The range scale, also called the interval scale, has the ordinal scale’s characteristics, but the distances between the numbers are known. What is more, the distances between each pair of the adjacent figures are equal. Zero natural, as a zero on the interval scale, is contractual in nature. An example of such scales application can be temperature measurement in Celsius degrees; zero, in this case, is the temperature of water’s change of state from liquid to solid and the other way around. Therefore, the numbers on this scale⁶ are the distances from the contractual zero, which prevents the use of the relationship aspect ratio (their division).

The numbers belonging to the ratio scale are the distances from zero. The quotient scale has properties of all weaker scales and of a natural zero point. All arithmetic operations, including multiplication and division, are thus allowed. Using statistical techniques is also possible. For instance, production capacity of natural or valuable units, employment size, wages, demand, prices, and so on, can be examples of a ratio measurement.

Conversion of the numbers from a stronger scale into the numbers belonging to a weaker one is possible; however, it involves partial loss of numerical information. Occasionally, when the numbers on the stronger scale carry excess information and therefore create the so-called information noise, such operations are necessary.

During measurement, possible errors have to be taken into account. Errors can be divided into two categories: random errors, that is, accidental ones and systematic (tendentious) ones. Random errors are an inherent feature of measurement. They result from imperfections of measuring tools as well as from imperfections of the person performing the measurement. Random measurement errors are characterized by a nominal distribution with a zero mathematical expectation.⁷ This means that positive and negative errors, during a long-term measurement, compensate each other.

Systematic errors result in a faulty measurement result, which signifies excess or insufficiency. This type of an error is caused by human interest in falsifying measurement results, usually done by, for example, providing a taxable income lower than the actual one or a company profit write-up, in hope for a higher reward.

It is also necessary to draw attention to two types of measurement: direct and indirect measurements. Direct measurement involves using a suitable measuring device to determine the measurements of features or things. For example, placing a few slices of cheese on a weight scale allows measurement of its weight. Using a graduated vessel filled with liquid allows measurement of its volume. Indirect measurement occurs in at least two stages. In the first, a physical measurement of features or things is conducted; then in the second, an appropriate system of weights is used to determine those measures in other units. For example, in the first stage, we determine the weight of the cheese. In the second, we use a system of weights, in this case the prices. This allows us to determine the value of that cheese in monetary units; in other words, it is transition from natural units onto economic ones.

A significant part of economic measurements is done indirectly, which is more risky than direct measurement. With indirect measurement, there is, at least, some accumulation of random errors. It is much worse when systematic errors accumulate during both stages. The value as well as suitability of the obtained statistical material may then be scant.