From picklel@mail.nih.gov Mon Jul 9 13:29:03 2001 Received: from mxu4.u.washington.edu (mxu4.u.washington.edu [140.142.33.8]) by lists.u.washington.edu (8.11.2+UW01.01/8.11.2+UW01.04) with ESMTP id f69KSt095544 for ; Mon, 9 Jul 2001 13:28:56 -0700 Received: from ims2.hub.nih.gov (ims2.hub.nih.gov [128.231.90.112]) by mxu4.u.washington.edu (8.11.2+UW01.01/8.11.2+UW01.04) with ESMTP id f69KSts03034 for ; Mon, 9 Jul 2001 13:28:55 -0700 Received: by ims2.hub.nih.gov with Internet Mail Service (5.5.2653.19) id <3GDA57L9>; Mon, 9 Jul 2001 16:28:54 -0400 Message-ID: <622E01CFB200F446B3E3EE069FA85A9123A060@nihexchange7.nih.gov> From: "Pickle, Linda (NCI)" To: "'waphgis@u.washington.edu'" Subject: RE: Zip Codes, ZCTA and areal interpolation. Date: Mon, 9 Jul 2001 16:28:53 -0400 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2653.19) Content-Type: text/plain; charset="iso-8859-1" I agree with the complaints about zip code data mentioned on these postings. Another problem with ZCTAs that has not been mentioned is that for the health surveys for which we only ask zip code and not address, or for which address is not available for confidentiality reasons, we can't substitute a question about ZCTA. No one will know their ZCTA #, but they can provide a zip code because it is needed for their mail. It would have been helpful if the Census Bureau could provide us with a "cross-walk" to look up ZCTA(s) from zip, but according to someone I talked to there, they have no plans to do so. Here at NCI we have planned to use zip code populations to calculate rates of phone calls to the 1-800-4CANCER help line, but I was told that (1) no 2000 populations by zip code will be produced and (2) Census is not providing a way to convert from ZCTA to zip population. We seem to be stuck with lots of numerator data but no available denominator. Can someone on the list confirm or correct this info re: 2000 zip pops? Thanks. Linda Williams Pickle, Ph.D. Senior Mathematical Statistician Statistical Research and Applications Branch Surveillance Research Program Division of Cancer Control and Population Sciences National Cancer Institute 6116 Executive Blvd., Suite 504 Bethesda, MD 20892-8317 (for overnight delivery, use Rockville, MD 20852) Branch phone: 301-435-7739; Fax: 301-480-2046 Email: PICKLEL@mail.nih.gov -----Original Message----- From: Susan Clemens [mailto:s-clemens@wildmail.com] Sent: Monday, July 09, 2001 4:04 PM To: waphgis@u.washington.edu Subject: RE: Zip Codes, ZCTA and areal interpolation. I have been skeptical of zip codes and/or ZCTA for some time now. Unfortunately, I am usually analyzing data collected by another agency so don't have the option of requesting more specific geographic information. Often a zip code is all that is collected or all that can be released due to patient confidentiality issues. I agree with the observation that zip codes are not consistent through time. As I understand it, perimeter changes are the norm when designating new zip codes. Also, zip + 4 numbers change more frequently than zip code perimeters and often have more to do with postal carrier staffing and routes vs population. Lastly, the ZCTA for the 1990 Census were developed using a different methodology than the 2000 ZCTA. Even if you are lucky enough to be in an area with no new zip codes you cannot compare 1990 and 2000 ZCTA data. So, where does this leave us? Is commercial zip code level data any more reliable that the ZCTA? Is any zip code level analysis for health valid given the constraints mentioned? Susan Clemens Humboldt Co Dept of Public Health ---- Begin Original Message ---- From: fpb01@health.state.ny.us Sent: Thu, 5 Jul 2001 15:44:38 -0400 To: waphgis@u.washington.edu Subject: RE: Zip Codes, ZCTA and areal interpolation. I believe that ZCTAs were created for the exact purposes given in Richard's original posting: to provide an easier way to deal with the mass of data that is geocoded to zip code only, and to provide a standard (and free) base map on which to map this data. ZCTAs are probably good enough for many purposes, such as marketing applications. They should also be good enough for certain kinds of proportional analysis. For example, you can probably safely compare a variable such as % smoking from a zip code-based survey with % completed high school from ZCTA. When it comes to using ZCTAs for calcuating incidence and mortality rates, though, I am pessimistic. Frank Boscoe NYS Department of Health "Richard Hoskins" @u.washington.edu on 07/05/2001 03:12:24 PM Please respond to waphgis@u.washington.edu Sent by:  WAPHGIS-owner@u.washington.edu To:   cc: Subject:  RE: Zip Codes, ZCTA and areal interpolation. Frank:  What was the purpose of the census bureau's developing the ZCTAs? Dick H -----Original Message----- From: WAPHGIS-owner@u.washington.edu [mailto:WAPHGIS-owner@u.washington.edu]On Behalf Of fpb01@health.state.ny.us Sent: Thursday, July 05, 2001 12:00 PM To: waphgis@u.washington.edu Subject: Re: Zip Codes, ZCTA and areal interpolation. I believe that conversion from Zip Codes to ZCTAs is, unfortunately, more than just an areal interpolation problem. The two entities are not the same thing - they differ along the edges. On a road that forms the boundary of a zip code, both sides of the road may be in the same Zip Code. This is reasonable and facilitates the delivery of the mail. But the two sides of the road will be in different census blocks and hence different ZCTAs (since ZCTAs are built from census blocks).  In heavily populated areas, the distinction is almost negligible, but not in rural areas. In fact, the census produced a graph which shows that zip code-ZCTA agreement is almost perfect in zip codes with more than 15,000 addresses, but under 80% where there were fewer than 250 addresses. In terms of health mapping, if a nursing home or hospital or other special type of address falls on a zip boundary, its entire caseload could be assigned to the wrong ZCTA, resulting in substantial error even in a heavily populated area. Frank Boscoe NYS Department of Health ---- End Original Message ---- Care2 make the world greener! http://www.care2.com - Get your Free e-mail account that helps save Wildlife! .