New Russia Barometer: Samples
New Russia Barometer I: 26 January-25 February 1992
The survey was conducted by VCIOM, the Russian Centre for Public Opinion Research, in consultation with Dr. Irina Boeva and Dr. Viacheslav Shironin, economists in the former USSR Academy of Sciences, and supported by a grant from the Centre for Research into Communist Economies. The universe consisted of Russians age 15 or older in the urban population, 73.3 per cent of the total population of Russia in the 1987 census.(1) A representative random sample of 34 cities and towns was selected; there was further stratification for areas, and for the selection of individuals for interview within households. Geographical dispersion is particularly important, for many surveys cover only one city or a few major cities, even though Moscow and St. Petersburg together constitute less than 15 per cent of the country's population. The effective sample consisted of 2,547 persons; a total of 2,106 interviews were completed. The survey was confined to urban areas because of difficulties at that time in carrying out field work in rural areas. Subsequent comparison with urban and rural respondents in NRB II showed that on the great majority of political questions there was no significant difference between urban and rural respondents, and the impact on economic behaviour and attitudes was predictable and, with few obvious exceptions, such as growing food, marginal. Interviews were in the household rather than by telephone, as half of households did not have a telephone. To correct for the under-representation of less educated persons, the sample was weighted by education. For full details, see Boeva and Shironin (1992: 40-43).
New Russia Barometer II: 26 June-22 July, 1993
This survey conducted by VCIOM and financed by a grant from the British Foreign Office Know How Fund to study popular response to privatization. The universe of the sample was the population of the whole of the Russian Federation age 16 and above, rural as well as urban. VCIOM stratified the Federation into 15 Russian regions. Each region was stratified into urban and rural areas, and the urban population stratified proportionate to population into oblast and krai capitals; peripheral towns of oblasts and krais; capitals of autonomous regions; and peripheral towns of autonomous regions. Within this framework, towns were randomly selected with a probability equal to their share of the total population. A total of 85 urban and rural areas were selected, and 153 primary sampling units were drawn with a probability proportionate to population. Within each primary sampling unit households were listed by addresses and up to 30 addresses drawn randomly. Within each household the person whose birthday was nearest the date of interview was selected as the respondent. There were a total of 1,975 interviews, and a response rate of 83 percent. VCIOM used a multiple regression analysis to derive appropriate weights for gender, age, education and town size. For full details, see Rose, Boeva and Shironin (1993: 56-58).
New Russia Barometer III: 15 March-9 April, 1994
The third New Russia Barometer survey was organized by the Paul Lazarsfeld Society, Vienna, in collaboration with the CSPP. Fieldwork was undertaken by MNENIE Opinion Service, Moscow, under the direction of Dr. Grigory A. Pashkov. Financial support came from a group of Austrian banks and a grant to the CSPP from the British Know How Fund. The questionnaire was piloted with 100 interviews in Rostov on Don and Petrozavodsk, 26-28 February 1994, and revised by Professor Richard Rose and Dr. Christian Haerpfer. The universe for the sample was the population of the Russian Federation age 18 and above. The Federation was first stratified into 14 different regions. At stage two each region was further stratified on a proportionate-to-population basis according to urban and rural divisions and differences in the administrative structure of each region. Within each region seven sampling points were selected with the number of interviews proportionate to population. At the third level, interviewers selected a sample of respondents on the basis of a quota set for each region by sex, age, level of education and nationality. The target number of responses was 3500; in the event, 3,535 interviews were completed. The validity of the interviews was randomly spot-checked by MNENIE and by Ernst Gehmacher, director of IFES, Vienna. Comparison of characteristics of survey respondents with census figures showed a good fit for the sample distribution by gender, age, education and region. Hence, no weighting was used. For sample details, see Rose and Haerpfer (1994: 54-55).
New Russia Barometer IV: 31 March-19 April, 1995
The fourth New Russia Barometer survey was conducted by VCIOM and financed
by a grant from the British Foreign Office Know How Fund. The universe was
the population of the whole of the Russian Federation age 16 and above. VCIOM
first stratified the Federation into 10 Russian regions. Each region was stratified
into urban and rural areas, and the urban population stratified proportionate
to population into oblast and krai capitals; peripheral towns of oblasts and
krais; capitals of autonomous regions; and peripheral towns of autonomous
regions. Within this framework, towns were randomly selected with a probability
equal to their share of the total population. A total of 62 urban and rural
areas were selected, and 85 primary sampling units were drawn with a probability
proportionate to population. Within each primary sampling unit households
were listed by addresses. An address was selected randomly as the starting
point for interviewing, and interviewers assigned with instructions to obtain
10 interviews. At each household the interviewer was instructed to ask for
the resident whose birthday was closest to the date of interview and, if the
respondent fell within a prescribed demographic matrix, to conduct an interview.
A total of 2,770 contacts were made, and 1,998 were interviewed face to face,
a response rate of 72.1 percent. The representativeness of respondents was
compared with Goskomstat data for the Russian population in 1993. The data
were weighted to match the regional census population in terms of gender,
age, education and town size. For full sample details, see Rose (1995c: 70-72).
New Russia Barometer V: 12-31 January 1996
The New Russia Barometer V survey was organized by VCIOM and financed by
the Centre for the Study of Public Policy. The survey was based on a multi-stage
stratified random sample. VCIOM first stratified the Federation into 11 macro-economic
regions. Geographical dispersion is particularly important. In macro regions
including "Autonomous" administrative units, the macro region was
subdivided into parts with autonomous and without autonomous units to take
into account ethnic composition of territories. A number of autonomous areas
in the far North and the North Caucuses (5 per cent of the total population)
were excluded from the survey. Each region was further stratified into urban
and rural areas, and the urban population stratified proportionate to population
into oblast and krai capitals; peripheral towns of oblasts and krais; capitals
of autonomous regions; and peripheral towns of autonomous regions. Within
this framework, towns were randomly selected with a probability equal to their
share of the total population. A total of 104 urban and rural places were
selected, and primary sampling units drawn with a probability proportionate
to population. Within each primary sampling unit households were listed by
addresses, and an address was selected randomly as the starting point for
interviewing, and interviewers instructed to seek an interview at every nth
house on the route. At each address the interviewer asked for the resident
whose birthday was closest to the date of interview to complete the questionnaire
himself or herself.
A total of 4,508 contacts were made and 2,426 persons were interviewed, a
response rate of 53.8 percent. The fall in the response rate is partly due
to declining interest in being interviewed, evident in established democracies
too, and partly due to this survey's very detailed political questions of
less interest to many respondents than the economics of everyday life. VCIOM
inspected all questionnaires, and verified ten percent of interviews by making
callbacks. When the representativeness of respondents is compared with Goskomstat
data for the Russian population in 1993, the sample shows a good fit for gender,
age, and urban/rural divisions. The data were weighted to match the regional
census population in gender, age, education, town size and, because of proximity
to the general election, party preference. Fuller details in Rose (1996a:
79-81).
A total of 214 interviewers were employed. VCIOM collected data on the voting
preference of its interviewers and they were widely dispersed across the political
spectrum. Yabloko was the only party favoured by more than a tenth of interviewers,
the Communist Party was second and Our Home is Russia third in interviewer
preferences.
New Russia Barometer VI: 25 July-2 August 1996
The sixth NRB survey was organized VCIOM and financed by the Centre for the Study of Public Policy. The fieldwork commenced three weeks after the second round of the presidential ballot and was concluded before the presidential inauguration. The universe of the sample was the adult population of the whole of the Russian Federation age 18 and above. The survey was a multi-stage stratified random sample. VCIOM first stratified the whole of the Federation into 22 regions. Each region was further stratified into urban and rural areas, and the urban population stratified proportionate to population into regional centres and other towns. Within this framework, towns were randomly selected with a probability equal to their share of the population. A total of 69 urban and rural settlements were selected on a proportionate to population basis. Within these locations, primary sampling units were randomly drawn. In each primary sampling unit households were listed by addresses, an address was selected randomly as the starting point, and 161 interviewers instructed to seek an interview at every nth house on the route. At each address the interviewer asked for a respondent matching an age by gender by education grid.
The timing of the presidential election shortly before the start of the summer holidays posed a dilemma. To maximize the number of people at home when an interviewer called would have meant postponing fieldwork until September. But given the uncertainties of the President's health and the political situation in Chechnya, as well as completely unforeseeable events, delay would have risked spoiling the study by introducing post-election events into an election survey, as we, as increasing decreasing the accuracy of reported voting. It was decided to minimize these risks by starting interviewing in late July, even though this invited a lower response rate during the holiday season. In the event, this meant that all interviews were concluded before the presidential inauguration made very visible the state of the president's health, and before quarrels about Chechnya had become a running story too. In all, 3,379 households were contacted with someone at home; in 965 cases no one at that point met the requirements of the interviewer's age by sex by education sampling grid. Of the 2,414 households where an interview was sought, there were 1,599 completed responses, 66.2 percent of the total. When the representativeness of respondents is compared with Goskomstat data for the Russian population in 1993, the sample shows a good fit for gender, age, and urban/rural divisions. VCIOM weighted the data to match the regional census population in terms of gender, age, education, town size and, because of the election, voting. For fuller details, see Rose (1996: 54-56).
New Russia Barometer VII: 6 March-13 April, 1998
The seventh NRB survey was conducted by VCIOM and financed with the assistance of a grant from the Social Capital Initiative of the World Bank, funded by the Development Fund of the Government of Denmark. The universe for the multi-stage stratified random sample was the adult population of the Russian Federation. VCIOM first divided the Russian Federation into 11 regions classified by seven social, economic and infrastructure attributes; in each it has a regional office. At the second stage, each region was stratified according to nine characteristics of population, administrative and urban or rural status. At the third stage, all the oblast, krai and autonomous capital cities were drawn with a probability proportionate to population. At the fourth stage, primary sampling units were chosen for other urban settlements and rural villages at least 50 kilometres distant from the area's chief city. From 28 different republics, oblasts and krais of the Russian Federation, 100 different areas were drawn, 26 republic, oblast or krai capitals, 36 medium or smaller towns, and 38 rural villages. At the fifth stage, each city was stratified into as many six areas according to economic and residential characteristics, and primary sampling units representative of different neighbourhoods within the city. In all, 191 primary sampling units were selected, averaging ten respondents each. At the sixth stage, the address of a household was selected as the starting point of a random route. An interviewer was instructed to call at every nth house from the starting point, until interviews were achieved proportionate to the unit's contribution to the sample. Within the household, one individual was selected as a respondent on the basis of a sex by age by education grid based on 1996 data of the official national statistical office, Goskomstat. Altogether, 5903 contacts were made with households, 1868 did not have a respondent matching grid requirements and in 1035 no one answered. In 3,000 households there was a respondent meeting the sampling specifications, and 2,002 interviews were satisfactorily completed, 66.7 percent of the total (Rose, 1998: 72-74).
New Russia Barometer VIII: 13-29 January 2000
The eighth NRB survey was organized by the Centre for the Study of Public
Policy and conducted by the Russian Centre for Public Opinion Research (VCIOM).
Fieldwork occurred between 13 and 29 January 2000, just after the Duma election
and the seasonal holidays. The universe consisted of the residents of the
Russian Federation age 16 and over, excluding soldiers, convicts and those
without a fixed address.
In its multi-stage sample VCIOM first stratified the Russian Federation into
10 large regions classified by seven social, economic and infrastructure attributes:
North with North-West, Central, Central-Tchernozemie, North Caucasus, Volgo-Viatka,
Volga, Urals, West Siberia, East Siberia and the Far East. Goskomstat population
statistics for 1999 were used to stratify the sample. Each region was assigned
a number of interviews proportionate to its share of the total Russian population.
Geographical dispersion is particularly important in Russia to avoid potentially
misleading results by concentrating interviews in a few major cities, since
Moscow and St. Petersburg have fewer people than rural Russia. At the second
stage, each region was stratified according to its administrative status,
characteristics of population and relation to autonomous republics. Moscow
and St. Petersburg were selected as independent strata. The total number of
interviews within a region was determined according to the proportion each
contributed to the region's population. At the third stage, rural districts
and urban centres were selected with a probability proportionate to their
population. At the fourth stage each was stratified by population characteristics,
e.g. villages within a rural district or wards or precincts within a city.
A total of 107 settlements belonging to 38 subjects of the Russian Federation
were included in the sample--33 oblast centres, 43 cities and towns, and 31
villages. Within these 107 settlements, at the fifth stage 195 primary sampling
units were selected, producing an average of ten respondents per unit.
A total of 192 experienced interviewers were employed; the median interviewer had taken part in more than 15 previous VCIOM surveys; was age about 40 and had a tertiary education. Within each primary sampling unit, households were selected by a random route method, in which the interviewer called at every nth house from a random starting point. Face-to-face rather than telephone interviewing is especially important in Russia, since less than half the population has a telephone. Interviewing in the home is a necessity, given the length of the questionnaire and weather conditions most of the year. Within the household, one individual was selected as a respondent on the basis of a sex by age by education grid based on official statistical data. Controlling for education is specially important to avoid over-representing respondents with above-average education.
Altogether, 4,326 addresses were visited. Of this total, 81 were non-residential buildings, in 378 there was no one in the household who meet the sample requirements; and in 812 there was no one at home. In 3,055 households there was a person meeting the sample specifications. The responses were as follows: Total contacts, 3,055; Refusals, 890; Bad health, unable to answer, 154; Interrupted interviews, 8; Total non-responses, 1,052. Completed interviews: 2,003, or 65.6 percent of total. Whereas 18 is the age of qualification for voting, VCIOM includes youths age 16 or over in its surveys. Because New Russia Barometer VIII is concerned with voting, 63 respondents under the age of 18 were excluded from the data set for the analyses reported here and elsewhere. The total respondents age 18 or over is thus 1,940. The profile of respondents was compared with official Goskomstat data, and the sample was weighted according to gender, age and education in order to bring it into almost exact mathematical representativeness of the population as defined by Goskomstat.
New Russia Barometer IX: 14-18 April 2000
The ninth New Russia Barometer survey was undertaken by the Centre for the
Study of Public Policy and fielded by the Russian Centre for Public Opinion
Research (VCIOM). Fieldwork occurred between 14 and 18 April 2000, promptly
after the presidential election, while the events of the election were fresh
in the minds of Russians and post-election events had not yet unfolded to
influence attitudes. The universe consisted of the residents of the Russian
Federation age 16 and over, excluding soldiers, convicts and those without
a fixed address.
A multi-stage stratified sample was interviewed. In the first stage, VCIOM
stratified the Russian Federation into 10 large regions classified by seven
social, economic and infrastructure attributes: North and North-West, Central,
Central-Tchernozemie, North Caucasus, Volgo-Viatka, Volga, Urals, West Siberia,
East Siberia and the Far East. Goskomstat population statistics for 1999 were
used to stratify the sample. Each region was assigned a number of interviews
proportionate to its share of the total Russian population. Geographical dispersion
is particularly important in Russia to avoid potentially misleading results
by concentrating all interviews in a few major cities, since altogether Moscow
and St. Petersburg have fewer people than rural Russia. At the second stage,
each region was stratified according to its administrative status, characteristics
of population and relation to autonomous republics. Moscow and St. Petersburg
were selected as independent strata. The total number of interviews within
a region was determined according to the proportion each contributed to the
region's population. At the third stage, rural districts and urban centres
were selected with a probability proportionate to their population. At the
fourth stage each was stratified by population characteristics, e.g. villages
within a rural district or a ward or precinct within a city. A total of 27
different oblast centres, 31 cities and towns, and 25 villages were included
in the sample. Within these 83 places, 160 primary sampling units were selected,
averaging 10 respondents in each PSU.
A total of 134 experienced interviewers were employed. Within each primary sampling unit, households were selected by a random route method, in which the interviewer called at every nth house from a random starting point. Face-to-face rather than telephone interviewing is especially important in Russia, since less than half the population has a telephone. Interviewing in the home is a necessity, given the length of the questionnaire. Within the household, one individual was selected as a respondent on the basis of a sex by age by education grid based on national statistical office data. As people with less education are more likely to refuse to be interviewed, controlling for education is specially important to avoid over-representation of respondents with above-average education.
Altogether, 3,567 addresses were visited. Of this total, 39 were non-residential buildings, in 985 there was no one in the household who met the sample requirements; and in 524 there was no one at home. In 2,019 households there was a person meeting the sample specifications. The responses were as follows: Total contacts, 2,019. Refusals, 367; Bad health, unable to answer, 39; Interrupted interviews, 13; Total non-responses, 419. Completed interviews: 1,600, 79.2 percent of total.
The profile of respondents was compared with official Goskomstat data, in contemporary Russia an approximation rather than a precise count of the population. The sample was weighted according to gender, age and education to match the population statistics of Goskomstat, and also weighted by voting in order to match the published election results.
New Russia Barometer 10: 17 June - 3 July 2001
The tenth NRB survey was organized by the Centre for the Study of Public
Policy with fieldwork by the Russian Centre for Public Opinion Research (VCIOM)
between 17 June-3 July 2001, a period of relative political calm. The universe
consisted of Federation residents age 18 and over excluding soldiers, convicts
and those without a fixed address.
Respondents were selected by a multi-stage, stratified sample in which the
basic units were all urban areas and rural districts. VCIOM stratified the
Russian Federation into 10 large regions by seven social, economic and infrastructure
attributes: North and North-West, Central, Central-Tchernozemie, North Caucasus,
Volgo-Viatka, Volga, Urals, West Siberia, East Siberia and the Far East. Geographical
dispersion is important to avoid concentrating interviews in a few major cities;
Moscow and St. Petersburg together have fewer people than rural Russia. Each
region was assigned interviews proportionate to its share of the population,
using Goskomstat population statistics for the year 2000.
At the second stage, each region was stratified according to its administrative
status, characteristics of population and whether it was an autonomous republic
of the Russian Federation. Moscow and St. Petersburg were also included independently
as strata. The total number of interviews within a region was determined according
to the proportion each contributed to the region's population. At the third
stage, rural districts and urban centres was selected with a probability proportionate
to their population. At the fourth stage each was stratified by population
characteristics, e.g. villages within a rural district or a ward or electoral
precinct within a city, for selection proportionate to population. A total
of 42 different subjects of the Russian Federation were included in the sample
and with them 33 oblast centres, 42 cities and towns, and 31 villages. Within
these 106 units, at the fifth stage 195 primary sampling units were selected,
producing an average of ten respondents per unit.
A total of 197 experienced interviewers were employed; the median interviewer
had taken part in more than 10 previous VCIOM surveys; was age about 40 and
had a tertiary education. Within each primary sampling unit, households were
selected by a random route method, in which the interviewer called at every
nth house from a random starting point. Face-to-face rather than telephone
interviewing is especially important in Russia, since less than half the population
has a telephone. Within the household, one individual was selected as a respondent
on the basis of a sex by age by education grid based on national statistical
office data. As people with less education are more likely to refuse to be
interviewed, controlling for education is specially important to avoid over-representation
of more educated respondents.
Altogether, 4,751 addresses were visited; 88 were non-residential buildings;
in 642 no one met sample requirements; in 767 no one was at home. In 3,254
households a resident met sample specifications: Total contacts: 3,254. Refusals,
1,060. Bad health, unable to answer, 160; Interrupted interviews, 34; Total
non-responses: 1,254. Completed interviews: 2,000. The profile of respondents
was compared with official Goskomstat data, in today's Russia an approximation
not a precise account of the population. The sample was weighted according
to gender, age and education.
New Russia Barometer XI: 12-26 June 2003
The eleventh NRB survey was organized by the Centre for the Study of Public
Policy with fieldwork by the Russian Centre for Public Opinion Research (VCIOM-Analytica)
between 12-26 June 2003, a period of relative political calm. The universe
consisted of the residents of the Russian Federation age 18 and over, excluding
soldiers, convicts and those without a fixed address.
Respondents were selected by a multi-stage, stratified sample in which the
basic units were all urban areas and rural districts. VCIOM-Analytica stratified
the Russian Federation according to the seven federal regions by social, economic
and infrastructure attributes: North-West, Central, South, Povolzhskii, Uralskii,
Siberia and the Far East. Each region was assigned a number of interviews
proportionate to its share of the Russian population, using Goskomstat population
statistics for the year 2002.
At the second stage, each region was stratified according to its administrative status, characteristics of population and whether it was an autonomous republic of the Russian Federation. Moscow, St. Petersburg and all cities with more than one million population were automatically included. The total number of interviews within a region was determined according to the proportion each contributed to the region's population. At that stage 100 cities, towns and rural areas were randomly selected with a probability proportionate to their population, and dispersed across 39 of the 89 subject territories of the Russian Federation. At the third stage each area was stratified by population characteristics, e.g. villages within a rural district or a ward or electoral precinct within a city, and primary sampling units randomly selected proportionate to population, to yield 160 primary sampling units with an average of 10 respondents per location. At the fourth stage, a location was chosen within each primary sampling unit as the starting point for interviewing, and households were then instructed to proceed by a random route to every nth house.
At each selected house, the interviewer sought a respondent whose characteristics
conformed to a gender by age by education grid, based on national statistical
office data. Within the household, one individual was selected as a respondent
on the basis of a sex by age by education grid based on national statistical
office data. Altogether, 4,052 addresses were visited; 56 were non-residential
buildings; in 986 addresses there was no one at home and in 915 no one appeared
to open the door. In 2,095 households a resident met the sample specifications,
and responded as follows: Refusals, 436; Bad health, unable to answer, 30;
Interrupted interviews, 28; Total completed interviews, 1,601 (76.4 percent).
The profile of respondents was compared with official Goskomstat data, in
today's Russia an approximation not a precise account of the population. The
sample was weighted according to gender, age and education to align it with
the population as defined by Goskomstat.
New Russia Barometer XII: 12-22 December 2003
The twelfth New Russia Barometer was conducted as part of the VCIOM-Analytica
post-Duma election survey between 12 and 22 December 2003. The universe for
the sample was the adult population of the Russian Federation age 18 and above,
excluding soldiers, convicts, the homeless and persons temporarily away from
home.
A multi-stage stratified sample was drawn. In the first stage the sample was
distributed among seven federal regions--Northwestern, Central, Southern,
Povolzhskii, Uralskii, Siberia, and Far East. Inside each region there was
then stratification among five types of settlement proportionate to number
of persons there age 18+ years. All cities with over 1 mn. population were
automatically included in the sample. There was then a further selection of
urban and rural settlements proportionate to population. A total of 100 locations
was then selected within these settlements, 66 urban and 34 rural, covering
39 of the 89 subjects (that is, administrative units) of the Russian Federation.
Within each of the hundred locations, one or more Primary Sampling Unit (PSU)
was then randomly selected, polling-districts in urban settlements or villages
in rural districts, from the list of all primary units according to their
proportion of the population. The object was to produce an average of 10 respondents
for each Primary Sampling Unit.
Selection of households in each PSU was by the random route method. Within
each household, only one person was eligible for interview. Selection was
made on the basis of an age by gender by education grid based on Goskomstat
statistics 2002. Respondents were interviewed in their home. The total number
of contacts: 2022. Non-respondents: Bad health, Unable to answer, 42; Refusals,
369; Interrupted interviews, 10; Completed interviews, 1601, or 79.2 percent
of individuals contacted.
The completed survey was weighted initially by gender, age and education
data from Goskomstat 2002. The weightings were defined as extreme values for
the sum of squares of deviations of weighted estimates from corresponding
precise values of proportions of social groups within the surveyed population.
A second stage of weighting was then conducted to take into account differences
between the official statistics of non-voting, which reflects inadequacies
in the electoral register as well as actual non-voting, and the reported non-voting
of respondents in the VCIOM-Analytica sample. The process also involved weighting
to reflect the official report of voting by parties in the list ballot.
New Russia Barometer XIII: 18-23 March 2004
The thirteenth NRB survey was organized by the Centre for the Study of Public Policy in collaboration with the Levada Centre, Moscow, staffed by former members of VCIOM, the All-Russian Centre for Public Opinion Research. Interviewing commenced within a week of the re-election of President Putin and all interviews were conducted between 18 and 23 March 2004. The universe consisted of the residents of the Russian Federation age 18 and over, excluding soldiers, convicts and those without a fixed address.
Respondents were selected by a multi-stage, stratified sample in which the basic units were all urban areas and rural districts. The Levada Centre stratified the Russian Federation according to the seven federal regions by social, economic and infrastructure attributes: North-West, Central, South, Povolzhskii, Uralskii, Siberia and the Far East. Geographical dispersion is important in Russia to avoid concentrating all interviews in a few major cities; Moscow and St. Petersburg together have fewer people than rural Russia. Each region was assigned a number of interviews proportionate to its share of the Russian population, using Goskomstat population statistics for the year 2002.
At the second stage, each region was stratified according to its administrative status, characteristics of population and whether it was an autonomous republic of the Russian Federation. Moscow, St. Petersburg and all cities with more than one million population were automatically included. The total number of interviews within a region was determined according to the proportion each contributed to the region's population: 100 cities, towns and rural areas were randomly selected with a probability proportionate to their population, and dispersed across 39 of the 89 subject territories of the Russian Federation. At the third stage each of these areas was stratified by population characteristics, e.g. villages within a rural district or a ward or electoral precinct within a city, and primary sampling units randomly selected proportionate to population, to yield 160 primary sampling units with an average of 10 respondents per location. At the fourth stage, a location was chosen within each primary sampling unit as the starting point for interviewing, and households were then instructed to proceed by a random route to every nth house to conduct face to face interviews.
At each selected house, the interviewer sought a respondent whose characteristics matched a gender by age by education grid based on data of the national statistical office. Within the household, one individual was selected as a respondent on the basis of a sex by age by education grid based on national statistical office data. As people with less education are more likely to refuse to be interviewed, controlling for education is specially important to avoid the over-representation of respondents with above-average education. Contacts were made with 2,130 individuals with characteristics fitting the sample design. Refusals, 482; Bad health, unable to answer, 42; Interrupted interviews, 4. Completed interviews, 1,602, 75.2 percent of individuals contacts. The average interviewer conducted ten interviews. Quality checks of questionnaires were undertaken by the Organization Department of the Levada Centre.
The profile of respondents was compared with official Goskomstat data, in today's Russia an approximation not a precise account of the population. The sample was weighted according to gender, age and education to align it with the population as defined by Goskomstat. In the second step weights were added for voting to take into account those favouring four named presidential candidates, other voters and non-voters.
New Russia Barometer XIV: 3-23 January 2005
The fourteenth NRB survey was organized by the Centre for the Study of Public
Policy in collaboration with the Levada Center, Moscow. Interviewing occurred
from 3 to 23 January 2005. The universe consisted of the residents of the
Russian Federation age 18 and over, excluding soldiers, convicts and those
without a fixed address.
Respondents were selected by a multi-stage, stratified sample in which the
basic units were all urban areas and rural districts. The Levada Center stratified
the Russian Federation according to the ten large economy-geographical regions
(North+North-West, Central, Central-Tchernozemie, North Caucasus, Volgo-Viatka,
Volga, Urals, West Siberia, East Siberia and Far East) of the Russian Federation.
Geographical dispersion is important in Russia to avoid concentrating all
interviews in a few major cities; Moscow and St. Petersburg together have
fewer people than rural Russia. Each region was assigned a number of interviews
proportionate to its share of the Russian population, using Goskomstat 2002
census statistics. Each region was stratified according to its administrative
status, characteristics of population and whether it was an autonomous republic
of the Russian Federation. Moscow, St. Petersburg and all cities with more
than one million population were automatically included. The total number
of interviews within a region was determined in proportion each contributed
to the region's population.
At the second stage 110 urban and rural places (35 regional centres, 39 towns
and 36 villages) belonging to 42 subjects of the Russian Federation) were
randomly selected with a probability proportionate to their population. At
the third stage each place was stratified by population characteristics, e.g.
villages within a rural district or a ward or electoral precinct within a
city, and primary sampling units randomly selected proportionate to population,
to yield Primary Sampling Units (PSU) with an average of 10 respondents per
Unit for a total of about 200 Primary Sampling Units. At the fourth stage,
a location was chosen within each PSU as the starting point for interviewing,
and households were then instructed to proceed by a random route to every
nth house.
Face-to-face interviewing is especially important in Russia, since upwards
of half the households is without a telephone. At each selected house, the
interviewer selected for interview the adult whose birthday came next. The
survey was conducted by means of self-completion of the questionnaire in the
presence of the interviewer in the respondent's home.
A total of 172 interviewers were employed, of which 31 percent had higher
education and 69 percent had secondary education. Five-sixths of the interviewers
had taken part in ten or more surveys.
The total number of addresses visited was 5,643. Of this total 87 were not
residential addresses, at 679 there was no one home and at 1,430 addresses
no one was willing to open the door. The total number of household contacts
was 3,447. Refusals, 1,151; Bad health, unable to answer, 169; Interrupted
interviews, 19, Total non-responses, 1,340; Completed interviews, 2,107. The
average interviewer conducted 12 interviews. Levada Center field supervisors
controlled one in five of the interviews, with 168 call backs at the respondent's
household, 234 cases checked by telephone and 30 by post. Quality checks of
questionnaires were undertaken by the Organization Department of the Center
in Moscow.
The profile of respondents was compared with official Goskomstat data, in
today's Russia an approximation not a precise account of the population. The
sample was initially weighted according to gender, age, education and reported
behaviour in the 2004 presidential election to align it with known national
statistics. In the second step weights were added for voting to take into
account those favouring four presidential candidates, other voters and non-voters.
New Russia Barometer XV: 13-23 April 2007
The fifteenth NRB survey was organized by the Centre for the Study of Public
Policy in collaboration with the Moscow-based Levada Center. Interviewing
occurred from 13 to 23 April 2007. The universe consisted of the residents
of the Russian Federation age 18 and over, excluding soldiers, convicts and
those without a fixed address.
Respondents were selected by a multi-stage stratified sample in which the
basic units were all urban areas and rural districts. The Russian Federation
was partitioned by its seven federal regions (North-Western, Central, Southern,
Volga, Urals, Siberia, and Far Eastern). Geographical dispersion is important
in Russia since Moscow and St. Petersburg together have fewer people than
rural Russia, and urban areas distant from Moscow are collectively more populous
too. Each region was assigned a number of interviews proportionate to its
share of the Russian population, using Goskomstat 2002 census statistics.
Then, each region was stratified according to its administrative status, characteristics
of population and whether it was an autonomous republic of the Russian Federation.
Moscow, St. Petersburg and all cities with more than one million population
were automatically included in the sample.
A total of 93 urban and 35 rural settlements belonging to 46 subjects of the
Russian Federation were randomly selected with a probability proportionate
to their population. Two or more Primary Sampling Units were selected at random
in each of the 128 urban settlements or rural villages. In Moscow there were
16 PSUs and in St Petersburg 9 PSUs, thus producing interviews in more than
275 randomly dispersed PSUs. Within each PSU an address was chosen as the
starting point for interviewing and, according to the area's type of housing,
interviewers were instructed to proceed by a random route to every 17th household
in blocks of flats or to every fifth house. Interviewing in the household
is especially important in Russia, since a substantial proportion of individuals
are still without a telephone. At each house in the sample, the individual
selected as the respondent was the adult whose birthday came next, subject
to controls for gender, age and educational level. The survey was conducted
by means of self-completion of the questionnaire in the presence of the interviewer
in the respondent's home.
Each household selected was visited up to three times on different days of
the week and at different times of day. The total number of addresses visited
was 3,989. Of this total, 43 were not residential addresses or unoccupied;
at 582 there was no one home after three calls and at 847 no one was willing
to open the door. The total number of household contacts was 2,517. The responses
were as follows: Refusals 807; Bad health, unable to answer 79; Interrupted
interviews 15. The total non-responses were 901. Ten interviews were rejected
during control and the completed interviews were 1,606.
Levada Center field supervisors controlled 28 percent of the interviews, with
75 call backs at the respondent's household and 380 respondents checked by
telephone. Quality checks of questionnaires were undertaken by the Organization
Department of the Center in Moscow.
The profile of respondents was compared with official Goskomstat data, in
today's Russia an approximation not a precise account of the population. The
sample was weighted according to gender, age, and education to align it with
national statistics.
New Russia Barometer XVI: 7-17 December 2007
Immediately after the Duma election on 2nd December the sixteenth NRB survey
was organized by the Centre for the Study of Public Policy in collaboration
with the Moscow-based Levada Center. Interviewing was between 7 and 17 December
2007. The universe consisted of the residents of the Russian Federation age
18 and over, excluding soldiers, convicts, persons temporarily visiting a
household sample, and those without a fixed address.
Respondents were selected by a multi-stage stratified sample in which the
basic units were all urban areas and rural districts. The Russian Federation
was partitioned by its seven federal regions (North-Western, Central, Southern,
Volga, Urals, Siberia, and Far Eastern). Geographical dispersion is important
in Russia since Moscow and St. Petersburg together have fewer people than
rural Russia, and urban areas distant from Moscow are collectively more populous
too. Within each region settlements were stratified into five categories in
proportion to their share of the population according to Goskomstat census
statistics. Then, each region was stratified according to its administrative
status, characteristics of population and whether it was an autonomous republic
of the Russian Federation. Moscow, St. Petersburg and all cities with more
than one million population were automatically included in the sample. A total
of 93 urban and 35 rural settlements belonging to 46 subjects of the Russian
Federation were randomly selected with a probability proportionate to their
population.
Within each of the 128 districts two or more Primary Sampling Units (PSU)
were selected at random; in Moscow there were 16 PSUs and in St Petersburg
9 PSUs. Thus, there were interviews in more than 275 randomly dispersed PSUs.
Within each PSU an address was chosen as the starting point for interviewing
and, according to the area's type of housing, interviewers were instructed
to proceed by a random route to every 17th household in blocks of flats or
to every fifth house,. Interviewing in the household is especially important
in Russia, since a substantial proportion of individuals are still without
a telephone. At each house in the sample, the individual selected for interviewing
was the adult whose birthday came next, subject to controls for gender, age
and educational level.
Each household selected in the sample was visited up to three times on different
days of the week and at different times of day. The total number of addresses
visited was 3,828. Of this total, 32 were not residential addresses or unoccupied;
at 935 there was no one home after three calls and at 831 no one was willing
to open the door. After allowing for these factors, the total number of household
contacts was 2,030. The responses were as follows: Refusals 397; Bad health,
unable to answer 23; Interrupted interviews 7. The total non-responses were
427. Two interviews were rejected during control, and the completed interviews
were 1,601.
Levada Center field supervisors controlled 21 percent of the interviews, using
138 call backs at the respondent's household and checking 304 respondents
by telephone. Quality checks of questionnaires were undertaken by the Organization
Department of the Center in Moscow.
The profile of respondents was compared with official 2002 Goskomstat data,
in today's Russia an approximation not a precise account of the population.
The sample was weighted according to gender, age, and education to align it
with national statistics.
New Russia Barometer XVII: 14-23 March 2008
Immediately after the presidential election on 2nd March the seventeenth
NRB survey was organized by the Centre for the Study of Public Policy in collaboration
with the Moscow-based Levada Center. Interviewing was between 14 and 23 March
2008. The universe consisted of the residents of the Russian Federation age
18 and over, excluding soldiers, convicts, persons temporarily visiting a
household sample, and those without a fixed address.
Respondents were selected by a multi-stage stratified sample in which the
basic units were all urban areas and rural districts. The Russian Federation
was partitioned by its seven federal regions (North-Western, Central, Southern,
Volga, Urals, Siberia, and Far Eastern). Geographical dispersion is important
in Russia since Moscow and St. Petersburg together have fewer people than
rural Russia, and urban areas distant from Moscow are collectively more populous
too. Within each region settlements were stratified into five categories in
proportion to their share of the population according to Goskomstat census
statistics. Then, each region was stratified according to its administrative
status, characteristics of population and whether it was an autonomous republic
of the Russian Federation. Moscow, St. Petersburg and all cities with more
than one million population were automatically included in the sample. A total
of 93 urban and 35 rural settlements belonging to 46 subjects of the Russian
Federation were randomly selected with a probability proportionate to their
population.
The 128 Sampling Units (PSU) included 16 PSUs in Moscow and 9 PSUs in St Petersburg.
Thus, there were interviews in more than 275 randomly dispersed PSUs. Within
each PSU an address was chosen as the starting point for interviewing and,
according to the area's type of housing, interviewers were instructed to proceed
by a random route to every 17th household in blocks of flats or to every fifth
house. Interviewing in the household is especially important in Russia, since
a substantial proportion of individuals are still without a telephone. At
each house in the sample the individual selected for interviewing was the
adult whose birthday came next, subject to controls for gender, age and educational
level.
To secure an interview, each selected respondent's household was visited up
to three times on different days of the week and at different times of day.
The total number of addresses visited was 3,739. Of this total, 49 were not
residential addresses or unoccupied; at 975 no one was home after three calls;
and at 761 no one was willing to open the door. After allowing for these factors,
the total number of household contacts was 1,951.
Levada Center field supervisors controlled 28 percent of the interviews, using
135 call backs at the respondent's household and checking 310 respondents
by telephone. Quality checks of questionnaires were undertaken by the Organization
Dept. of the Center in Moscow.
The profile of respondents was compared with official 2002 Goskomstat census
data, in today's Russia an approximation not a precise account of the population.
The sample was weighted according to gender, age, education, and voting behaviour
to align it with national statistics.
(Fuller statistical details of single NRB samples are given in the appendix of the SPP paper in which full results are reported.)
1. For technical reasons, some NRB samples have included all citizens age 15 upwards. When this has occurred, respondents under the age of 18 are excluded in order to make results represent the electorate.