I’m trying to figure out what “Family market value” means for several of the poverty measures including “SCHLLUNCH” and “FOODSTAMP”, but the documentation is vague and the questionnaire text is not available. Does this mean how much they spent, or what is the market value of what the family spent, or how much value the family places on what they spent? Of maybe none of these options -
Thanks!
The documentation is vague on these variables. We are now in the process of upgrading the description of these values. In general, “family market value” represents the dollar value of the benefit in question. Therefore these values identify the equivalent dollar value of just giving cash rather than the benefit. These values are enumerated slightly differently across the SCHLLUNCH, HOUSSUB, and FOODSTAMP variables. FOODSTAMP is reported directly by the respondent. (See section 5.11 in the ASEC technical documentation.) SCHLLUNCH is imputed based on how many individuals within a family receive school lunches. HOUSSUB is also imputed based on information about the market value of housing units in the family’s geographic area. I hope this helps.
Is the data of these variables monthly or yearly?
Good question. Although the SCHLLUNCH and FOODSTAMP variables are in terms of annual values, the HOUSSUB variable is in terms of monthly values. This is a bit confusing, so do note that the IPUMS CPS team is planning on revisiting how these variables are presented. When any revision is made it will be noted here.
Thanks for clarifying this information, Jeff. Is there any chance you could link to the particular ASEC Technical Documentation file that lays out explicitly which of these variables are imputed? You mention here that section 5.11 of the ASEC Technical Documentation explains this, but the hyperlink intended for that documentation just links to the FOODSTAMP variable documentation on IPUMS.
One additional question, Jeff. The documentation for SCHLLUNCH states that, " SCHLLUNCH reports the family market value of school lunches. Unlike FMVSCHLUNCH, the values of SCHLLUNCH are self-reported." Is this in contradiction with your previous comment where you stated that the SCHLLUNCH value is imputed? If the values are self-reported, wouldn’t this mean that they were not imputed?
Ah, sorry about the incorrect link above. Here is a link to the 2019 ASEC Codebook. You can find these documents on this page on IPUMS CPS.
Thanks for raising this question about FMVSCHLUNCH. After checking with the IPUMS CPS Team about this detail, FMVSCHLUNCH is a variable that does not exist. So, I can confirm that SCHLUNCH values are imputed based on how many individuals within a family received school lunches. We will edit the variable description for this variable in the next data release.
A belated thanks for this information, Jeff! I have one additional question related to how the imputed value for SCHLLUNCH is determined. I dug a bit into the documentation you provided, and I wasn’t able to find the answer.
When the CPS imputes these SCHLLUNCH values, are the imputed dollar amounts adjusted for the average cost of lunch in the state or school district (I doubt it’s this fine-grained) where the data was collected? To illustrate by example, the average cost of a reduced or free lunch is likely different in New York versus say Mississippi. So when imputing the value of SCHLLUNCH for a family of four (two adults, two kids) in New York the dollar amount should be higher than for the same family in Mississippi just due to cost of living differences, right? Further, I understand that the pricing of a hot lunch at school is highly decentralized; often it’s decided by individual school boards. In sum, do you know whether the CPS applies a single value per free or reduced lunch per child recipient across all households, or do they somehow adjust for cost of living?
According to this document:
For this study (and for Census Bureau estimates), the average (over the 48 contiguous states) reported school lunch payment rates , for schools in which less than 60 percent of the lunches served during the second preceding school year were served free or at a reduced price, are assigned to each student . The appropriate per-meal value (for either free, reduced, or paid meals, depending on the level of eligibility) is multiplied by the participation- adjusted number of children between the ages of 5 and 18 , and then by 167, the number of days students are assumed to be in school. This is the same number of days used for estimating NSLP benefits in SPM resources by the Census Bureau.
So it seems these are not imputed differently by state.
Thank you @Matthew_Bombyk!
@Matthew_Bombyk, looking again at the document you linked to, this paper by Garner, Gudrais, and Short seems to be referring to the process of imputing the SPM school lunches value so SPMLUNCH. Do you have any sense of whether this is also the procedure by which the CPS imputes the non-SPM school lunches value so SCHLLUNCH? I’m trying to find specific documentation on this, or additional documentation to support Garner, Gudrais, and Short’s imputation formula, and I’m unable to locate any.
Matt is out of the office, but I am able to answer your question. There is not a lot of detailed documentation on how school lunch values are imputed in ASEC data. This 2019 report on the supplemental poverty measure offers a bit of information in the appendix. While the resource I shared and the paper that Matt provided are about SPM, SPMLUNCH should simply change who is being assigned a value for school lunch, but not how Census is imputing the value. SPMLUNCH reports the value of school lunch by resource units as defined by the supplemental poverty measure (SPM). Because documentation on the topic is minimal and the imputation is performed by Census, you may want to email the CPS folks at the Census Bureau if you are interested in more information.