Confusion about coding of MIGRATEP

[Stephen_Pitts]

Dear Ipums International forum,

I am looking at internal migration in the Mexico 1960 census, which is coded slightly differently from the censuses from 1990 onward.

I created a data extract with MIGRATEP (nature of migration), GEO1MIG_P (origin state), and GEO1_MX (present state).

If MIGRATEP == 20 (move to different state), then GEO1MIG_P should be different from GEO1_MX, since the person started in one state and moved to another state.

Yet I found a few rows where this is not true (see screenshot). There are 933 rows like this. My extract has 502,800 records.

I am thinking of dropping these rows, but I wonder if there is something I am missing here.

Thanks.

[Matthew_Bombyk]

You are correct that these 933 should have a value of 10 (moved within same major administrative unit) for MIGRATEP. The IPUMS-I team will make this change, but in the meantime you can recode these cases manually, using the same criteria you used to identify them. Thank you for pointing out this error in the data! We like to show our appreciation for our data users who discover errors like this by sending them an IPUMS mug. Please send your mailing address to ipums@umn.edu and we will send you one in the mail.

[Matthew_Bombyk]

Following up on this, these 933 cases have now been fixed in the current version of the database. @Stephen_Pitts