Comparability — Argentina [top]

In the 1991 sample, roughly half the cases in the sample lack data for this and several other variables. Users should apply the alternative weight variable AR1991A_0434 when using one of these variables.

Comparability — Bangladesh [top]

All samples consistently refer to the person's main field of employment. The 2011 sample has only a few general categories.

Comparability — Belarus [top]

This variable is only available in 2009.

Comparability — Benin [top]

The 1979 and 1992 samples use the same industry classification but with differing detail. The later samples use different classifications. The age universe of respondents also differs between 1979-1992 and 2002-2013.

Comparability — Bolivia [top]

The reference period is consistent across samples, but the universe differs over time. The 1976 and 1992 samples include some or all unemployed persons, while the 2001-2012 samples do not. The 2012 sample uses the ISIC rev. 4 classification, the 1992-2001 samples use the ISIC rev. 3 classification, and the 1976 sample uses a distinct 4-digit classification.

Comparability — Botswana [top]

The coding structure is roughly comparable across the Botswana samples, but with different level of detail. The industry data have a 1-digit coding structure in 1981, 2-digits in 1991, and the same 4-digits classification in 2001 and 2011.

The reference period was the month prior to the census date for the 1981-1991 samples, and the week prior to the census date for the 2001- 2011 samples.

Comparability — Burkina Faso [top]

The data records the person's type of economic activities. Industry is not reported in the 1985 and 2006 samples.

Comparability — Cambodia [top]

The 1998-2004 samples use one classification, and the 2008-2013 samples use another.

Comparability — Cameroon [top]

Data on industry is not available for the 1976 and 1987 samples.

Comparability — Canada [top]

The universe and reference period (the week prior to the census) are identical across all samples.

The industry classification structure used in the Canada samples changed over time. The 1971 sample uses the 1970 Standard Industrial Classification, whereas the 1981- 2001 samples use the 1980 Standard Industrial Classification. Both are national classifications systems. The 2011 sample used the NAICS structure.

Comparability — China [top]

The samples use different national classification systems over time. The 1982-1990 samples ask about the current industry (at the time of census), while the 2000 sample refers to industry in the week prior to the census.

Comparability — Costa Rica [top]

The universe for 1973-1984 is persons age 12+ who ever worked, but in 2011 the universe becomes persons who were employed.

Comparability — Cuba [top]

The 2002 and 2012 samples use the same industry classification and have the same universe. The reference period for both samples is the week before the census.

Comparability — Côte d'Ivoire [top]

The universe is consistent across samples; however, the level of detail varies.

Comparability — Dominican Republic [top]

The data report the activity of the establishment or sector in which the person works or worked. While the universe is consistent across samples in terms of age, the reference period and employment categories included in the universe vary. Samples use 3 different classification systems: 1960, 1970-1981, and 2002-2010

Comparability — Egypt [top]

The universe statement is consistent across samples, but the classification systems differ. The industry data is classified using ISIC rev. 2 in 1986, ISIC rev. 3 in 1996, and ISIC rev. 4 in 2006.

Comparability — Fiji [top]

All samples refer to the person's industry during the week prior to the census. The classification differs for each sample. The universe statements also differ with regard to the respondent's age and employment status.

Comparability — Finland [top]

The 2010 sample uses a national classification, the Standard Industrial Classification TOL 2008.

Comparability — France [top]

There are differences between samples in the minimum age of respondents.

The classification systems differ across samples.

Comparability — Germany [top]

The samples use different classification systems.

Comparability — Ghana [top]

In both samples, the reference period is the 7 days preceding the census night. The age universe differs across samples, including persons age 7+ in 2000 and persons age 5+ in 2010. The broad industry categories are similar across samples, but the 2010 sample uses a much less detailed system.

Comparability — Greece [top]

The 1971-1981 samples use the same classification system, while the classification system is also common across 1991-2001. The 2011 sample follows the ISIC Rev. 4 coding structure.

The reference period is the week before the census for all years, except for 1981 that specify the last time the person worked.

The 2011 sample does not have an age universe, but previous samples only reported information for persons aged 10 or more.

Comparability — Guatemala [top]

Each sample uses a different classification. 2002 offers less detail than earlier years. The age universe differs across samples.

Comparability — Guinea [top]

The variable is only available for the 1983 and 2014 samples. The universe and industry classification differ across the two samples, although the reference period in both cases is the week prior to the census.

Comparability — Haiti [top]

The 1971 census asked the industry question, but the coded data are lacking many industries and are therefore not included in INDGEN. See the unharmonized variables.

Comparability — Honduras [top]

The industry classification differs across each sample. The 2001 sample uses the 1988 ISCO classification. The 1988 census asked an industry question, but the microdata lack labels.

Comparability — Hungary [top]

The universe and classification systems differ across samples. The industry data use ISIC rev. 3 in 2001 and ISIC rev. 4 in 2011. Responses refer to the reference week in 2001 and to the current industry (at the time of census) in 2011.

Comparability — Iran [top]

The classification systems differ across samples. The 2006 and 2011 data report a person's industry during the previous week.

Comparability — Ireland [top]

The Irish samples use separate classifications for 1971, 1981-1996, 2002-2006, and 2011-2016. The 2011 and 2016 samples use the ISIC rev. 4 classification, although the 2011 sample provides 3-digit codes while the 2016 sample provides 2-digit codes. There are minor universe differences across samples.

Comparability — Israel [top]

All samples use the same standard industrial classification, with only slight modifications over time. There is a small change in the age universe between the 1972 and 1983-2008 samples.

Comparability — Italy [top]

The censuses each provide a two-digit classification, but the coding scheme differs.

Comparability — Jamaica [top]

The data report a person's industry during the previous week.

The 1982 sample records the industry of everyone who ever worked. The latter samples limit the universe statement to persons currently in the labor force.

The 1991 and 2001 samples use the same classification. The 1982 sample uses a different classification.

Comparability — Laos [top]

Note that persons working in agriculture were not asked the industry question in the 2005 sample. See the source variable LA2005A_0418 (FARM2) to identify these workers. The question was asked to all employed persons age 10+ for the other samples.

Comparability — Liberia [top]

The universe differs across samples, including persons age 10+ in 1974, and persons age 6+ in 2008. Both samples included persons who worked in the last 12 months in the universe, although the response categories for economic activity were more detailed in 2008 and therefore explicitly included those who were paid employees, self-employed or contributing family workers. The samples use different levels of detail in their classification systems.

Comparability — Malawi [top]

The only difference in universe statement between samples is the change in age: from age 10 or older in the 1987-1998 samples to age 6 or older in the 2008 one.

The 1987 and 1998 samples use the same classification system, although the 1998 sample contains more detail. The 2008 sample uses a different classification system.

Comparability — Mali [top]

All samples report main industry during the month prior to the census. They use different codes, but record similar major industrial categories.

Comparability — Mauritius [top]

The industry classification is inconsistent across the 1990-2011 samples. The 2011 sample uses the NSIC rev. 2 classification. The 1990 and 2000 samples have various unharmonized source variables with different level of detail for industry codes. The universe is identical across the three samples, and the reference period in each sample is the week prior to the census.

Comparability — Mexico [top]

The coding structures differ across samples, except for the years 2010, 2015, and 2020 that use the same industry classification. The industry data are coded according to national classification systems.

Comparability — Morocco [top]

The universe is identical across the 1982-2014 samples and the industry classifications are mostly comparable, with only minor discrepancies. The 1994-2004 samples use a classification consistent with ISIC rev. 3, while the 2014 sample uses a classification that is consistent with ISIC rev. 4.

Comparability — Myanmar [top]

The 2014 sample uses the ISIC rev. 4 classification. The reference period in the 2014 sample is the past 12 months.

Comparability — Nepal [top]

The samples use different classifications.

Comparability — Pakistan [top]

The data record the person's industry during the week prior to the census. The 1981 and 1998 samples do not report industry information.

Comparability — Palestine [top]

There are minor universe differences across the 1997-2017 samples. The 1997 and 2007 samples use the ISIC rev. 3 classification, while the 2017 sample uses the ISIC rev. 4 classification. All samples record the person's main industry during the previous week.

Comparability — Papua New Guinea [top]

The 1980 census only addressed the question to urban respondents.

Comparability — Paraguay [top]

The universe and classifications change over samples. The 2002 sample uses a four-digit coding scheme, while all other samples use a 3-digit coding scheme.

Comparability — Peru [top]

The classification system is identical across samples. The 1993 sample records the person's industry during the last week or at the last time they were employed. The 2007 sample reports a person's industry during the week prior to the census. The 2017 sample reports the economic activity of persons age 5+ who worked within the last week or those who had a job, business, or worked on a farm or family business.

Comparability — Philippines [top]

The industry classification differs across each sample. The minimum age universe differs across samples.

Comparability — Poland [top]

The classification systems differ across samples.

The 1978 source data assigned industry also to non-workers based on who "supported" them, but IND includes only data for economically active persons.

Comparability — Puerto Rico [top]

All samples use the 1990 classification system. The samples report the person's industry during the last week or at the last time they were employed (up to five or 10 years prior to the census). See the specific universe statements.

Each sample also has alternative industry variables that are coded according to the classification used in the respective census years.

The universe for 1980-2005 and 2015-2020 is persons age 16+ who worked within the last five years, but the universe changes to all persons for the 2010 sample.

Comparability — Romania [top]

The classification system used and the level of disaggregation available differ across samples. The universes change over time but all samples refer, in broad terms, to the economically active population.

The 1977-1992 samples report the current industry (at the time of the census), while the 2002-2011 samples use the week prior to the census as reference period.

Comparability — Saint Lucia [top]

The data report the person's industry during the week prior to the census. The 1980 sample does not have industry information.

Comparability — Senegal [top]

The 1988 and 2013 samples record the person's main industry during the last year. The 2013 sample uses the ISIC rev. 4 classification, while the 1988 sample uses a classification that is not compatible with the ISIC structure. There are minor universe differences across the two samples. The 2002 sample lacks an industry variable.

Comparability — South Africa [top]

The 2001 and 2007 samples use the same 3-digit coding scheme. The 1996 sample, instead, uses a 2-digit coding scheme. The first two digits of this latter scheme are comparable to those of the other South Africa samples.

Comparability — Spain [top]

Retirees are included in the 1991 universe and excluded from the 1981, 2001, and 2011 samples. Therefore, there are significantly more persons with industry responses in 1991 than in the other three years. The classification scheme changes across samples.

Comparability — Suriname [top]

The 2004 sample uses a coding structure similar to the ISIC rev. 3.1 classification, while the 2012 sample uses the ISIC rev. 4 classification. The age universe differs across the two samples. The reference period in both cases is the week prior to the census.

Comparability — Switzerland [top]

The samples are consistent over time and report current industry.

Comparability — Tanzania [top]

This variable in the 2002 and 2012 samples record a person's main industry category. It applies to persons who have worked in their current industry for at least 7 days in the time before the census. The 1988 Tanzania sample does not include an industry variable.

Comparability — Thailand [top]

All samples report the person's main industry during the last year. The universe statement for age varies slightly between the 1970-1980 samples and the 1990-2000 ones.

Comparability — Togo [top]

The samples use different classifications and have different universes. The 2010 sample provides much more detail.

Comparability — Trinidad and Tobago [top]

The Trinidad and Tobago samples follow the ISIC rev. 2 coding structure. The reference period is the week prior to the census.

Comparability — Uganda [top]

Data on industry is only available for the 2002 sample, as the 1991 and 2014 samples lack an industry variable.

Comparability — United Kingdom [top]

The 1961 sample reports the industry of persons age 15+ who are employed, while the 1971 includes persons age 15+ who were ever employed.

Comparability — United States [top]

The classifications are consistent for 1960-1970, 1980-1990, 2000-2005, and 2010-2020, with some minor variations in availability of specific categories. The universe is consistent across years.

Comparability — Uruguay [top]

The universe differs across samples. In the 1963 sample, this question was answered by persons aged 8 years or older who were in the labor force; when the person was inexperienced unemployed, the industry indicated was coded with the person's profession. For the 1985 and 1996 samples, the industry was reported by persons aged 12 years or older who were employed or experienced unemployed. In 2006, the industry was reported by persons aged 14 or older who were employed. Industry is not available for the 1975 sample.

Comparability — Vietnam [top]

The universe varies considerably across samples. The 1989 sample includes persons employed at any point during the previous year. The 1999 sample includes persons employed for six months or more during the previous year. The 2009 sample includes persons who are currently employed. The 2019 sample includes persons who did paid work in the last 7 days or have a job to return to within 30 days.

The samples use different classification systems.

Comparability — Zambia [top]

The universe is consistent across samples, except exclusion of absent household members in the 2010 sample. The 1990 and 2000 samples use the same industry classifications, but the 2010 sample uses a different classification system.