art photo



I downloaded a US Census Extract recently and have been running some descriptive statistics. In terms of race, Brazilian median household income immigrants mostly identify as white, "some other race" and "two or more races". I would like to find out how they classified in the latter 2 categories. Is there any way I can do this? Also, if it is just a matter of creating an extract with specific variables (e.g. "OTHER", NUMRACE), would it be possible to add these variables to my existing dataset? Or would I need to do a completely new data extract with all of the variables?


RACE3 has individuals classified by various combinations. For instance, [see codes on page 213 in documentation]

48 White; Some other race
68 White; Black or African American

There are also separate race recodes (e.g., WHITE, BLACK,etc. but they don't always help identify the "other race." For instance:

0 No
1 Yes, alone or in combination with one or more other races

If a person is coded 68 on RACE3 they will be coded Yes on the WHITE variable and Yes on the BLACK variable.

Other race means that the individual marked the box next to "Other." If they filled in something like "German" the Census Bureau would have recoded the individual to "White" but if they put in an Hispanic choice like Mexican or Dominican, the "Other" choice would remain.

As to running another extract. . . Yes, you will have to select all your variables again, unless you took off Serialno and Pnum. If you take those variables off, you can make an ID:

ID = Serialno*100 + Pnum;

Then if you need to pull off any more variables, you can merge via ID.

Annotated Resources:

Direct Links:

Related Question Groups: