search

bdecon / econ_data

Python 3 examples of using economic data APIs and working with economic microdata. Includes bd CPS.
74 stars 38 forks source link

bd CPS: issue from use of 2000-02 revised weights #249

Closed bdecon closed 4 years ago

bdecon commented 4 years ago

Recently tried to calculate the headship rate, following the example from this FEDS Note (See bd_CPS_headship.ipynb).

One issue is that the revised person weights for 2000 to 2002, which reflect new data gathered in the 2000 census, were included in the bd CPS but the bd CPS did not update the corresponding household weights. There are two options here:

  1. Use original person weights
  2. Update household weights

The second option seems better. Usually the household weight is just the person weight of the "householder". Will need to play with the data a bit to make sure. If this doesn't work, go with option 1.

bdecon commented 4 years ago

From IPUMS:

The final household weight equals the weight of the wife in a husband-wife household and it equals the weight of the reference person in all other households.

bdecon commented 4 years ago

This is resolved by some changes to the revision_reader notebook. I generate a new household weight equal to the person weight of the person who supplied the original household weight.