Thanks, Ed. I have been working LEHD data for 2 weeks and comparing it with other sources (REMI, Claritas, WP, etc). All sources has different number for 2010, for example, which one is the good one ? LEHD has all jobs, primary jobs, private jobs, etc. I found the use primary job data give close number with other sources, am I on the right track ? I have not compared LEHD data with 'just released' commuting data set, I guess it would be the same and LEHD has better resolution. What do you think ? Thanks, Hary On Fri, Mar 8, 2013 at 10:22 AM, Ed Christopher <edc(a)berwyned.com> wrote:

Just to provide some BASIC items of differences on data sources for home/work FLOWS. ACS and CTPP based on ACS: This is a sample SURVEY of households and household members. The ACS is an address-based sample. The response rate to ACS is lower than to Census 2000 "long form" and the field follow-up for non-respondents to the mail-back form is for a SAMPLE of non-respondents. All the responses are weighted to come up with a total estimate. Respondents are asked to provide their workplace location, which are then geocoded by the Census Bureau. About 25 % of workers do not provide sufficient information to code to a census block and their workplace is imputed using a routine using 4 variables: industry, occupation, travel mode and travel speed. See the recent CTPP Status Report article by Melanie Rapino on the details of the imputation process. LEHD: is from ADMINISTRATIVE records, based on Employment Security (unemployment), and augmented with federal employment records. The base is from files submitted to the Census Bureau from State Departments of Employment Security who maintain the QCEW files (formerly called ES-202). LEHD does not currently include self-employed, or military employment. In this case, the workplace location is the address(es) provided by businesses. If the business has multiple locations in a county, many states require that each location be reported. Only in Minnesota is the employee record tied to a specific workplace location. That is, in all other states, for businesses with multiple sites, the link between an employee's home location and the workplace location is imputed. That is, an algorithm selects among the multiple worksites which is the most likely workplace for that worker. The home location may be moved to protect individual confidentiality, but largely the moves are within the same census tract. The link between the unemployment (QCEW) and home location is primarily from IRS tax records, using an encrypted version of SSN. Sometimes people use a different home address for IRS than where they are actually living, for example young people or college students may use their parents' address. And similarly, sometime people report on the ACS form a workplace location where they may not actually be working, for example, a construction worker might report the business address of his employer. So, there is plenty of room for inconsistencies. Private lists of businesses and employment totals: Typically these are compiled using web resources and telephone surveys. Often the list guarantees that each business has been contacted within 3 (or 5) years to get the most current employment figures. Because these firms have been compiling these lists and checking them for so many years, they have been able to refine their databases. However, these lists do not provide FLOW between work and home, just workplace location totals. Please let me (and the listserv) know if I have made any mistakes above. Elaine Murakami ________________________________________ From: ctpp-news-bounces(a)chrispy.net [ctpp-news-bounces(a)chrispy.net] on behalf of hprawiranata mitcrpc.org [hprawiranata(a)mitcrpc.org] Sent: Friday, March 08, 2013 11:49 AM To: ctpp-news(a)chrispy.net Subject: Re: [CTPP] Commuter Flow Data Concerns Thanks, Ed. I have been working LEHD data for 2 weeks and comparing it with other sources (REMI, Claritas, WP, etc). All sources has different number for 2010, for example, which one is the good one ? LEHD has all jobs, primary jobs, private jobs, etc. I found the use primary job data give close number with other sources, am I on the right track ? I have not compared LEHD data with 'just released' commuting data set, I guess it would be the same and LEHD has better resolution. What do you think ? Thanks, Hary On Fri, Mar 8, 2013 at 10:22 AM, Ed Christopher <edc(a)berwyned.com> wrote: _______________________________________________ ctpp-news mailing list ctpp-news(a)ryoko.chrispy.net http://ryoko.chrispy.net/mailman/listinfo/ctpp-news