PINC vs HINC for Income by Gender

Hi ACS Folks,

I am using PUMA data in order to get a bit more dis-aggregation by race (dis-aggregated Asian races) and household type and in order to compare incomes and poverty status to the UW Self Sufficiency Standards. Note, I am ONLY working with data for the State of Oregon, as I work for the state government.

I'm attempting to calculate median HINC, PINC and FINC in my software (SPSS) and then would like to visualize the individual response, with the weights, in Tableau. The weights (both PWGT and WGTP), when I add them on, put me within $1,500 of the ACS S1903 standard table for Family Median Income and Household Median Income. I cannot,  however, find a table to compare income by gender on, nor am I sure which variable in PUMS to use (HINC or PINC). In the S1903 Median Income ACS Standard Table, does anyone know whether I would use non family member household by gender to verify if my estimate is close? When I use PINC, the estimates are very small, as the median PINCP for women in Oregon is $19,000 and for men it is $32,000. There is no ACS table that I can find that uses PINC....unless the non-family member household at the bottom of S1903 uses PINC.

When I use HINC and calculate median HINC using the gender of the reference person (RELP=0), the estimates are higher than "female headed householder" and "male headed householder" in the ACS S1903 standard table.

I'm wondering if anyone has other experience with this. Additionally, is it standard practice to leave in the negative entries for the incomes? I wondered if that's why it's bringing my PINC down...there are almost 1400 negative income entries in the state of oregon.

Any thoughts on this would be appreciated.

Thanks!

Parents
  • Sorry for a really late potential response...but I can think of 2 potential reasons for the higher median HINC:
    1) Make sure that you are excluding non-response values (I know in IPUMS, HHINCOME value of 9999999 = N/A) -- I often forget to do this on my first pass!
    2) When you do the analysis by the gender of the reference person, you could be including married households, as well. (Which are distinct from female-headed household and male-headed household, as these are unmarried either single parent or cohabiting HH with families).

    And yes, it is standard practice to leave in negative entries for income.
Reply
  • Sorry for a really late potential response...but I can think of 2 potential reasons for the higher median HINC:
    1) Make sure that you are excluding non-response values (I know in IPUMS, HHINCOME value of 9999999 = N/A) -- I often forget to do this on my first pass!
    2) When you do the analysis by the gender of the reference person, you could be including married households, as well. (Which are distinct from female-headed household and male-headed household, as these are unmarried either single parent or cohabiting HH with families).

    And yes, it is standard practice to leave in negative entries for income.
Children
No Data