Connecting people around the 2010 U.S. Census and beyond
Incomplete/ Duplicate Mailed-In Forms
Three questions for people who are "in the know" on this:
1. If someone mails in a blank or incomplete form, does this get followed up on by field workers at some point? I read that blank forms are taken care of during VD Check, but I thought I'd include them in this question anyway.
2. During NRFU, I encountered a number of HUs where the resident said he had already mailed in the form. What steps does the Bureau take to prevent double counting of these?
3. Is any further action taken on EQs that were turned in as 99-POP UNKNOWN during NRFU close-out?
These are just some things I've been wondering. Thanks in advance for your help.
Views: 2
Replies to This Discussion
-
Permalink Reply by Peachtree FOS on -
1. According to the office manuals, addresses whose mailed forms were blank or incomplete are in the VDC workload.
2. The mailback questionnaire and the EQ label will have the same geocodes (state, county, block, map spot). The computer will recognize this after data capture and select one to represent the address (or, in unclear cases, will probably kick it to a human for either a decision or further inquiry about which data are correct).
3. Probably not. They're not in the VDC supplemental caseload as far as I know. If there's no better data (e.g., a mailback questionnaire), a population is imputed to the address based on neighboring HUs.
-
Permalink Reply by Non Cense on
-
Duplicates are identified by case id, when they get to processing center, NRFU VDC has begun, no actually inemeration will be done, except to verify vacant, deleted residents. The case loads will be compressed to less CLDs but more AAs. .Hope this helps
-
Permalink Reply by Sheri on
-
I've been trolling the internet to find methodology information. . . funny, I can't access the FDCA library from my computer but I get frequent hits to documents that are "for internal use only" that I can read. . . the "need to know" philosophy of the Census frustrates me because I do like knowing. . . maybe it's based on the idea that workers need to stay within the parameters of their function but still.
I have enough statistical background to get some of the gist of the stuff I find, and basically there is no single step for duplicate or blank or whatever questionnaires. There are many different statistical and data-crunching operations whose ultimate purpose is to:1) turn the raw data into usable form and 2) provide estimates of statistical numbers (e.g. false positive rates, false negative, variance, on and on) that are used to determine reliability, provide information that will be used in future statistical design, etc etc.
Short version being: don't freak out about bad data because it's not fatal. But get the best data you can. Note to Peachtree--I came across stuff about imputation, it's based on all kinds of stuff depending on what information is being imputed. But I wish our office had management people who even KNEW the word "imputed". I think it would be helpful if each LCO was linked with a local government statistical expert who could from time to time assist them to understand some of the basic methodology concept. We ended up doing a fair amount of bad work just because of the lack of that understanding. Last week a QA clerk entered NPC-bashing tirade in the notes for a MaRCS case, with the blessing of the AMQA, because they think the NPC clerks don't know what they are doing when they defer a case back to us, when in fact it looks like the NPC clerks are (to the same extent as any other operation) doing what they are supposed to.
-
Permalink Reply by Sheri on
-
We have exactly the same number of AA's, since the VDC materials go directly into the NRFU binders. Our training stuff said AA's would be grouped with multiple AA labels on each binder, but we did not do it that way and it may be the regional guidelines were revised. All the NRFU binders had been stored in labeled boxes, came out of those boxes, had the stuff added to them, and went back in the same boxes with the CLD numbers changed based on the final CLD assignments. Unless you mean more AAs in a CLD
© 2012 Created by Evan Goldin.
Powered by
