Comparability — Cambodia [top]

The universe differs across samples.

Comparability — Canada [top]

The samples all refer to work in the previous full calendar year. The data were reported as number of weeks worked. For 1971, the weeks were reported into broad intervals, and these are coded to the midpoints for WRKMTHS. The other samples reported exact numbers of weeks, which are converted to months: from one to four weeks is coded to 1 month; the 3rd, 6th, 9th, and 12th months are each assigned 5 weeks, and other months are 4 weeks. See the unharmonized input variables for the original number of weeks reported.

Comparability — Colombia [top]

The samples for Colombia are generally comparable, but the age universe differs. The 1973 census was specific that a month worked was one in which the person worked for 15 or more days. The other censuses were not explicit.

Comparability — Guatemala [top]

The data refer to the number of months the person worked during the 1963 calendar year. The "less than one month" category uses a different source variable to identify work within the last four weeks, and the number of cases may be inflated.

Comparability — Haiti [top]

The 1982 sample refers to work in the previous full calendar year. The data were reported as number of months worked.

Comparability — Honduras [top]

The data refer to work in the previous 12 months. In 1961, the data were reported as broad intervals: less than 3 months, 3 to 9 months, and 9 months or more. These categories were coded to the midpoint of the interval.

Comparability — Israel [top]

In the 1995 and 2008 samples, the original categories before recoding were 1-3, 4-6, 7-9, 10-11, and 12 months. In the 1995 sample, a month of work was one in which the person worked at least one week; the 2008 sample only requires the person to have worked one day to count that month. Both the 1995 and 2008 censuses indicate that months when the person was absent from their job (due to vacations, maternity leave, etc.) should be counted.

Comparability — Jamaica [top]

Number of months worked are coded into midpoints and are rounded to the nearest month in 1991 (See the unharmonized variable for the original month ranges).

Comparability — Mauritius [top]

The variable for the 1990 sample is constructed from the number of weeks worked in the preceding calendar year; the enumeration instructions indicate to count any week in which the person worked even 1 hour, including periods of approved leave.

Comparability — Nepal [top]

The data are comparable between years. The values for 2011 are calculated from four source variables identifying work under different types of employment.

Comparability — Puerto Rico [top]

The 1970 Puerto Rico sample has a different universe statement from the other four samples (1980-2005). In the former sample, it reads: "Persons age 14+ who worked last year", while in the latter it reads: "Persons age 16+ who worked last year".

The 1982 sample has an "Unknown" category, but no data months 9, 10, 11 and 12. The 2020 and 2015 samples were asked to people age 16+ who worked last year. The 2020 census identified persons age 16+ working within the last five years, not new workers.

Comparability — Saint Lucia [top]

In the 1980 sample, some of the months are merged (2-3 months; 4-5 months; 6-7 months; 8-9 months; and 10-11 months). The 1991 sample, instead, has one category per month. Comparability, then, is not perfect. Also, the 1980 sample has the category "unknown", while the 1991 sample does not.

Comparability — Trinidad and Tobago [top]

In 1970 and 1980, most of the months are merged (2-3 months, 4-5 months; 6-7 months; 8-9 months; and 10-11 months). The data were assigned to the lower value of the interval (i.e. 2-3 months is coded as "2 months" in WRKMTHS.)

Comparability — United States [top]

The data refer to work in the previous calendar year, or the previous 12 months in 2005.

In 1960, 1970, 2010, and 2015 the data were reported as broad intervals: 1 to 13, 14 to 26, 27 to 39, 40 to 47, 48-49, and 50-52. These categories were coded to their rough midpoints in terms of months.

For 1980-2005 and 2020 the data were reported as number of weeks worked. These were converted to months: from one to four weeks is coded to 1 month; the 3rd, 6th, 9th, and 12th months are each assigned 5 weeks, and other months are 4 weeks. See the unharmonized input variable for the original number of weeks reported.

The age universe changes from 14+ in 1960-1970 to 16+ thereafter.