PERWT is an 8-digit numeric variable with 2 implied decimal places. See the variable description.

## Description

PERWT indicates the number of persons in the actual population represented by the person in the sample.

For the samples that are truly weighted (see the comparability discussion), PERWT must be used to yield accurate statistics for the population.

NOTE: PERWT has 2 implied decimal places. That is, the last two digits of the eight-digit variable are decimal digits, but there is no actual decimal in the data.

## Comparability — Index

## Comparability — General

PERWT is available for all samples, but only some samples have weights that differ among persons (i.e., many of the samples are "flat" samples in which the use of weights is optional).

The samples for which weights must be used to generate representative statistics are:

Brazil 1970, 1980, 1991, 2000, 2010

Canada 2001, 2011

Colombia 1985, 2005

Dominican Republic 1970, 1981

Ecuador 1974

Egypt 1986, 1996

Ethiopia 1994, 2007

France 2006, 2011

India 1983, 1987, 1993, 1999, 2004, 2009

Indonesia 1971, 1976, 1980, 1985, 1990, 1995, 2005

Iran 2006, 2011

Jamaica 2001

Kenya 1969, 1979

Mexico 1995, 2000, 2010, 2015

Mongolia 1989

Mozambique 1997

Netherlands 1960, 1971, 2001

Nigeria 2006, 2007, 2008, 2009, 2010

Pakistan 1973, 1981

Palestine 1997, 2007

Panama 1980

Paraguay 2002

Philippines 1990, 2000

Puerto Rico 1990, 2000, 2005, 2010

South Africa 1996, 2001, 2007, 2011

South Sudan 2008

Spain 1981, 1991, 2011

Sudan 2008

Tanzania 1988, 2002, 2012

Thailand 1970, 1980, 1990, 2000

Uganda 1991

United States 1990, 2000, 2005, 2010

Uruguay 2006

Venezuela 1971, 1990

Vietnam 1989, 1999, 2009

For the other samples, a weight was calculated from the sampling fraction. In those samples PERWT may not yield a precise count of the actual population to the extent that the theoretical sampling fraction was imperfect.

NOTE: The 1991 Argentina sample has an alternate weight variable that must be used to create accurate statistics for some variables. See the Argentina comparability discussion.

NOTE: Egypt 1986 is a recovered sample with missing geographic areas. Weights were constructed based on village and district population totals. Detailed notes about missing areas can be found in the geographic variables for Egypt.

NOTE: Ethiopia samples of 1984 and 2007 have special weight issues. In 1984, parts of the country are missing; weights do not correct for the missing geographic areas. The 2007 Ethiopia sample includes records from both the long and short forms, but only long form records are weighted (short form records are weighted zero). See the Ethiopia comparability section.

Some samples use fractional weights (to two decimal places), while most use only whole numbers (the first 6 digits of PERWT).

NOTE: Some samples from early years are missing geographic areas. These samples include Brazil 1960, Ethiopia 1984, Haiti 1982, Indonesia 1995, Indonesia 2005, Iraq 1997, Malaysia 1970 and 1980, Pakistan 1973, and Romania 1977.

NOTE: Mexico 1995, 2000, 2010, and Mongolia 1989 have clustered sample designs.

## Comparability — Argentina [top]

The 1980 and 1991 samples are weighted; the 1970 and 2001 samples are flat.

Several important variables in 1991 require the use of an alternative weight: the unharmonized variable AR1991A_0434. That alternative weight -- not PERWT -- must be used to generate accurate statistics for industry, occupation, previous residence and birthplace. Those variables have missing responses for roughly half the population, and AR1991A_0434 corrects for this.

## Comparability — Armenia [top]

The samples are flat: every record represents 10 persons in the population.

## Comparability — Austria [top]

The samples are flat: every record represents 10 persons in the population.

## Comparability — Belarus [top]

The samples are flat: every record represents ten persons in the population.

## Comparability — Bolivia [top]

All samples are flat: every record represents 10 persons in the population.

## Comparability — Botswana [top]

The samples are flat: every record represents ten persons in the population.

## Comparability — Brazil [top]

In 1960, the weight variable was missing from the microdata. The IPUMS weights are therefore theoretical weights based on the census design: 25 percent of households were to receive the long-form questionnaire from which the microdata sample is derived. All persons receive an equal weight corresponding to the overall sample density for the country in 1960.

In all other years, the weight varies among individuals and within households.

## Comparability — Cambodia [top]

The samples are flat: every record represents 10 persons in the population.

## Comparability — Canada [top]

The 2001 and 2011 samples are weighted; the others are flat. The 2001 and 2009 samples use fractional weights.

## Comparability — Chile [top]

All samples are flat: every record represents 100 (1960) or 10 (1970-2002) persons in the population.

## Comparability — China [top]

The samples are flat: every record represents 100 persons in the population.

## Comparability — Colombia [top]

The 1985 and 2005 samples are weighted; the others are flat. The 2005 sample uses fractional weights.

## Comparability — Costa Rica [top]

All samples are flat: every record represents an equal number of persons in the population (16.7 or 10, depending on the year).

## Comparability — Cuba [top]

The sample is flat: every record represents 10 persons in the population.

## Comparability — Dominican Republic [top]

Samples for 1960, 2002, and 2010 are flat: every record represents an equal number of persons in the population. Samples for 1970 and 1982 are differentially weighted.

## Comparability — Ecuador [top]

The 1974 sample uses fractional weights. The other samples are flat: every person represents an equal number of persons in the population.

## Comparability — Egypt [top]

The 1986 sample is weighted and uses fractions. This is a partially recovered data file; many geographic areas of the country are missing in part or in whole. Weights reflect village totals for villages where available. For details about missing geographic areas, see the geography variable documentation for Egypt.

The 1996 sample is weighted and uses fractions; the 2006 sample is flat.

## Comparability — El Salvador [top]

The samples are flat: every record represents 10 persons in the population.

## Comparability — Ethiopia [top]

The 1984 sample from Ethiopia is not representative of the country-level population. The file is a 1 in 10 sample of enumerated households. However, due to famine and conflict, parts of the country are missing from the sample. Detailed documentation about the nature of the undercount and affected geographic areas is available in a report.

Regions were sampled at different rates in 1994. Use of weights is especially important since Somali and Affar regions were sampled at lower densities.

The sample from 2007 includes records from both the long and short forms. When combined, the total set of records constitutes a flat 10% sample, but few variables are available. The weight applies to long form respondents only (short form respondents have a weight of zero). Use of the weight is essential for most analyses and yields point estimates that correspond closely to published reports for all variables.

## Comparability — France [top]

The 1962-1999 samples are flat: every record represents an equal number of persons in the population. The 2006 and 2010 samples are weighted and use fractions.

## Comparability — Germany [top]

All samples are flat: every record represents an equal number of persons in the population.

## Comparability — Ghana [top]

The samples are flat: every record represents 10 persons in the population.

## Comparability — Greece [top]

All samples are flat: every record represents an equal number of persons in the population.

## Comparability — Guinea [top]

All samples are flat: every record represents 10 persons in the population.

## Comparability — Hungary [top]

All samples are flat: every record represents an equal number of persons in the population.

## Comparability — India [top]

All samples are weighted. The 1993 and later samples use fractional weights. The weighted population count in all years is smaller -- in some years significantly smaller -- than the official census population counts.

## Comparability — Indonesia [top]

The samples for 1971-1995 and 2005 are weighted. The 2000 and 2010 samples are flat: every record represents 20 persons in the population.

## Comparability — Iran [top]

The samples are weighted and use fractions.

## Comparability — Ireland [top]

All samples are flat: each record represents the same number of persons in the population.

## Comparability — Israel [top]

All samples are flat: every record represents 10 persons in the population.

## Comparability — Italy [top]

The sample is flat: every record represents 20 persons in the population.

## Comparability — Jamaica [top]

The 1982 and 1991 samples are flat; the 2001 sample is weighted.

## Comparability — Jordan [top]

The sample is flat: every record represents 10 persons in the population.

## Comparability — Kenya [top]

The 1969 and 1979 samples are weighted. The later samples are flat: every record represents 10 or 20 persons in the population.

## Comparability — Kyrgyz Republic [top]

The sample is flat: every record represents 10 persons in the population.

## Comparability — Liberia [top]

Both samples are flat: every record represents the same number of persons in the population.

## Comparability — Malawi [top]

The samples are flat: every record represents 10 persons in the population.

## Comparability — Malaysia [top]

The samples are flat: every record represents 50 persons in the population.

## Comparability — Mali [top]

All samples are flat: every record represents 10 persons in the population.

## Comparability — Mexico [top]

The 1995, 2000, 2010, and 2015 samples are weighted; the others are flat. The state of Chiapas lacked person weights in the 1995 sample; we substituted the dwelling weights for that state.

## Comparability — Mongolia [top]

The 2000 sample is flat; the 1989 sample is weighted. The 1989 sample design was highly clustered. All provinces are covered, but within provinces entire districts were included in the sample, while other districts were not sampled at all. We have weighted the data at the province level using published data on age structure, but for some applications the data may not be entirely representative of the population.

## Comparability — Morocco [top]

The samples are flat: every record represents 20 persons in the population.

## Comparability — Mozambique [top]

The 1997 sample is differentially weighted. The 2007 sample is flat: every record represents 10 persons in the population.

## Comparability — Netherlands [top]

The samples are weighted and use fractions.

## Comparability — Nicaragua [top]

The samples are flat: every record represents 10 persons in the population.

## Comparability — Nigeria [top]

The samples are differentially weighted.

## Comparability — Pakistan [top]

The 1973 and 1981 samples are weighted; the 1998 sample is not. The 1973 weight is a crude inflator based on province and urban-rural status.

## Comparability — Palestine [top]

The Palestinian samples are weighted and use fractions.

## Comparability — Panama [top]

The 1980 sample is weighted; the others are flat.

## Comparability — Paraguay [top]

The 1962, 1972, 1982, and 1992 samples are flat: every record represents 10 or 20 persons in the population depending on the sample year. The 2002 sample is differentially weighted and uses fractions.

## Comparability — Peru [top]

Both samples are flat: every record represents 10 persons in the population.

## Comparability — Philippines [top]

The 1990 and 2000 samples are weighted; the 1995 sample is flat.

## Comparability — Poland [top]

The samples are flat: every record represents approximately 10 persons in the population.

## Comparability — Portugal [top]

All samples are flat: every record represents 20 persons in the population.

## Comparability — Puerto Rico [top]

The 1990-2010 samples are weighted; the earlier samples are flat.

## Comparability — Romania [top]

All samples are flat: every record represents 10 persons in the population.

## Comparability — Rwanda [top]

The samples are flat: every record represents 10 persons in the population.

## Comparability — Saint Lucia [top]

Both samples are flat: every record represents 10 persons in the population.

## Comparability — Senegal [top]

The samples are flat: every record represents 10 persons in the population.

## Comparability — Sierra Leone [top]

The sample is flat: every record represents 10 persons in the population.

## Comparability — Slovenia [top]

The sample is flat: every record represents 10 persons in the population.

## Comparability — South Africa [top]

All samples are weighted and use fractions. The 2007 sample has multiple weights; we used the adjusted weight.

## Comparability — South Sudan [top]

The sample is weighted and uses fractions.

## Comparability — Spain [top]

The 1981, 1991, and 2011 samples use fractional weights. The 2001 sample is flat.

Several provinces in 1981 were sampled at roughly 5 times the density of other provinces, making the use of weights particularly critical in that year. The over-sampled provinces are Alava, Guipuzcoa, Navarra, and Vizcaya.

## Comparability — Sudan [top]

The sample is weighted and uses fractions.

## Comparability — Switzerland [top]

All samples are flat: every record represents 20 persons in the population.

## Comparability — Tanzania [top]

All Tanzania samples are weighted and use fractions.

## Comparability — Thailand [top]

All samples are weighted, reflecting a stratified sample design for each census.

## Comparability — Trinidad and Tobago [top]

The samples are flat: every record represents 10 persons in the population.

## Comparability — Turkey [top]

The samples are flat: every record represents 20 persons in the population.

## Comparability — Uganda [top]

The 1991 sample is weighted. The 1991 person weight data were flawed; the household weights have been substituted for them instead. The 2002 sample is flat.

## Comparability — Ukraine [top]

The samples are flat: every record represents an equal number of persons in the population.

## Comparability — United Kingdom [top]

The samples are flat: every record represents an equal number of persons in the population.

## Comparability — United States [top]

The 1990-2010 samples are weighted; the earlier samples are flat.

Alternative weights, such as replicate weights or sample line weights, are available in the source variables for the 2010 sample.

## Comparability — Uruguay [top]

The 2006 sample is weighted. The 1963-1996 and 2011 samples are flat, with each record representing 10 persons in the population.

## Comparability — Venezuela [top]

The 1971 and 1990 samples are weighted; the 1981 and 2001 samples are flat.

## Comparability — Vietnam [top]

All samples are weighted. In 1989, men and women have separate weights. The weights are identical for all persons of that sex within a given enumeration area.

## Comparability — Zambia [top]

The samples are flat: every record represents 10 persons in the population.

## Universe

- All persons

## Availability

- Argentina: 1970, 1980, 1991, 2001, 2010
- Armenia: 2001, 2011
- Austria: 1971, 1981, 1991, 2001, 2011
- Bangladesh: 1991, 2001, 2011
- Belarus: 1999, 2009
- Bolivia: 1976, 1992, 2001
- Botswana: 1981, 1991, 2001, 2011
- Brazil: 1960, 1970, 1980, 1991, 2000, 2010
- Burkina Faso: 1985, 1996, 2006
- Cambodia: 1998, 2008
- Cameroon: 1976, 1987, 2005
- Canada: 1971, 1981, 1991, 2001, 2011
- Chile: 1960, 1970, 1982, 1992, 2002
- China: 1982, 1990, 2000
- Colombia: 1964, 1973, 1985, 1993, 2005
- Costa Rica: 1963, 1973, 1984, 2000, 2011
- Cuba: 2002
- Dominican Republic: 1960, 1970, 1981, 2002, 2010
- Ecuador: 1962, 1974, 1982, 1990, 2001, 2010
- Egypt: 1986, 1996, 2006
- El Salvador: 1992, 2007
- Ethiopia: 1984, 1994, 2007
- Fiji: 1966, 1976, 1986, 1996, 2007
- France: 1962, 1968, 1975, 1982, 1990, 1999, 2006, 2011
- Germany: 1970-1971, 1981, 1987
- Ghana: 1984, 2000, 2010
- Greece: 1971, 1981, 1991, 2001, 2011
- Guinea: 1983, 1996
- Haiti: 1971, 1982, 2003
- Hungary: 1970, 1980, 1990, 2001, 2011
- India: 1983, 1987, 1993, 1999, 2004, 2009
- Indonesia: 1971, 1976, 1980, 1985, 1990, 1995, 2000, 2005, 2010
- Iran: 2006, 2011
- Iraq: 1997
- Ireland: 1971, 1979, 1981, 1986, 1991, 1996, 2002, 2006, 2011
- Israel: 1972, 1983, 1995
- Italy: 2001
- Jamaica: 1982, 1991, 2001
- Jordan: 2004
- Kenya: 1969, 1979, 1989, 1999, 2009
- Kyrgyz Republic: 1999, 2009
- Liberia: 1974, 2008
- Malawi: 1987, 1998, 2008
- Malaysia: 1970, 1980, 1991, 2000
- Mali: 1987, 1998, 2009
- Mexico: 1960, 1970, 1990, 1995, 2000, 2005, 2010, 2015
- Mongolia: 1989, 2000
- Morocco: 1982, 1994, 2004
- Mozambique: 1997, 2007
- Netherlands: 1960, 1971, 2001
- Nicaragua: 1971, 1995, 2005
- Nigeria: 2006-2010
- Pakistan: 1973, 1981, 1998
- Palestine: 1997, 2007
- Panama: 1960, 1970, 1980, 1990, 2000, 2010
- Paraguay: 1962, 1972, 1982, 1992, 2002
- Peru: 1993, 2007
- Philippines: 1990, 1995, 2000
- Poland: 1978, 1988, 2002, 2011
- Portugal: 1981, 1991, 2001, 2011
- Puerto Rico: 1970, 1980, 1990, 2000, 2005, 2010
- Romania: 1977, 1992, 2002, 2011
- Rwanda: 1991, 2002
- Saint Lucia: 1980, 1991
- Senegal: 1988, 2002
- Sierra Leone: 2004
- Slovenia: 2002
- South Africa: 1996, 2001, 2007, 2011
- South Sudan: 2008
- Spain: 1981, 1991, 2001, 2011
- Sudan: 2008
- Switzerland: 1970, 1980, 1990, 2000
- Tanzania: 1988, 2002, 2012
- Thailand: 1970, 1980, 1990, 2000
- Trinidad and Tobago: 1970, 1980, 1990, 2000, 2011
- Turkey: 1985, 1990, 2000
- Uganda: 1991, 2002
- Ukraine: 2001
- United Kingdom: 1991, 2001
- United States: 1960, 1970, 1980, 1990, 2000, 2005, 2010
- Uruguay: 1963, 1975, 1985, 1996, 2006, 2011
- Venezuela: 1971, 1981, 1990, 2001
- Vietnam: 1989, 1999, 2009
- Zambia: 1990, 2000, 2010