Statistics is a science that, using probability calculations, studies collective phenomena and processes (social, natural, etc.) in quantitative terms, in order to study and describe these phenomena and processes, as well as to discover their regularities. manifestations. Having become acquainted with the main categories and methods of statistics as a science, you can understand how it works.
Statistics helps to get the necessary information from the available amount of data, which can be quite large. The information can be used to understand available data (descriptive statistics) or to discover new information about events and their relationships (logical statistics).
Types of statistics share the basic concepts and categories of statistics. It is almost impossible to briefly describe them, since this discipline operates with huge amounts of data, methods and principles for processing them.
The process of getting information from data is called statistical inference about some statistical parameters or even whole probability distributions. This is the more general view taken by non-parametric theory in statistics.
In classical applied statistics, the subject of statistics and the main categories of statistics (summarized below) are more clearly defined, so it is preferable to build a statistical model from which conclusions can be drawn; in most cases, this model has not been tested, which can lead to erroneous conclusions.
Statistics as a science is widely used by other branches of science, such as physics, biology, psychology, economics, sociology and others.
To learn as much as possible about phenomena and processes, including economic and social ones, as well as to determine patterns and relationships using statistical methods and techniques, you must first characterize and describe the current situation as accurately as possible. This is done by collecting data that reflects the actual state of affairs, that is, through statistical observation.
The necessary data can be obtained from various sources:
- from specially organized statistical studies are primary data (for example, census or survey results);
- from an existing information system (for example, from the current records of economic agents, banks and various databases of some central andlocal government) - such data is called secondary.
For example, in September 2008, the General Directorate of Regional Statistics in Bucharest published a statistical bulletin for the second quarter of 2008: natural migration of the urban population, incomes of workers, the number of unemployed, the main industrial goods produced in Bucharest, the dynamics of trade turnover by enterprises from main activity in the field of retail and market services and much more. All of this published information is a secondary source of statistical data.
Statistical observation: content, necessity, tasks
Statistical data - one of the basic concepts and category of statistics - is necessary in order to start the process of proper statistical research to assess the effectiveness of social and economic processes, to formulate alternative ways of making decisions, etc. Thus, data can be considered as necessary information to form judgments about decisions in specific situations.
For statistical analysis to be useful for decision making, the input data must be correct and fit the purpose. Therefore, determining the data needed and how it is collected is extremely important. That is why it is important to know the subject of statistics, the main categories of statistics and its methods
If there are errors in the data, if they are ambiguous and misleading, even the most sophisticated and sophisticated processing methods will be inefficient and notbe able to compensate for shortcomings; it is obvious that the results in this case will not be correct and useful.
Any process of statistical research begins with statistical observation. This is part of the main categories of statistics as a science. The way it is organized and conducted additionally affects other stages of the statistical research process, since obtaining reliable, real, accurate data determines the quality of the stages of processing, analysis and interpretation of the results.
Statistical observation is the first stage of statistical research, which consists in the systematic and unified collection and registration of statistical data on all features included in the research program.
Observation must meet quantitative and qualitative requirements:
- fulfillment of quantitative conditions (required amount of data) means receiving at a predetermined time the entire amount of data necessary to achieve all the goals of the statistical study;
- Fulfillment of quality conditions guarantees the authenticity of the collected data, so that the results obtained from the processing of this data are as accurate as possible and lead to the best decisions.
If a statistical study is required to obtain statistical data, then it should be carried out in accordance with a specific plan to ensure that the most accurate and relevant results are obtained with the minimum material and financial resources.
Statistical observation does not happen randomly, because data collection requires some effort, especially if the goal is to obtain reliable data. Typically, statistical observation relies on a predetermined plan (or program) developed in accordance with the purpose of the statistical study.
A complete observation plan, which is based on the subject and categories of statistics, usually includes the following elements:
- observation target;
- observed object: a collection of objects with similar characteristics that are being monitored, for example, agricultural firms, machine-building enterprises, city residents, students or schoolchildren;
- unit of the object of observation: a constituent element of the object of observation, the carrier of the main characteristics of the object, that is, a separate company, enterprise, family, etc.;
- time and place of observation;
- determining characteristics to be monitored;
- data entry form;
- organizational arrangements and instructions.
Surveillance plan in the narrow sense contains only a list of all characteristics to be recorded, required indicators, etc.
Here it is worth considering separately the components of the statistical observation plan, since they represent the main categories of statistics.
Target of surveillance
The purpose of observation is subordinate to the general purpose of statistical research and continues to influence all otherselements of the monitoring program. Setting goals to be achieved through research is the starting point for developing a plan.
At this stage, the definition, the establishment of all statistical units that will be included in the study.
This is done using known nomenclatures, existing classifications, or previous research in the area (if any). The coverage of the observed group depends on the observation method to be used:
- if this is a general observation method, then the object of observation will consist of all units of the group;
- if the method of partial observation is used, data will be collected only on a part of the collective units of the group, which will constitute the object of observation.
However, in both cases, in order to correctly determine the object of observation, it is necessary to determine the space, time and organizational coordinates of the group of interest.
Unit of observation
This is a single unit to which observable characteristics refer, whether simple (whether employee, facility, citizen, etc.) or complex (when considering a team, family, or organization).
Sometimes the unit of observation is not the reporting unit. For example, in the case of a survey conducted within an economic agent, the reporting unit would be the economic agent, and the reporting unit could be an employee, department, team, or product.
Time and placeobservations
Setting the observation time requires attention to two aspects:
- of the time to which the recorded data refers (this can be one "critical moment" or period of time; in the first case, the phenomenon is recorded statically, in the second case dynamically).
- data logging time - usually a well-defined time interval; it is desirable that it be as short as possible so that data recording is carried out as quickly as possible.
For example, during the population and housing census in March 2002, the critical moment was 00:00 on March 18, and the period when the data was recorded was March 18-27. The unit of observation was the household (complex unit).
A place of observation is, as a rule, a place where a phenomenon is registered, where it is observed and studied.
List of observed characteristics
Determining what characteristics are being monitored forms the monitoring program in the narrow sense. It requires the establishment of all variables to be recorded in order to ensure that the phenomenon of interest is studied in all aspects provided for by the objectives of the study, it is desirable to avoid redundancy of information.
Observed characteristics can appear in various forms:
- in the form of indicators in statistical reports compiled by economic agents;
- in the form of answers to questions in the questionnaire in the case of specially organized statistical studies.
Correct drafting of the form
In order to obtain all the data needed for the study and successfully fulfill the purpose of observation, the design of questionnaires must be done in such a way as to ensure a logical structure and well-formulated questions.
The questionnaire should be structured in such a way as to follow a logical sequence of transitions from one question to another, from one topic to another. If the logical structure is not followed, the respondent may become disoriented, which in turn will affect the answers.
The American Institute of Public Opinion Gallup, which is engaged in statistical research around the world, believes that when formulating a question in a questionnaire, it is necessary to follow several basic rules:
- Make sure the respondent knows the topic being researched. Example: “Are you aware of plans to build a business center in District X?”
- Find out the general attitude of the respondent on the issue under consideration. Example: "Do you think this business center is needed in this area?" (Yes/No/Difficult to answer).
- Find out answers to questions on specific issues related to the main question. Example: “Do you think the new business center will affect the neighborhood?” (Yes/No).
- Find out your own points of view. Example: “If you are against the construction of a business center, the main reason will be: a) there are a lot of buildings in this area; b) the construction will violate the integrity of the landscape; c) the project would mean the destruction of a park or playground forchildren; d) another reason.”
In the questionnaire, you can find open questions (where the respondent can give any answer, for example: “What is your profession?”) Or closed questions (where the respondent is given several possible answers from which he can choose one or more). Questions can also be factual (for example: "What size shoes do you wear?") or subjective, aimed at the respondent's opinion (for example: "What do you think about the government's intention to raise VAT?").
Methods of statistical observation
The main statistical categories in statistics certainly include various methods of statistical observation and research.
There is a wide range of observation methods, used together or separately, to obtain real, complete and objective statistical information. Observation methods are one of the basic concepts and categories of statistics. They are classified according to the following criteria:
1. According to the terms of registration of data on various phenomena and observation processes:
- Current observations, when phenomena and processes are continuously monitored and data characterizing them are constantly recorded, for example, in "statistics of civil status" such demographic phenomena as births, deaths, marriages, divorces are continuously recorded.
- Periodical observations, when data on phenomena or economic and social processes are recorded at regular intervals, such as population censuses, agricultural censuses.
- One-time observations whendata relating to phenomena or processes are recorded intermittently, sometimes for a specific purpose, which is why they are also called “organized” (for example, a public opinion poll about the attitude of citizens towards a new normative act that has entered into force).
2. According to the degree of coverage of units of the object of observation:
- Continuous observations, when all units of the statistical population are subject to observation. For example, population and housing censuses.
- Partial observations, when only part of the units of the statistical population is to be observed.
3. According to the way the data is collected, the observations are:
- Primary (direct) when data is obtained by collection, recording directly from statistical units (eg census, opinion poll).
- Secondary (indirect), when data is taken from existing documents (for example, from accounting records).
The census as a statistical observation is a continuous periodical observation method. It has been carried out since ancient times. Even the Romans and Egyptians practiced similar studies.
From a simple number of inhabitants, the census has expanded into other areas of the social and economic life of the population. For example, censuses of housing, livestock, industry, agriculture, trade, etc. began to be carried out.
This is a very labor-intensive type of research, requiring high costs, huge staff, detailed organizational arrangements and complex processing of recorded data.information.
The population census is one of the main sources of data in demographic statistics and provides information about the size and structure of a country's population at a given point in time. It is initiated by the state and regulated by regulations, but the responsibility for its organization and implementation lies with the National Institute of Statistics, which also processes the collected data.
Basic principles of statistical research
The census is based on respect for the principles of universality, simultaneity and comparability.
Compliance with the principle of universality requires the registration of all persons falling under the main categories of the population in statistics and under the jurisdiction of the state. Thus, in the course of the population and housing census in March 2002, all citizens residing in the country were registered, regardless of whether they were in the country or temporarily abroad, as well as persons of other nationalities or stateless persons who were temporarily reside in the territory of the state.
According to the principle of simultaneity, the collected information reflects the existing situation simultaneously for all respondents (critical moment), although data recording takes some time. The critical moment is usually chosen in winter, when the observed population is in a state of stability, uniformity, not subject to fluctuations and random movements (vacations or holidays are excluded in this sense).
Based on the census, wewe obtain the indicators of volume and static structure associated with the critical moment. Nevertheless, population studies are surprising in identifying changes in volume and structure from one census to another, which requires the principle of comparability of data both in time and in territorial terms.
Statistical survey is a method of partial specially organized observation, one of the most frequently used methods of modern statistical observation, since the subject, method and main categories of statistics are observed, while saving material, financial and human resources. The method often replaces the general large-scale observation, which is much more difficult to organize and conduct.
In the case of a statistical survey, the sample selects an object consisting of a certain part of the total units. The pattern may or may not be chosen randomly.
Random selection of units assumes that each statistical unit has an uncomputable chance of being included in the sample, that there is no unit that cannot be subjected to the selection process, and that there is no unit that can be predominantly selected. Randomly selected samples are representative of the entire population, meaning they have the same basic characteristics as the general population.
Non-random selection refers to any method of selecting statistical units in a sample on a preferential basis.
After samplingis formed, all the elements included in the observation program are processed, and statistical indicators are obtained at the sample level.
Questioning is one of the most frequent and simple methods and main categories of statistics. It is a method of partial observation with a more random character. It is similar to, but different from, a statistical survey:
- non-fulfillment of the sample representativeness condition.
- Higher non-response rates reported directly, by phone, by e-mail, or by post, as questionnaires are voluntary.
Other methods and main categories of statistics
Observation of the main array. It is a method of partial, specially organized observation, which includes the formation of a non-representative sample by selecting the most significant units from a group.
Monographic observation. A method of partial observation, which is a detailed, in-depth characterization and description of one or more statistical units (a monograph of an enterprise, county, locality, etc. can be compiled).