Comparability — Argentina [top]

The age universe differs between samples.

The classification of wage workers changes from one based on white and blue-collar (manual versus non-manual) in 1970 to one based on public versus private employment in later years.

Comparability — Armenia [top]

The census grouped all persons working in a farm household together. CLASSWK codes the first person with such a response as an own-account agricultural worker; subsequent persons are coded to "unpaid family worker."

Comparability — Austria [top]

The data are comparable. The 2001 sample separately identifies government workers.

The source data for Austria assigned class-of-worker information to non-workers based on who "supported" them. CLASSWK reports data only for economically active persons.

Comparability — Bangladesh [top]

The data are comparable across samples, with the only difference being the identification of servants in 2001.

Comparability — Belarus [top]

The 1999 sample includes additional categories that are grouped in "other" in 2009.

Comparability — Benin [top]

The samples are comparable across years with slight difference in specified categories in 1979. The age universe differs between 1979-1992 and 2002-2013.

Comparability — Bolivia [top]

The universe is similar across samples. There are minor discrepancies in the level of detail used to classify wage/salary workers, but the categories are otherwise comparable across samples.

Comparability — Botswana [top]

Self-employed workers are separated into those with or without employees in 2001 and 2011, so it is possible to distinguish employers from those working on their own account. However, the 1991 sample does not identify whether the self-employed have employees or not, and these cases have been classified only as "working on own account."

The 2001 and 2011 samples differentiate between employees paid in kind and those paid in cash. Payment in kind may be made for an exchange of goods or services for work performed. Employees paid in kind were categorized under "other wage and salary" in CLASSWK.

Comparability — Brazil [top]

The available categories for Brazil change over time as different things were emphasized. The 1960 and 1970 censuses used similar classifications; 1980 added detail on sharecroppers and migrant workers with and without legal contracts; 1991 added distinctions for domestic workers and public sector workers and retained only a single migrant worker category; and 2000 differentiated types of unpaid work and dropped categories for sharecroppers, migrants and domestics. Some years distinguished between public and private sector employees. The 2000 and 2010 samples also distinguish between workers with and without legal contracts.

The treatment of sharecroppers presents a noteworthy inconsistency for Brazil. Sharecroppers are coded among self-employed workers in 1960 and 1970. In those years, sharecroppers with employees were combined with other "employers". In 1980, sharecroppers with employees are identified separately as a subgroup of employers. The remaining sharecroppers in 1980 are differentiated between those with autonomy (self-employed) and those without autonomy ("employees") in production or freedom to cede their contract (see the enumeration instructions). Both of these latter types of sharecroppers are coded in the IPUMS within the larger "self-employed" grouping in 1980 and 1991. The 2000 census did not distinguish sharecroppers from other types of workers. The 2000 census did introduce a separate category for subsistence workers producing for their own consumption. The 2010 census no longer included categories for sharecroppers or subsistence workers, and provided no information on how they should be categorized in the enumeration instructions.

Comparability — Burkina Faso [top]

Categories for this variable are comparable across samples, but there are age differences in the universe of respondents.

Comparability — Cambodia [top]

The universe is comparable except for the age restrictions imposed in 2008 and 2019.

Comparability — Cameroon [top]

The data are largely comparable across samples, but there are differences in the universe of respondents with respect to household type, age, and residential status.

The 1976 sample does not distinguish paid from unpaid apprentices, and they are coded to unpaid in this variable.

Comparability — Canada [top]

Unpaid family workers are combined with wage and salary workers in 1991-2001. The unharmonized source variables for all years distinguish between incorporated and unincorporated self-employed workers.

The universes are comparable across samples. The census question refers to those who worked in either the previous or current calendar year.

Comparability — Chile [top]

The universe differs across samples both in terms of minimum ages and with respect to employment status. The 1982 sample, which directed the question to all persons 15 and older, is the most different from other years.

The 1960-1982 samples identify white and blue collar wage-salary workers, who were combined in 1992-2002.

Comparability — Colombia [top]

The Colombian censuses do not pose problems in terms of universe comparability. The 1993 census combined white and blue collar employees, while in the two previous censuses they were separated.

The 2005 sample has a broader universe than the other sample years.

Comparability — Costa Rica [top]

The codes are fully comparable across samples. Unemployed persons were asked the question in 1973 and 1984, but only employed persons were in the universe for 2000 and 2011.

Comparability — Côte d'Ivoire [top]

Unemployed persons were asked the census question in 1988, but only employed persons were in the universe for 1998.

Comparability — Dominican Republic [top]

The universe changes over time. In 1960, the question was asked of persons age 10 and older who were employed between January 1st and August 7th, 1960. In 1970, the question was asked of persons age 10 and older who were employed or unemployed the week prior to the census. In 1981-2010, the question was asked of persons age 10 and older who had ever worked. Respondents who had worked before but were not working during the week prior to the census reported their status in their last job.

The 1981 sample distinguishes between public sector, private sector, and domestic employees. The other samples do not make this distinction.

Comparability — Ecuador [top]

The universe of persons for the question changes over time, with the minimum age for the question dropping from 12 in 1962-1974, to 8 in 1990, to age 5 in 2001 and 2010.

The samples offer differing categorization of wage-salary workers. The 2010 sample introduced a separate category of "partner"; this was combined with employers for the integrated variable. See the unharmonized source variable to recover this information.

Comparability — Egypt [top]

The data are comparable. There is a substantial drop in reported self-employment in 2006 with respect to 1986 and 1996.

Comparability — El Salvador [top]

The 2007 sample includes experienced unemployed persons in the universe. Categories differ slightly between censuses.

Comparability — Fiji [top]

Response categories differ between samples as do the age universes. The 1976 response, "villager", is interpreted as subsistence worker.

The 1986 and 1996 data do not distinguish self-employed workers from day laborers; thus, those samples are not included in CLASSWK. See the relevant unharmonized variables for those samples to make other distinctions, such as identifying unpaid workers.

Comparability — Finland [top]

The 2010 sample only distinguishes between wage-salary and self-employed workers, and does not contain any additional detail. No information is provided about unpaid labor.

Comparability — France [top]

There are no major problems of comparability in terms of universe and category availability, although 1999 differs somewhat. The 2006 sample has more detail on temporary employment than previous samples; 2011 offers a basic classification.

The "public companies" includes Social Security workers. Government workers in 1999 were classified somewhat differently than earlier samples, but have been put into the earlier categories as best as possible to maintain consistency. The full original detail is available in the unharmonized source variable.

The age universe differs slightly across samples. All samples exclude soldiers (non-civilians) except for 1999, when soldiers are mostly classified into the civil servant category.

Comparability — Germany [top]

The universe includes persons currently employed but is slightly modified with regards to exclusion of students and apprentices across years.

The level of detail differs across samples.

The age universe differs slightly across samples. It includes persons age 15 and older in 1970 and 1987 samples, but persons age 14 and older in the intervening samples.

Comparability — Ghana [top]

The universe varies across samples for young children.

Comparability — Greece [top]

The age universe changed in 2011, but all samples are targeting person who are economically active or seeking a job.

Comparability — Guatemala [top]

The age universe differs across samples. Category availability differs between 1964-1973 and later years.

The 1994 and 2002 samples distinguish between self-employed workers with a workplace and those without one, such as street vendors. The detail is retained in the source variables.

Comparability — Guinea [top]

The age universe varies between the 1983 and 1996-2014 samples. The categories are generally comparable across samples, although the 1996 and 2014 samples provide more detail about wage and salary workers than the 1983 sample.

Comparability — Haiti [top]

Response categories are largely comparable across Haiti samples, although the 2003 sample incudes a separate category for apprentices. The universes differ across samples with respect to age; the 1971 census included persons age 5+, while the later census included persons age 10+. All samples limit the universe to persons who worked during the reference period, but each sample has a substantially different reference period.

Comparability — Honduras [top]

The 1974 sample offers less category detail than other years. The age universe for 2001 and 2013 is lower than earlier years.

Comparability — Hungary [top]

The 2001 and 2011 census had a response for working members of non-cooperative partnerships. This is coded among "own-account" workers in CLASSWK and results in a higher proportion in that category than in other samples. It is unclear where they would have been coded in earlier censuses.

The 2001 census reports a much smaller labor force than the earlier samples, with a larger NIU category.

Comparability — Indonesia [top]

The universe is mostly consistent across samples, with the following differences: it includes the experienced unemployed for 1971, excludes those who had a job but did not work for 1976, and the minimum age is lowered to 5+ for 2000.

From the 1980 sample onward, the data separate own-account workers into persons with temporary or unpaid help and those without. For the 2005 and 2010 samples, the data also separate employees into regular wage-earners ("permanent job") and casual employees, defined as those that work for multiple employers for daily contracts/wages ("occasional/temporary job").

Comparability — Ireland [top]

Samples are fully comparable with the exception of 1971. In 1971, the question was asked of persons age 14 and older and excluded retired or unemployed persons, but in later years included persons age 15 and older who ever worked. Beginning in 2002, the census question excluded absent persons. CLASSWK is not available for the 1979 sample.

Comparability — Israel [top]

The 1995 sample combines salaried workers and members of cooperatives, while the 1972-1983 and 2008 samples classify them separately.

All Israeli samples identify Kibbutz members.

Comparability — Italy [top]

For both the 2001 and 2011 samples, the "employer" category is constructed using a separate variable on whether the person has employees.

Comparability — Jamaica [top]

In all samples, the universe includes persons age 14 + who worked or had a job in the week prior to the census.

The "other" category includes foreign consultants resident and working in Jamaica, but whose class of worker is unknown.

Comparability — Kyrgyzstan [top]

The 1999 census provided more detailed response categories than the 2009 census, both in types of self-employment and in identifying unpaid workers. Some additional details in 1999 are available in the source variable for that sample.

Comparability — Lesotho [top]

The samples are comparable across years.

Comparability — Liberia [top]

The universe changes over time. In 1974, the question was asked of persons age 10 or older who worked in the past year. In 2008, the question was asked of persons age 6 and older who worked in the past year.

Comparability — Malawi [top]

For all Malawi samples, class of worker is consistently interpreted as position in employment the week preceding the census. The question excludes visitors and includes persons age 6+ in 2008, but includes persons age 10+ in 1987, 1998, and 2018.

In 1987 and 1998, class of worker is obtained from a question on the person's activity the week preceding the census. The 2008 and 2018 samples include a separate question about the class of worker. In 2008, people working at a family farm or business are coded as "own account, agriculture" for comparability with earlier samples.

Comparability — Malaysia [top]

The samples are comparable over time.

Comparability — Mali [top]

The samples differ in their categorization of employees. Apprentices are a single category in 1987 and 2009, but are differentiated into paid and unpaid workers in 1998. The 1987 apprentices are coded as unpaid apprentices, where the majority of apprentices are in 1998. The 2009 enumeration instructions specify that the apprentice category refers only to unpaid apprentices.

Comparability — Mauritius [top]

The universes and categories are fully comparable across samples, with the exception of an additional category for members of cooperatives in the 2011 sample.

All employees in the 1990-2011 samples are classified in the general category for wage and salary workers, regardless of whether they are paid by the month, fortnight, week, day, or job. These distinctions are retained in the corresponding unharmonized source variables.

Paid and unpaid apprentices/trainees are grouped together in the 1990-2011 samples, and are classified as "apprentice, unpaid or unspecified" (even though some of these workers may be remunerated).

Comparability — Mexico [top]

The universe for CLASSWK varies considerably among the Mexican samples. The 1960 census did not specify a population to be enumerated; rather, any person whose situation did not match with any categories presented was assumed to be not working and/or inexperienced unemployed and not in the universe. In 1970 the question was asked of all persons age 12 and older who worked at least one month during 1969, regardless of whether they had a job at the time of the census. This is an important difference from the 1990-2000 and 2015-2020 censuses, which applied the question to persons employed at the time of census, and from the 2010 census which applied to all economically active persons.

Census documentation specifies that the unpaid workers are family workers from 1960-1990, but explicitly includes non-family members from 1995-2015. The 2010, 2015, and 2020 censuses include a response for paid "assistants/helpers" -- the definition of these cases is unclear and has been grouped with "white or blue collar" employees.

Comparability — Morocco [top]

The universes are nearly identical, with the exception in 1982 when persons who had worked any day in the previous 42 days were included; later samples only consider current workers.

The 2004 and 2014 samples distinguish public from private salaried employees. All samples preserve data on the location of self-employment (separate premises, at home, or traveling) in the unharmonized source variables.

The 2014 sample combines some categories separately available in previous samples. Employers and members of cooperatives are grouped together and classified as "employers", while apprentices and unpaid family workers are grouped into "unpaid workers". The 2014 source variable also provides less detail about own-account workers compared to older samples.

Comparability — Nepal [top]

The samples are comparable across years.

Comparability — Netherlands [top]

Samples are comparable across years. Unpaid family workers are combined with wage and salary workers in all years.

Comparability — Nicaragua [top]

The 1995 and 2005 samples distinguish two types of employees -- day laborers and regular employees -- and also include a category for production cooperative members. The 1971 and 1995 census documentation specifies that the unpaid workers are members of the family, whereas in 2005 it explicitly notes that they may be related or unrelated.

Comparability — Palestine [top]

In 1997-2017 censuses, the question was asked of persons who were employed or experienced unemployed during the week preceding the census. The age universe changes from persons age 10+ in the 1997 sample to persons age 7+ in the 2007-2017 samples. There is an additional difference between the 2007 and 2017 samples, as the latter includes parts of Jerusalem that were annexed by Israel in 1967, while the former does not.

The categories are comparable across samples, with the exception of an additional distinction between regular wage employees and irregular wage employees in the 2017 sample.

Comparability — Panama [top]

Universes are fully comparable across samples, but the available categories differ, especially among types of wage and salary workers.

Comparability — Papua New Guinea [top]

The samples for 1990-2011 are comparable. The 1980 sample lacks detail for self-employed workers but provides additional detail among wage and salary workers. The 1980 census only addressed the question to urban dwellers.

Comparability — Paraguay [top]

The age universe changes from 12+ in 1962-1982 to 10+ in 1992-2002.

Comparability — Peru [top]

The data are fully comparable between samples apart from a slight difference in universes.

Comparability — Philippines [top]

The samples (1995-2010) are comparable across years.

Comparability — Poland [top]

The 2002 sample distinguishes between employers and own-account workers, but only the latter are included in 1978 and 1988. Own account workers in agriculture (farmers) are not separately identified in 2002. For this sample, the proportion of persons who are members of cooperatives drops significantly, given it only refers to those in agricultural productive cooperatives.

The category for wage or salary workers includes those employed in socialized and non-socialized economies in 1978-1988, and both part-time and full-time paid employees in 2002. In all Poland samples, a commission worker refers to a person obtaining payment from a commission according to the amount of goods sold or services provided; these cases are also classified as "wage or salary worker".

In 1978 and 1988, a "cottage worker" refers to an individual who processes raw materials and transforms them into a ready-made product in their own houses. These workers are categorized as "blue collar (manual)". Cottage workers differ from own-account workers because it is unknown whether such workers sell the goods on their own or send them back to the company hiring them.

The 1978 and 1988 source data assigned status in employment also to non-workers based on who "supported" them, but CLASSWK reports data only for economically active persons.

Comparability — Portugal [top]

The age universe changes from 12+ in 1981-1991 to 15+ in 2001-2011. The data are otherwise comparable.

Comparability — Puerto Rico [top]

The data are fully comparable between samples apart from a minor age universe difference in 1970 and the addition of non-profit work as a category starting in 1990.

Comparability — Russia [top]

The maximum age universe differs across samples. The 2002 sample identifies a category for self-employed workers in cooperatives and unpaid that is not included in 2010.

Comparability — Rwanda [top]

The samples (1991-2012) are comparable across years with slightly difference in one category. The category "employee" for 2002 is provided with detailed information i.e. "fulltime or part time employee". The age universe changes from 10+ in 1991, 6+ in 2002, to 5+ in 2012.

Comparability — Senegal [top]

The data are comparable across samples apart from minor differences. The 2002 sample adds an unpaid trainee category. In addition, the 2013 sample distinguishes between permanent and temporary salaried workers, while the earlier samples include them in the same category.

Apprentices in all three samples are coded as unpaid workers, though the instructions for the 1988 and 2002 samples indicate that some were remunerated.

Comparability — South Africa [top]

The universe differs among samples. The 1996 census addressed the question to persons age 15 and older from private households only, while 2001 asked it of persons age 10 and older. In 2007, the universe includes a top age restriction as well, including employed persons from age 15 to 74. All universes include employed persons, but 2001 also includes additional persons identified as economically active.

In 1996 there a single response for "works in a family business" could overlap with any of the other responses. This response is coded to "unpaid family worker" but it explicitly includes paid family workers as well. The 2001 and 2007 censuses separately identify paid and unpaid family workers, with paid family workers being coded with wage and salary workers.

Comparability — South Sudan [top]

Unpaid family workers are treated as a separate category from "unpaid working for others". Other unpaid workers besides family workers include "apprentices or trainees who are not paid, participants in food for work programmes and anyone else who works for others but does not receive money."

Comparability — Spain [top]

The universe differs with each census. The 1991 sample includes retired persons in the universe making its NIU category notably smaller than the other samples. In 2011, the question was asked of person's age 16+ that were employed or worked previously.

Comparability — Sudan [top]

Unpaid family workers are treated as a separate category from "unpaid working for others". Other unpaid workers besides family workers include "apprentices or trainees who are not paid, participants in food for work programmes and anyone else who works for others but does not receive money."

Comparability — Suriname [top]

The age universe differs across samples, as the 2004 sample records responses from employed persons age 12+ and the 2012 sample records responses from employed persons age 15 to 64.

The categories are comparable across samples, although the 2012 sample provides more detailed groupings of wage/salary workers and also includes an additional category for unpaid volunteers.

Comparability — Switzerland [top]

The data are comparable across samples apart from the differentiation of employers from other self-employed workers in 2000.

Apprentices are coded as paid workers in Switzerland.

Comparability — Tanzania [top]

The 1988 Tanzanian sample does not distinguish between types of own account workers. In 2002 and 2012 samples, 'own account workers' are divided into agriculture or non-agriculture. Non-agricultural works are classified in CLASSWK as 'own account, other.'

Comparability — Thailand [top]

The age universe changes from 11 years and older in 1970-1980 to 13 years in 1990-2000. The data are largely comparable, but some additional categories arise in the later samples. The 1990 and 2000 samples distinguish between government officers and other government employees, and they are coded to white collar and blue collar public workers in CLASSWK.

Comparability — Togo [top]

The samples are comparable across years. The 2010 sample offers a few more details, which are preserved in the source variables.

Comparability — Trinidad and Tobago [top]

For all samples, the category of wage or salary workers includes persons working in the private or the non-governmental sector.

The enumeration instructions documents across samples do not specify consistently whether an apprentice or learner receives an income or not; thus, these cases have been classified as unpaid workers.

Comparability — Turkey [top]

The samples are comparable.

Comparability — Uganda [top]

For the Uganda samples, CLASSWK is based on a question designed to capture employment status, and only secondarily class of worker. As a result, some of the categories typically found in CLASSWK (for example, employer) are not represented in the 1991 and 2002 samples. The 2014 sample provides separate categories for employer and own-account worker, as well as additional categories for unpaid workers, compared to the older samples.

Comparability — United Kingdom [top]

The universe differs between samples, with 2001 applying a maximum age.

Comparability — United States [top]

The universe for 1960 and 1970 includes persons age 14+ who had ever worked in the previous 10 years, regardless of their current employment status; for 1980 to 2020, the age cutoff is 16+ and respondents had to have worked in the past 5 years. All years distinguish between various classes of public (government) and private employment and identify unpaid family workers. The 1970-2020 censuses distinguish between employers and other self-employed persons whose businesses/farms were incorporated and those whose businesses/farms were not.

Comparability — Uruguay [top]

The 1975-1996 samples are comparable. The 1963 sample asks the question of young children and 2006 has a higher age threshold than other years for respondents. There are also differences in category availability across samples.

Comparability — Venezuela [top]

The universe differs among samples. The 1971 sample includes persons who have ever worked in the past who are not institutionalized or disabled; 1981 includes persons age 12+ who worked in the past year; and 1990 includes persons age 12+ who were in the labor force.

The 1971 and 1990 samples distinguish members of cooperatives from other self-employed persons. All samples differentiate between private and public workers, with 1981 making a further distinction between white and blue collar workers within sectors.

The 2001 sample distinguishes permanent from contract employees.

Comparability — Zambia [top]

The universe changes somewhat over time. The 1990 and 2000 enumeration instructions indicate that this question should have been asked only of present household members and visitors, but class of worker is reported for almost all absent household members; they are included in the universe.