A special sample survey of the household economy is being conducted. Problems of using the results of sample surveys of households to model the structure of their expenditures

Another additional, but extremely important source of data on population and demographic processes are special sample surveys. Their role is difficult to overestimate. And the main thing is not even savings compared to population censuses, which is usually discussed in this regard. The main point is that special sample survey programs tend to be much more detailed and in-depth than population census programs. They allow you to deeply and comprehensively explore many issues of demographic change and the factors that cause them, to obtain information that is simply impossible to obtain in other ways. That is why such surveys, based on a sampling method, are widely used in the world, incl. and in our country.

It should be borne in mind that under the word "survey" Demographers usually mean completely different procedures: on the one hand, we are talking about traditional statistical operations for collecting information on a particular issue or problem, in which the sampling method may or may not be applied. On the other hand, we are often talking about sociological research proper, in which the use of the sampling method and the survey method is mandatory. The word itself "survey" in the latter case, it is rather a tribute to the statistical tradition than a term that reflects the essence of the matter.

The differences between statistical surveys and sociological research are of a fundamental nature, since in the first case we are talking about collecting information about facts (events, characteristics, etc.) and only in the second place about opinions (the so-called "statistics of opinions"). Sample parts of the 1989 population census and the 1985 and 1994 microcensuses can serve as an example of such surveys, in which, along with information on the number of children born, opinions on the expected and desired number of children were also recorded.

In the second case, we mean a deep comprehensive study, which is complex in nature and aimed at identifying not only the external side of the phenomenon under study, but also its driving forces, both external (factors) and internal (needs, attitudes and motives for the corresponding behavior). This fundamental difference should not be forgotten when speaking of "examinations".

In modern demography, both surveys in the traditional statistical sense and sociological studies are used and complement each other. At the same time, surveys were historically the first, which, apparently, is the reason that the so-called. "traditional" demographers prefer to use this term.

Demographic surveys have a rather long history. The first attempts to conduct them date back to the 20-30s. last century. One of the first surveys of opinions about the size of the family was conducted by demographers of Kharkov University under the leadership of S.A. Tomilina (1877-1952) back in 1927 50

Tomilin Sergey Arkadyevich (1877-1952), Ukrainian scientist, specialist in the field of social hygiene, sanitary statistics, history of medicine, demographer, professor (1926), doctor of medical sciences (1936). He graduated from the medical faculty of Moscow University (1901). Taught at medical schools educational institutions Kharkov and Kyiv. ...Under the leadership of Tomilin, the staff of Kharkiv University conducted one of the world's first studies of opinions on family size (1927), which is close in meaning to identifying the desired number of children... Population. Encyclopedic Dictionary. M., 1994. S. 531.

In the United States, the Gallup Institute of Public Opinion began in 1936 regular surveys of the ideal family size from a nationwide sample. Data from these surveys over several decades has made it possible to estimate the extent to which Americans' ideas about the ideal number of children in a family have decreased 51 .

Ideal number of children -

an individual's idea of ​​the best number of children in general, without taking into account the specific life situation and personal preferences. Received in surveys as an answer to the question: “How many children is it best to have in a family?” Desired number of children - the number of children that an individual would prefer to have in his family based on his own inclinations. They are received in surveys as an answer to the question: “How many children would you like to have in a family under all the necessary conditions for this?”

Expected number of children - the number of children that an individual intends to have in his family, given the specific life situation and personal preferences. Received in surveys as an answer to the question: "How many children are you going to have in your family?" or “How many more children are you going to have in the near future?”

A new development of surveys of opinions about the size of the family was received after the Second World War. In 1955 and 1960 In the United States, two nationally representative surveys were conducted, known as the American Family Development Surveys (GAF-1 and GAF-2). They examined opinions about the preferred number of children and the degree of their implementation after 5 years. In the GAF-1 and GAF-2 studies, for the first time, along with the ideal number of children, opinions were also revealed about the expected and desired number of children. In the same years, similar surveys were conducted in about 30 other countries 52 .

In our country, surveys of opinions on family size resumed after a long break in the 1960s. 20th century Employees of the Department of Demography of the Research Institute of the Central Statistical Bureau of the USSR in the period 1966-1984. 8 surveys were conducted on the basis of the budget network, the results of which were published in a number of monographs and articles 53 . Similar studies were carried out by other authors. In total for the 60-80s. of the last century, several dozens of such surveys were conducted in various regions of the USSR. In modern Russia, this kind of survey of opinions about the size of the family is regularly conducted by the All-Russian Public Opinion Research Center (VTsIOM). In addition, as mentioned above, questions about the expected and desired number of children in the family were included in the 1985 and 1994 microcensus programs.

Demographic surveys are also carried out at the international level. Among them, the most famous is the World Fertility Survey (WFS), conducted in 1974-1982. by the International Statistical Institute and the International Union for the Study of Population Problems in 41 developing and 21 developed countries 54 and the Demographic and Health Survey (DHS) carried out in the early 1990s. in 59 developing countries by the American Development Resources Institute 55 .

As regards the actual sociological research, their history is shorter. In fact, in our country, it began only in the 70s. of the last century, when the Center for the Study of Population Problems of the Faculty of Economics of Moscow State University. M.V. Lomonosov, the first truly sociological and demographic study "Moscow-1978" was conducted (headed by A.I. Antonov), the program of which was based on the theory of reproductive behavior. This study was preceded by two pilot sociological surveys conducted by the same group of authors in 1976 in Moscow and Vilnius. The results of the study "Moscow-1978" were published in a number of articles, pamphlets and monographs 56 .

Later, according to a similar program under the guidance of A.I. Antonov and V.A. Borisov, a number of sociological and demographic studies were carried out in Moscow, Saratov and Ufa 57 . According to a similar program in the second half of the 1980s - the first half of the 1990s. conducted a number of studies of the Ural family of A.I. Kuzmin 58 . A feature of these studies was the orientation towards a comprehensive study of reproductive behavior in the unity of its attitudes, motives and results. Accordingly, they “used a set of indicators, supplemented by new methods for measuring the degree of realization of the need for children, the degree of coincidence of the attitudes of both spouses, as well as a number of new methods created on the basis of the “semantic differential” technique 59*.

Sociological studies of reproductive behavior were continued in 1999-2001. by the staff of the Department of Sociology of the Family of the Sociological Faculty of Moscow State University. M.V. Lomonosov (headed by A.I. Antonov).

The eighties of the last century were the time when sociological studies of mortality problems began in our country. At the same time, the attention of researchers was focused on the study of self-preserving behavior - a concept that was introduced into sociological demography by analogy with reproductive behavior**. The beginning of the study of self-preservation behavior was laid by studies conducted in 1980-1983. at the Center for the Study of Population Problems, Faculty of Economics, Moscow State University. M.V. Lomonosov under the leadership of A.I. Antonov and then continued at the Institute of Sociological Research of the USSR Academy of Sciences (now the Institute of Sociology of the Russian Academy of Sciences). In 1985-1993 studies on a similar program were carried out in the Urals by A.I. Kuzmin. The results of these studies are reflected in a number of publications***.

4. METHODOLOGY OF STATISTICAL POPULATION SURVEY
4.1. HOUSEHOLD SURVEYS
4.1.1. HOUSEHOLD BUDGET SURVEY METHODOLOGY
Goals and objectives of statistical observation

Household budget surveys are multipurpose in nature. Its main tasks are defined as obtaining weights for the calculation of the consumer price index and data for compiling the accounts of the household sector in the system of national accounts. Traditionally, the budget survey is also a source of statistical data on the distribution of the population according to the level of material well-being, on the level of poverty and food consumption.

The household budget survey is conducted in all republics within the Russian Federation, territories and regions using a sample method and covers 49,175 households.

The survey is based on the principles of voluntary participation of selected households.

Study population definition (survey scope)

The general population in the selection is made up of all types of households, with the exception of collective households (the part of the population consisting of persons who are long-term residents of hospitals, nursing homes, boarding schools and other institutional institutions, monasteries, religious communities and other collective residential premises).

Definition and procedure for the formation of the statistical basis of observation,
definition of sampling unit, observation unit

The information array of the 1994 population microcensus served as the basis for constructing a sample of households.

the presence (absence) of household, garden, suburban area, vegetable garden

Distribution of examined persons:

for living in a household of a certain size

from 1 person, 2 people, 3 people, 4 people,
5 people, 6 people, 7 or more people

according to the age

0-6 years old, 7-12 years old, 13-16 years old, 17-29 years old, 30-39 years old, 40-49 years old, 50-54 years old, 55-59 years old, 60-64 years old,

65 years and older

men; women

by source of livelihood

Form No. 1 (section 1, question 2, column 2)

The sum of the values ​​of indicators R1V212 - R1V252

Expenses for the purchase of jewelry

Form (section 5c,
column 4);
form (section 1c,
column 4)

Amount by code 941

Expenses for the purchase of building materials for construction and overhaul

Form (section 5b,
column 5); form (section 1b, column 5)

Amount by codes

Payment for construction and overhaul services

Form (Section 6,
column 4); form (section 2, column 4)

Amount by code 502

Intermediate consumption expenditure and gross fixed capital formation

Summary indicator

e + f + h + i + k

Taxes, fees and other obligatory payments

Form No. 1 (section 1, question 1, column 2)

The sum of the values ​​of the R1V112 indicator

Other expenses

Form No. 1 (section 1, question 1, column 2; question 3, column 2; question 5); form
(section 6,
column 4); form (section 2,
column 4)

The sum of the values ​​​​of indicators R1V122-R1V152 (repayment of a loan, repayment of loans, payment of alimony, insurance and membership fees, given free of charge to relatives and friends), R1V362 (purchase of other real estate), R1V511 (other expenses) in form No. 1 minus the amount of codes 961- 965 (insurance services) according to forms No. 1-b

Cash expenses

Summary indicator

1 + 2 + 3 + 4

Amount of savings made

Form No. 1 (section 1, question 9)

Sum of R1V911 indicator values

Amount of loan and spent savings

Form No. 1 (section 1, question 7)

The sum of the values ​​​​of the indicator R1V711

Growth of financial assets

Summary indicator

l - m

Cash income

Summary indicator

I + II

In-kind food value

Form (section 2,
column 4)

The sum of estimates at the average purchase prices of in-kind receipts of food [(codes 101-108, 121-134,141-145, 151-163, 171,172, 181-189, 201-210, 221-227, 241, 242, 244, 261-263 , 271, 272) x kz ]

The value of subsidies and benefits provided in kind

Form No. 1 (section 3, question 17)

Sum of R3V172 indicator values

The value of natural receipts

Consolidated
index

Gross income

Consolidated
index

III+ IY

The value of household donated food

Form No. 1-a (section 3,
column 3)

The sum of estimates at the average purchase prices of natural food transfers [(codes 101-108,121-134, 141-145,151-163, 171,172,181-189,
261-263, 271.272) x kz ]

Final consumption expenditure

Summary indicator

1 + IY - R

Available
resources

Summary indicator

I + l+ IY or Y +m

Development sections and the order of formation of grouping features
based on survey results

The results of the survey are developed for the Russian Federation as a whole and for the regions included in the survey plan, in the following sections:

I. Geographically, determined on the basis of the address part of the survey forms:

households in urban areas;

households in rural areas;

II. According to the form of quantitative expression of the developed indicators:

absolute data are summary results of the survey, obtained by summing up the weighted data of individual household budgets covering the entire population of respondents or its part belonging to one or another group;

average data per 100 household members are calculated by dividing the absolute data by the weighted number of current members household, determined on the basis of data of monthly registration of the number of persons living in the household. Cash persons include all members of the household, except for those who are absent for a long time (on a business trip, called up to the ranks Russian Army students in boarding schools, etc.);

relative data are given as percentages and calculated from absolute data;

III. In groupings according to a number of socio-economic characteristics. Grouping features in the development of survey results are:

1) composition of households:

2) socio-demographic typology of households:

The source of information for the formation of these types of households is the Household Register, which contains information on all cash members of the household as of the end of the quarter (section 3 of the Household Budget Survey Questionnaire, lines 1, 4);

3) indicators characterizing the level of welfare of households.

The system of indicators characterizing the welfare of households based on the results of the survey includes:

cash income;

cash expenses;

gross income;

available resources.

Groupings are constructed using the method of ranking individual household budgets in ascending order of the average per capita value of the attribute used as the basis for assessing well-being. Using this method, the surveyed households and the population in them are grouped:

by decile (10 percent) groups of the surveyed population;

according to the interval series of the distribution of households (population) depending on the size of the welfare indicator;

When constructing groupings by decile groups, the order is applied in accordance with which the weighted data on the number of persons in households are ranked as the average per capita indicators of well-being increase and are summed up cumulatively to obtain the total number of the surveyed population. This number is taken as 100%. The sum of all households, where 10% of the total number of the surveyed population is concentrated, refers to the corresponding decile group of the population, distributed as well-being indicators increase.

The key characteristic for assessing the uneven distribution of household welfare indicators is the coefficient indicating the ratio between the total values ​​for the most and least well-to-do population groups. In developing household budget survey data

as such a characteristic, the coefficient of funds is used, the calculation of which is carried out according to the following formula:

When constructing groupings according to the interval series, the order is applied in accordance with which the weighted data on the number of persons are ranked as indicators of the level of well-being increase and are summarized within the boundaries of the interval series. The boundaries of the interval series are reviewed annually.

When constructing a grouping by category with a welfare level below the subsistence level (poor households), an order is applied in accordance with which the weighted data on the number of persons are ranked as the indicators of the level of well-being increase and are summarized within the boundaries up to the subsistence level.

Among the characteristics that explain the position of the category of households with a level of well-being below the subsistence level, the following indicators are used in developing the results of the household budget survey:

the percentage ratio of this category of households (the population in them) to the total number of households (the population);

shortage of funds for this category of households, necessary to bring their level of well-being to the subsistence level.

The first group of indicators includes poverty and extreme poverty rates.

Poverty ratio:

,

The second group of indicators includes indices of the depth and severity of poverty. Both indices characterize the ratio of the deficit in the welfare level of households (i.e., the difference between one of the welfare indicators and the subsistence minimum) to the subsistence minimum per person. The difference between the two indicators is that when calculating the depth of poverty index, a simple ratio of deficit to the subsistence minimum is taken, and when calculating the severity of poverty, it is raised to the second power, which ultimately makes this characteristic more sensitive to the level of well-being of the poorest.

Borders confidence interval for the mean value of the feature.

Research University graduate School economy Moscow Problems of using the results of sample surveys of households to model the structure of their expenditures When analyzing and modeling the structure of household expenditures, dependence on available statistical information is critical. An analysis of the dynamics of expenses both for individual products and for large groups clearly shows a fairly strong volatility in the structure of expenses. To model the structure of household spending in Russia at this stage...

If this work does not suit you, there is a list of similar works at the bottom of the page. You can also use the search button

Matytsin M.S.

Research University

Higher School of Economics, Moscow

Problems of using the results of sample surveys of households to model the structure of their expenditures

When analyzing and modeling the structure of household expenditures, the dependence on available statistical information is critical. When forecasting such a structure, it is not possible to confine ourselves to macroeconomic data only - a more accurate tool is needed. An analysis of the dynamics of expenses, both for individual products and for large groups, clearly shows a fairly strong volatility in the structure of expenses. To take it into account, it is necessary to involve additional microeconomic characteristics and parameters of the object under study, which can be obtained using the results of sample surveys.

However, in most known works, due attention is not paid to the problems of formation and primary processing data. In fact, such modeling cannot be implemented without using the results of special surveys of household budgets, the organization and conduct of which can only be carried out by state statistics bodies or by the combined forces of several scientific organizations.

At this stage, two main sources of information are used to model the structure of household expenditures in Russia - the Sample Household Budget Survey (HBS) - a quarterly survey conducted by Rosstat, and the Russian Economic and Health Monitoring (RLMS).

1. The HHBS is characterized by a large number of observations, high frequency and breadth of issues covered.

These data have a number of restrictions on their use for the purpose of building a model of the structure of household expenditures. Some time ago, Rosstat opened free access for all interested researchers to survey microdata. However, they full use hampered by a number of factors. Firstly, not primary data from diaries and surveys are provided, but a set of grouped characteristics in the context of households. Thus, from the point of view of analyzing the structure of expenditures, the absence of data on purchases of individual goods is critical - only information is provided on expenditures on groups of products of varying degrees of aggregation. That is, the data is available only in value terms and there is no information on purchase prices (or physical volume), which is necessary as in assessing most demand models ( AIDS -model, Rotterdam, translog models), and to study the relationship between these prices and the level of income, including testing the hypothesis of price endogeneity.

From the point of view of panel analysis, panel repairs carried out by Rosstat can also become a significant problem. In this case, it is not so much the natural depletion of the panel with the subsequent addition of new observations that is critical, but the assignment of numbers of retired observations to new families, without an explicit indication of such a replacement. As a result of data comparison in different periods it turns out that a household with the same number can have very different characteristics, which, for general reasons, should remain relatively stable. So in any two adjacent periods there is a sufficiently large number of families (up to several percent), for which the size of the family (by 3-5 people) is very different, as well as the number of the 10% income group to which they belong. Such jumps, most likely, testify to the replacement of the observation under this number with a new one.

Significant problems arise with the distribution of sample data to the entire population of families. Although Rosstat offers a mechanism for re-weighting survey results to better match the parameters of the general population, even the weighted results deviate strongly from macro statistics. One of the basic parameters - the average income - differs by tens of percent (see table). Moreover, even indicators calculated on the basis of a sample survey are subsequently significantly adjusted. So, for example, the income differentiation coefficients of the population (Gini coefficients and funds) in the aggregated results given by Rosstat are quite different from the value of this indicator published by Rosstat in official reference books:

2. RLMS, which has a section on household expenditures on various goods in its structure, is the second source of information that partly fills in the gaps. official statistics. The RLMS survey is conducted in the form of annual household interviews and, unlike the HHBS, does not involve completing diaries or journals over a period of time.

From the point of view of the analysis of the structure of expenditures, information on purchases of food products, tobacco and alcohol (56 categories in total) is key, not only in value terms, but also in kind, which makes it possible to obtain individual data on purchase prices.

Data are presented on spending on a range of non-food items (clothes, children/adults, appliances, fuel, etc.) and services. Due to the natural complexity of measuring the corresponding parameter, these data are not given in physical terms, but only in value terms.

Information on expenditures for the purchase of certain types of goods and services is collected on the basis of various time horizons. Thus, expenses for the purchase of food products are recorded only for the period of the week preceding the interview, which obviously introduces too much random factor - far from all (including regularly consumed) food products are bought by the family every week.

Thus, a number of items may be omitted from the survey when the household completes the relevant section, simply because of the short survey horizon. This leads to significant inconsistencies in the range of purchases for each family in neighboring survey waves and makes intertemporal comparisons much more difficult.

During the analysis of the data, two main problems were identified that affected the results and the accuracy of the estimation. The first is an incomplete system of answers. That is, a situation where the respondent indicated that he purchased a particular food product, but "forgot" to indicate the quantity or cost (or, more rarely, both). This misrepresentation slightly underestimates the total cost, as the number of omissions is relatively low: 97.7% of respondents indicated both the cost and the quantity of the item purchased.

The second problem, in contrast to the first, is more substantive in nature (whereas the first is more technical) and can be described as “looseness” of the data. Many households, especially families with relatively low incomes, purchased only a small part of the range of commodity items in the survey during the reporting period (for example, for 2006 data, on average, families purchased 15.6 items out of 56, or 28% of the range). This behavior may also be related to the short survey horizon of one week. This problem can be partially solved by moving from calculations at the level of individual indicators to work with homogeneous groups of households.

Unlike HHBS in RLMS when the panel is repaired, the numbers are not repeated, so the use of intertemporal comparisons is more correct, including for the structure of expenditures of individual households. Such an analysis (on the example of 2005-2006) confirms the conclusions about the “looseness” of the data. With the horizon of the survey on food purchases only one week long, for the same families, not only the structure, but also the nomenclature of purchases in neighboring waves of the survey is quite different. So, only 10 families (out of 4 thousand) completely repeated their purchases in two neighboring years. The average match value is 78%. Although this value does not seem too low, it significantly distorts the values ​​of the price indices. Due to the discrepancy between the nomenclature of family purchases in the two periods, the calculation of the Laspeyres price index is significantly underestimated (the value is 0.69 in 2005-06), and the Paasche price index is overestimated (2.84).

Such a distortion is manifested not only in the deviation of the indices from adequate values, but also in their ratio. From the chart ( rice. one ) shows that the Laspeyres and Paasche indices are very weakly dependent, so the choice of the index can radically affect the conclusions about the price dynamics of an individual food basket.

Rice. 1. The ratio of individual price indices according to the data RLMS 2005-2006

When analyzing the dynamics of prices and expenditures of individual families, observations with enormous gaps are revealed. So, not only the level of purchase prices can differ hundreds of times for one family in two adjacent periods, but also the total amount of expenses and the amount of food expenses can differ by 150-170 times.

The described "looseness", the presence of obvious inaccuracies (including omissions in answers) and significant scatter in individual data do not allow working directly with the initial information on individual households when modeling the structure of expenses. To analyze and model such a structure, including for forecasting purposes, it is necessary to single out groups of households that are homogeneous in terms of the level of expenditures or other characteristics.

This approach turns out to be quite effective. In particular, it is possible to investigate the relationship between the level of expenses and purchase prices by calculating "spatial" price indices. More details about the implementation of this method are given in the report of Ershov E.B. and Matytsina M.S. " Economic theory and Statistical Practice of Analysis of Consumer Behavior of Households” within the framework of this Congress.

Modeling and medium-term forecasting of the structure of household spending in the Russian economy is an important task, especially relevant during the crisis. However, such modeling, which is impossible without taking into account microeconomic factors at the level of individual families or homogeneous groups, is significantly hampered by the lack of free access to the necessary statistical information.

To partially fill the gaps in statistics, it is necessary to combine information from various sources, including combining the results of the HBS and RLMS , which is associated with a number of both technical and substantive problems. An analysis of individual sample indicators testifies in favor of the need to combine observations into homogeneous groups for more correct modeling of the expenditure structure. A separate problem is the coordination of the results of models at the level of sample surveys with macrostatistics and their distribution to the general population.

1. Deaton A. The Analysis of Household Surveys: A Microeconomic Approach to Development Policy, World Bank Publications, 1997.
2. Consumer price index manual: theory and practice (ILO, World Bank, IMF, Eurostat), 2004.
3. Bondarev A. Estimation of demand functions for groups of food products in the Russian economy for 1999-2004, M.: IET, 2004.
4. Köves P. Theory and practice of economic analysis. Moscow, Finance and statistics, 1990.
5. Dumnov D.I. On the methodology and organization of sample surveys of households. Economic science in modern Russia, No. 2, 2002.

PAGE \* MERGEFORMAT 1

Approved

by order of the Chairman

Committee on Statistics

Ministry of National

economy of the Republic of Kazakhstan

Methodology for conducting a sample survey of households

Chapter 1. General Provisions

1. Methodology for conducting a sample survey of households (hereinafter referred to as the Methodology) refers to a statistical methodology formed in accordance with international standards and approved in accordance with the Law of the Republic of Kazakhstan dated March 19, 2010 "On State Statistics" (hereinafter referred to as the Law).

2. The methodology establishes the main aspects and methods for analyzing the sample and general population of households and is intended for use by structural divisions of the Committee on Statistics of the Ministry of National Economy of the Republic of Kazakhstan (hereinafter referred to as the Committee).

3. A subset of households is selected for sample surveys, and observations are made or data are collected within this subset. The results obtained are extrapolated (distributed) to the entire population as a whole.

4. The main advantages of using the sampling method in modern statistics are:

1) reducing the time for statistical observations (surveys);

2) reducing the information load on respondents;

3) significant savings in labor costs, material and financial resources during the examination;

4) significantly accelerated receipt of the results of the study compared to a complete survey.

5. This Methodology uses concepts in the meanings defined in the Law, as well as the following definitions:

1) Ppanel observation method- a method of collecting information, in which a certain group of units of analysis is periodically polled for a relatively long time, and the subject of the study remains constant;

2) the general population - a complete group of all units of analysis, whose characteristics are subject to evaluation;

3) representativeness - the correspondence of the characteristics of the sample to the characteristics of the population or the general population;

4) mathematical expectation - the average value of a particular characteristic in all possible samples, as well as the weighted average of all possible results with a weight of probabilities reflecting the possibility of occurrence in each result;

5) parameter - a value calculated from all values ​​in the set of the general population, that is, a descriptive measurement of the general population;

6) stratum - division into special layers of units (respondents) with the same or similar indicators;

7) sampling plan - a set of specifications that define the general population and sample units, as well as the degree of probability of possible samples;

8) sampling set (sample) - a set of cases (subjects, objects, events, samples), using a certain procedure, selected from the general population for participation in the study;

9) sample size - the total number of observation units in the sample.

Chapter 2Planning andsampling process

6. When planning a survey, it is necessary to determine the geographic areas to be covered and the population to be surveyed.

7. When determining the statistical population, it is necessary to identify the population group from which the sample is formed. Outlying areas with few households or residents are removed from the sampling frame because their coverage is too expensive. They represent only a small fraction of the population, their impact on population figures is very small. The report on the results of such a survey clearly indicates the exclusion of these areas.

8. The process of forming a sample for conducting a survey consists of several stages:

definition of the general population;

establishing a sampling frame;

choice between probabilistic and improbable methods of selection;

definition of sampling plan;

determination of the sample size;

direct sampling according to the plan.

Chapter 3Definition of populationand sampling frame

9. Population census data or information system, the statistical register of housing stock (hereinafter - IS HRHF) is the main source for the formation of a sampling frame for household surveys in the Republic of Kazakhstan. Population census data serve as a means of providing information on the size, composition and geographical distribution of the population, in addition to socioeconomic and demographic characteristics. The population census collects information on each individual in the household and on each set of dwellings throughout the territory. In order to avoid cases of non-receipt of data from respondents, IS SRZHF is used to form the sample. IS SRZHF was created for the purpose of generating and accumulating data on residential buildings and dwellings for housing stock statistics and sampling for household surveys.

10. Accounting units in IS SRZHF are all residential buildings and residential premises (apartments) located on the territory of the Republic of Kazakhstan.

These include:

residential premises (apartment);

single-family (individual) house;

semi-detached house;

three or more apartment buildings.

Each house and apartment has an identification number (hereinafter - ID).

In addition, IS SRZHF contains the following data: apartment ID, classifier of administrative territorial objects (KATO), street, house number, apartment number, total area, living area. Actualization of the data contained in the IS SRZHF is carried out daily.

Chapter 4Sampling strategy and methods

11. To select elements from the general population, the following methods of probabilistic selection are used:

systematic random sampling (step sampling);

a sample with a probability proportional to size (hereinafter referred to as VLOOKUP).

12. Simple random selection (sample) provides an equal probability of being selected for each element of the general population. There are the following varieties of this method:

repeated random selection;

repetitive random selection.

13. Non-repeated random selection gives more accurate results of sample observation compared to repeated, since with the same sample size, observation covers more units of the general population. In cases where non-repetitive sampling is not possible, re-sampling is used.

14. The essence of systematic random sampling is the selection from the base of the element, starting with the first element, which is selected at random.

For example, when forming a systematic sample with a size of 500 elements from a general population of 15,000 employees of an organization.

First, a random start is determined, then a selection step. (15,000/ 500=30, selection step is 30).

15. The VLOOKUP selection method improves the estimation accuracy if the auxiliary size variable used to determine the probabilities is approximately proportional to the features being studied. When using the CDF method, there is a greater likelihood that units with large features will fall into the sample. The sampling method is often used in household surveys to select areas where the probability of including items in the sample is proportional to the size of the resident population in the area of ​​sampling.

Section 1. Stratified sampling

16. In designing a household survey, a widely used technique is stratification for the survey population prior to sampling. It serves the purpose of classifying a population into subpopulations based on additional information that is known about the population. For example, territorial characteristics or gender and age categories, type of area, number of residents, type or type of structure, building. The main principle of formation of strata (stratification) is heterogeneity between strata and homogeneity within strata. Urban and rural areas are formed as two separate strata for the household survey. Urban and rural populations differ from each other in many aspects (type of employment, source and size of income, average household size, birth rate) while persons belonging to one of these subgroups have similar characteristics. The probability of selection with a starified sample using non-repetitive random selection is calculated by the following formula:

https://pandia.ru/text/80/295/images/image005_10.png" width="20" height="24 src=">- the size of the general population in the stratum.

17. The advantages of stratified sampling are:

Δ is the marginal sampling error.

23. To determine the sample size, the following parameters of the population are estimated:

1) The arithmetic mean (for example, household income and expenses, the number of people living in households) is calculated for all units of the general population and is called the general average () and is calculated using the following formula.

https://pandia.ru/text/80/295/images/image027_3.png" width="16" height="25 src="> - the sum of the i-stratum indicator.

2) Population variance is defined as the mean of the squared deviations of all individual observations from their mean.

Population variance is calculated using the following formula:

The square root of the variance is called the standard deviation or standard deviation and is calculated using the formula:

3) If the error is expressed as standard error ( m), then the following formula is used to determine the sample size:

RSE is the relative standard error of the sample.

If the final population adjustment is not taken into account, the formula for determining the sample size will be as follows:

24. Once the sample size has been determined, the sample should be allocated to strata if it is a stratified sample or to clusters if it is a cluster sample. The sample distribution is made by the same sample size in each stratum (uniform distribution), or distributed in other ways. In order to determine the distribution of the sample to different strata, there are two important criteria that affect how the sample size in the strata is determined:

The first criterion is convenience: a method of proportional distribution is chosen in which the sample size in i-th stratum is calculated by the formula:

ni- sample size i- strata;

i = 1,2,…,h;

Ni- number of households in i-th stratum, while i = 1,2…., h.

The second criterion is accuracy: the method of optimal distribution is chosen, which gives the smallest mean square error (standard error) of the sample.

25. Where the costs of sampling from different strata are the same, the optimal distribution formula is called the Neumann distribution. In this case, the sample size in i th stratum is determined by the formula:

https://pandia.ru/text/80/295/images/image034_2.png" width="16" height="29">, then its basis weight, denoted as (spread factor), is calculated by the formula:

32. The problem of non-response from the sample unit for subjective reasons in household surveys is solved by adjusting the sample weights. The calculation of the adjusted non-response weight for the i-th sample unit is calculated using the following equation:

https://pandia.ru/text/80/295/images/image038_1.png" width="51 height=29" height="29"> is the number of actually reported.

Calculation of the final adjusted weight in case of non-response for i The th sampling unit is calculated using the following equation:

https://pandia.ru/text/80/295/images/image040_1.png" width="34" height="23 src="> is the initial basis weight;

DIV_ADBLOCK117">

Estimate of the standard error of the sample. Possible discrepancies between the characteristics of the sample and the general population are measured by the standard error (mean error) of the sample. The sample standard error is determined by the following formula:

m - standard error;

General dispersion;

Sample size.

36. Sample standard error shows the absolute values ​​of the error. The relative standard error (coefficient of variation) is used to determine the estimated value in fractions. This coefficient is expressed as a percentage and is calculated by the formula:

https://pandia.ru/text/80/295/images/image048_0.png" width="25" height="30">.png" width="25" height="30 src=">- | £ D, from which it follows that x - D £ https://pandia.ru/text/80/295/images/image050.png" width="113" height="32 src=">

The sum of the indicator of the sample i-strate;

The average value of the indicator of the sample population i-strat.

The development of market entities is not possible without marketing research. The scope of this kind of research is constantly expanding.

Any marketing research should be based on objectivity, accuracy and thoroughness. Objectivity means that the research process takes into account all possible facts, and the final conclusions are not formulated until the data is collected and analyzed. The accuracy of research results depends on the careful choice of research tools and methods of their application.

Marketing research is aimed at identifying and evaluating the key factors and trends underlying the strategy and tactics of market participants.

An important guideline for marketers is compliance with following principles marketing research:

• - Systematic - research should be conducted systematically, and not be sporadic one-time;
• - Consistency - research should cover the entire market and the entire structure of hierarchical market processes, factors, their dynamics and relationships;
• - Complexity - means, on the one hand, that the study includes a set of actions or processes (data collection, processing, analysis), on the other hand, that an integrated approach is applied to the study of objects of their relationships with other processes and objects;
• - Connectedness and purposefulness - the direction, scope, depth, detailing of the research being carried out should be organically linked to the goals and objectives of the participant in this market, reflecting his real needs for specific analytical information;
• - Plurality of sources of information - it is advisable to receive information not from one, but from several sources, which allows you to have comprehensive "overlapping" data and thereby clarify, verify information, discard doubtful data;
• - Universality - studies can be carried out in order to meet any market participant's need for rational decision information;
• - Scientific - research should be characterized by accuracy, objectivity, conditionality. Insufficiently objective, unfounded studies lead to incorrect, distorted recommendations.

In addition to the basic principles, marketing research methods are also important, which allow solving a wide range of marketing tasks using various modern techniques. There are: general scientific, analytical and prognostic research methods, methodological techniques borrowed from different fields of knowledge.

Population censuses are the primary source of information about the population. Population census - the process of collecting demographic, economic and social data characterizing each inhabitant of a country or territory as of certain moment time.

The last All-Russian population census was conducted in 2002 as of 0000 hours on October 9th.

Household budget surveys are multipurpose in nature. Its main tasks are defined as obtaining weights for the calculation of the consumer price index and data for compiling the accounts of the household sector in the system of national accounts. Traditionally, the budget survey is also a source of statistical data on the distribution of the population according to the level of material well-being, on the level of poverty and food consumption.

The household budget survey is conducted in all republics within the Russian Federation, territories and regions using a sample method and covers 49,175 households.

The survey is based on the principles of voluntary participation of selected households.

The general population in the selection is made up of all types of households, with the exception of collective households (the part of the population consisting of persons who are long-term residents of hospitals, nursing homes, boarding schools and other institutional institutions, monasteries, religious communities and other collective residential premises).

The information array of the 1994 population microcensus served as the basis for constructing a sample of households.

The use of this database was due to a number of advantages, which include the following:

the availability of a ready-made sampling frame on machine media, which eliminates one of the significant cost items associated with compiling the sampling frame;

exclusion from the sampling procedure of a number of intermediate stages of selection related to the selection of administrative units and the determination of the specific place of residence of households;

the presence on machine media of socio-economic and demographic information about the population (a specific household, person) both in the context of the district, enumeration area, and in the whole region.

The final unit of selection is a household, which is a set of persons living in the same dwelling or part of it, both related and not related by kinship, jointly providing themselves with everything necessary for life, fully or partially pooling and spending their funds. A household may consist of one person living independently.

The household is also the unit of the survey.

The sample for budgetary surveys is formed on the principles of representativeness of the category "the entire population" within a particular region of the Russian Federation (republics within the Russian Federation, territories and regions).

The construction of a territorial sample of households is based on a two-stage model of probabilistic (random) sampling using a stratification procedure at each of the selection stages. The stratification procedure is aimed at forming a representative sample of households that adequately reflects the territorial features of the stratification of the population, its demographic and socio-economic structure, objectively representing structural changes in the transition to a market economy.

The layering procedure consists of two levels:

• 1. Territorial stratification (in two layers) taking into account the place of residence of the population:
  • - urban area;
  • - countryside;
• 1. Formation of sublayers in each layer, taking into account the structure of socio-economic and demographic characteristics. The system of such features includes grouping features presented in Table 1.

Table 1. Grouping characteristics for the formation of a sample population of households

The implementation of the survey program is carried out by maintaining the following survey forms:

form No. 1-a "Diary of accounting for daily household expenses";

form No. 1-b "Journal of accounting by the household of purchases of non-food products and services received";

form No. 1 "Questionnaire of the survey of household budgets".

The survey is based on direct questioning (interviewing) of household members and keeping household records of current consumption expenditures.

The reference period for a single household survey covers a quarter. Within three months, differentiated data collection procedures were established depending on the composition of the indicators.

At the end of the annual survey cycle, an annual survey of households is conducted. The program of the annual survey includes the collection of data on the housing conditions of households, the availability of durable items in households, the turnover of livestock in personal subsidiary plots, and the level of education of household members.

The procedure for collecting data on diary and journal surveys is organized on the principles of household rotation within one observation area.

For this purpose, the set of households surveyed by each interviewer is divided into three rotational subgroups, the grouping criterion for the formation of which is the size of the household.

The survey procedure should ensure that each group of households falls under different types of data collection during the month: weekly diary entries, 2-3 weekly and monthly journal entries.

Correction and addition of diary and journal entries can be made by the interviewer when he visits the household only with the consent of the subjects.