Theory of statistical observation

Statistical observation is the initial stage of statistical research and largely determines the success of the entire work. On how complete and high-quality the collected data will be, largely depend on the conclusions to which the researcher will come.

Statistical observation should be organized as systematic, mass and systematic.

The systematic nature of statistical observation lies in the fact that it is prepared and implemented according to a detailed plan developed in advance, which covers the organization and technique of collecting statistical information, controlling its quality and reliability, and the design of final materials.

The mass nature of statistical observation means that it is organized and aimed at covering as many cases of manifestation of this phenomenon or process as possible.

The systematic nature of statistical observation is determined by the fact that it should not be carried out spontaneously, but systematically, if possible at regular intervals.

Statistical observation is a scientifically organized, systematic and systematic process of collecting mass data on various economic and social processes, phenomena and facts.

The main task of statistical observation is to obtain reliable statistical data for the management and planning of the national economy of the Republic of Belarus, for forecasting the economic and social development of the country, region, individual enterprise.

Statistical observation can be carried out through reporting and through specially organized surveys. These are the main organizational forms of statistical observation.

In the first case, statistical information is received from individual enterprises in the prescribed forms within the established time limits. Completion of reports is carried out by each enterprise on the basis of operational-technical or accounting data. Reporting is distinguished by time periods: daily, decadal, monthly, quarterly, semi-annual, annual. All these types of reporting, except for the annual one, are united by one name – the current one.

According to the method of transmission, reports are divided into postal and telegraphic, as well as transmitted through a computer network.

The collection of information through reporting in the Republic of Belarus is carried out by the state statistics bodies. The Ministry of Statistics and Analysis develops and approves reporting forms and their content.

However, there is no need to constantly take into account all phenomena of social life without exception (the cost of fixed production assets, the family budget). In such cases, the collection of information about the phenomenon of interest is carried out through a specially organized statistical observation in the form of censuses or special surveys. The information obtained as a result of specially organized statistical observation complements the reporting data, allows a more detailed study of individual aspects of public life.

Statistical observation differs in the types and methods of conduct. Types of statistical observation can be classified according to a number of characteristics.

Depending on the time factor, observation can be continuous (current) and discontinuous. Current observation is carried out continuously as the fact under study arises (statistical reporting). Discontinuous, in turn, is divided into one-time, if the observation occurs from case to case, as necessary, and periodic, if it is repeated at regular intervals (year, 5 years, 10 years, etc.). Censuses are an example of intermittent observation.

Final indicators for a certain period are possible only on the basis of current observation (information on the number of births per year, on the products released, etc.).

In accordance with the coverage of the units of the observed object, a distinction is made between continuous and non-continuous statistical observation. With continuous observation, the task is to obtain information about all units of the population under study (population census). Continuous observation is resorted to in cases where it is physically impossible to carry out continuous observation (population budget), when observation entails damage or destruction of the observed unit (food quality research, testing the strength of parts), when there is a limitation in time or means, etc.

There are the following types of non-continuous observation: observation of the main array, questionnaire, sample, monographic.

Observation of the main array involves the exclusion of insignificant units from the composition and the study of its main part (production of spare parts for cars – collection of information on large enterprises, excluding private workshops).

Questionnaire observation is organized as follows. The organization, having set itself the task of studying a particular issue, sends (or distributes) to a certain circle of consumers special questionnaires in which questions are contained. Filling out and returning questionnaires is voluntary.

Sample observation is a type of continuous survey in which a certain number of units are selected from the entire population under study by lot or other method, for which all the signs of interest to the researcher are registered and indicators are calculated, which are then extended to the initial set.

A monographic observation is a detailed, thorough study and description of one unit (one worker, one brigade, one enterprise). Monographic observation is usually carried out by scientific organizations.

According to the way the collection of statistical observations is organized, the following methods of data collection are distinguished: expeditionary method, turnout, postal or telegraphic.

In the expeditionary method of observation, registrar persons are sent to the units about which information should be obtained and receive the necessary information on the spot (enumerators carry out population censuses).

But you can oblige each person to appear at the census station and there give the necessary information about himself. This is a safe method (information about births, deaths, etc.).

Postal (telegraphic) data collection means that the necessary information is requested and transmitted by the relevant communication authorities or directly by teletype.

Registration of the necessary information in statistical observation can be made on the basis of different sources. In cases where the registrars obtain the necessary information through direct personal inspection, measurement, weighing, they talk about registration on the basis of direct observation (counting the balances of inventory in the warehouses of enterprises during inventory).

In other cases, the information is recorded on the basis of a survey. So, in the census of the population, all information about each person is recorded from the words of the respondent.

In some cases, information is recorded only on the basis of the data recorded in the document. This method of recording information is called documentary. It is the basis for filling out the reporting (registration of the child in the registry office on the basis of a certificate from the maternity hospital).

Statistical observation is carried out according to a pre-developed plan. In the Republic of Belarus, by Order No. 131 of 20 October 6, 1997, the Minister of Statistics and Analysis approved the procedure for carrying out plans for statistical work (observations). All issues solved in terms of statistical observation can be divided into two groups: programmatic and methodological and organizational.

Program and methodological issues include: determining the purpose of observation, the object of observation, the unit of observation, the development of an observation program.

When planning any statistical observation, first of all, it is necessary to accurately formulate its purpose. Depending on the task of statistical observation, the purpose of observation is formulated. It should be specific, relevant in content and aimed at collecting the necessary data for planning, managing and analyzing the economic activities of the enterprise, as well as for studying the patterns of development of the enterprise. Next, the object of observation is determined. To determine the object of observation means to accurately establish the boundaries of the studied population, to decide what or who should be examined in the process of observation. When determining the object of observation, its boundaries in time, space and material essence must be strictly drawn.

Material boundaries (boundaries by content) define the groups of the population to be registered.

Boundaries in space (boundaries of the place) establish the place where information is collected.

Time boundaries are established depending on the characteristics of the objects under study, as well as on the purpose of the study. If the answers to the questions of the program are received for a certain period of time, then they talk about the registration period (GDP for 1999 amounted to 2880612 thousand rubles). Information on a certain date is a critical moment of registration (the total number of officially registered unemployed in the Republic of Belarus as of 1.09.2003 amounted to 139.4 thousand people).

The object of observation is always a certain set consisting of individual elements, subject to statistical observation.

The unit of observation is the primary element of the object of observation, characterized by a number of features that are registered in the process of observation (the census object is the totality of all inhabitants, the unit of observation is each person).

The reporting unit is the primary cell from which information about the units of observation should be received.

Next, a statistical observation programme is being developed. The monitoring programme is understood as a list of those questions to which answers must be obtained in the process of observation.

The characteristics that the individual units of the population may differ from one another may be of a different nature. They can be quantitative (age, length of service, weight, etc.), and then the individual units differ from each other in the time of this feature, the presence or absence of a particular feature. Signs can be qualitative (gender, marital status, occupation), and then the individual units of observation differ from each other in the presence or absence of a particular quality.

When drawing up the program, special attention should be paid to the wording of the questions. They should be formulated clearly, concisely and without differing interpretations. The more specific the question, the greater the guarantee of its correct understanding and obtaining an accurate answer.

Filling in any statistical document is carried out in accordance with the instructions. The instruction contains written instructions and explanations for the implementation of the monitoring program. The instruction reflects the purpose and objectives of statistical observation, information about the object and unit of observation, the time and timing of the observation, the design of the results and the timing of their submission to the relevant organizations, gives instructions on how to understand a particular question of the program, provides examples of possible answers and the procedure for filling out the form.

Questions of the program and answers to them are recorded in special statistical forms, which can have different names (questionnaire, questionnaire, form, form, list).

Statistical forms are of two types – individual and list. In the first case, the form is started for each unit of observation separately. In the second case, one form is drawn up for several units.

Organizational issues of the plan of statistical observation include such important points as the definition of the subject of observation, place and time, form and method of observation.

The definition of the subject of observation boils down to the answer to the question: who will carry out statistical observation? In some cases, these may be statistical bodies with their own personnel employees. In the case of large-scale observation (population censuses), the general public may be additionally involved in statistical observation. Sometimes the population itself may be involved in the observation.

When planning statistical observation, it is necessary to solve the question of the time of observation (the population census is conducted in winter). Next, the question of the place of observation is decided (in the census – this is the place of residence).

In terms of statistical observation, it should be decided which observation will be applied in terms of coverage (continuous or non-continuous), what time is chosen for observation, how data will be collected, and on the basis of which sources the information will be recorded.

No matter how carefully the plan of statistical observation is thought out, no matter how precisely the instructions of the person collecting information are guided by all the instructions, observation errors may occur with any statistical observation.

D.P. Zhuravsky (1810-1856) for the first time in his work “On the Sources and Use of Statistical Information” raised questions about the reliability and suitability of official statistical data for the analysis of economic activity. He pointed out that until the accuracy of these data was ensured, statistical work would be useless to society.

Depending on the nature, stage and causes of occurrence, several types of observation errors are distinguished (Table 1).

Table 1 Resource requirements by component

Classification of observational errors

Features of classification

Types of errors

Nature of errors



Stages of occurrence

Registration errors;

Errors in preparing data for machine processing;

Errors in the process of machine processing.


Measurement errors;

Errors of representativeness;

Intentional errors;

Unintentional errors.

By their nature, errors are divided into random and systematic. Random errors are called errors, the occurrence of which is due to the action of random factors (the respondent could make a reservation or make a mistake both in the direction of exaggeration and in the direction of understatement).

Systematic errors (errors of poorly configured measuring instruments; errors resulting from an unclear formulation of the observation question) are of great danger.

Depending on the stage of occurrence, registration errors are distinguished; errors in preparing data for machine processing; errors that appear at the stage of machine processing.

Registration errors include those inaccuracies that occur when recording data in a statistical form or when entering data into a PC (accidental misprints or deliberate distortion of the values of individual indicators, non-compliance with the form of the form, errors that occur when typing data on the keyboard).

The occurrence of errors in the preparation of data for processing and in the processing itself is associated with the poor-quality state of primary documents or machine storage media – floppy disks, improper storage of information or equipment malfunctions.

Measurement errors are associated with certain errors that occur with a single statistical observation.

Errors of representativeness are characteristic of non-continuous observation and are associated with the fact that the value of the studied feature in the selected population differs from the value of this feature in the entire studied population, i.e. these errors are associated with incomplete coverage of the units of the studied population.

Intentional errors arise from the deliberate distortion of data.

Unintentional errors are random and can be considered as misses in the work and are the result of low qualifications of employees, negligence and dishonesty in work.

Since errors can always occur in the process of observation, all the collected material must be subjected to control in order to eliminate errors.

The main types of data reliability control are logical and arithmetic.

Logical control consists in comparing the answers to interrelated questions of the observation program.

Counting (arithmetic) control is reduced to checking the general and group digital totals and comparing them. Its task is to detect and correct incorrect results of numerical indicators.

All detected errors should be fixed. To do this, you have to make control surveys, inquiries by mail, by phone, etc.