Can you please confirm exactly which SCIAN vintage each of the Mexico 2000, 2010, and 2020 IND source variables uses? (and thus the comprehensive variable IND). What about 1990?
The SCIAN is the same thing as the NAICS: the North American Industry Classification System. The acronym is SCIAN in Spanish.
The 2020 Mexico census uses the 2018 SCIAN vintage (see page 281 of the Censo de Población y Vivienda 2020 Marco conceptual). The 2010 Mexico census uses the 2010 SCIAN vintage (see page 90 of the Marco conceptual del Censo de Población y Vivienda 2010).
The 2000 Mexico census uses the 1997 SCIAN Mexico, a Mexico-specific version of the SCIAN. This was the first version of the SCIAN used in Mexico. Note that in the 2000 census, although the SCIAN comprises six levels of aggregation, the nature of the census data necessitated coding at the second level—the activity subsector—with certain modifications introduced to facilitate comparisons between the results and the first level of the 1990 Census (see page 155 of the 2000 Marco conceptual).
According to the ILO, the 1990 Mexico census uses the Standard Industrial Classification (SIC).
You can read more about the use of the SCIAN in Mexico in “Implementation of the North American Industry Classification System (NAICS) in Mexico.”
The IND variable is not recoded, meaning it preserves the sample-specific codes. Because IND is not recoded, the codes may not be comparable across samples. As noted above, the industry classification system used varies from sample to sample, the same code may not have the same meaning across coding schemes. The variable INDGEN has been recoded to provide harmonized, comparable industry codes that can be used for any sample.