The variable HEATVAL quantifies the value an individual receives from the Federal Low-Income Home Energy Assistance Program. The description of the variable in IPUMS states that the variable measures the value of the Assistance Program for only six months out of the year: “the total value of energy assistance the household received since October 1 of the previous year.” However, if one looks at the technical documentation from the Census (page 108, or 7-1 of the Glossary), it appears as though after 2011, the variable shifts to record the annual value of energy assistance:
The energy assistance questions were asked for the first time in 1982. In 2011, the question was revised to include assistance for cooling as well as heating expenses, and the reference period was expanded from: (a) receipts since October 1 of the previous year; to (b) receipts for the entire previous calendar year.
If this is true, I have one suggestion and a question. First, that the IPUMS documentation on HEATVAL be changed to reflect the fact that a) the reference period shifts after 2011 to encompass a full 12 months (rather than 6) of energy assistance; and b) that the questionnaire shifted to include cooling assistance. Presumably both changes to the questionnaire would have significant impact on respondents’ reported value.
Second, does the Census provide any guidance on how to deal with this variable before 2011? If the variable captures only six months of energy assistance, would a user of this data prior to 2011 multiply the value by two to annualize it? Or is this too presumptuous?
Thanks for bringing this to our attention. Interestingly, the data dictionaries included in the ASEC codebooks retain the “since October 1” specification in 2011-2015 and only calls out heating costs in the glossary definition of “energy assistance program”, but also include the note you mentioned about covering the full year and updated questionnaire text for heating/cooling and the full year. I have passed this information along to the IPUMS CPS team with a request to update the documentation (including the comparability text) accordingly.
I am not aware of guidance from the Census Bureau on how to address this different reference period. Without this guidance, I will leave it up to you to determine if the assumption that costs should be doubled is reasonable. A few things to consider when making this determination:
- The six months covered approximately correspond to winter in the United States; doubling this assumes that heating expenses are constant throughout the year or a reasonable proxy for cooling expenses (i.e., in ASEC samples that ask about energy assistance “since October 1,” the variable does not include cooling expenses). This assumption might be more reasonable in the Midwest than the South; it is also highly dependent on how a home is heated/cooled.
- Note that for a small subset of those in the ASEC (the oversamples), the ASEC interview does not take place in March (see Table 2 of this paper for details on ASEC interview months for the oversample). This would NOT affect the 2011-forward data, as the question asks about the preceding calendar year rather than the preceding 12 months. However, the phrasing of the question prior to 2011 might be affected by this.
- While I would assume the reference period is October 1 through the interview date, the questionnaire text in the codebook notes “Enter annual amount only” even for the pre-2011 samples. I do not see any guidance in the codebooks on how to interpret this note. I have a request out to the IPUMS CPS team for any information they have, and will follow up with the Census Bureau depending on what I hear back from my IPUMS colleagues who work with CPS data.
I am sure you were hoping for a more definitive answer, but hope this helps. If you would like to follow up with the Census Bureau directly for any guidance on the comparison, I encourage you to contact them at email@example.com.