[CTPP] Urgent: Proposal for reporting travel time and departure time in PUMS 2000.

In the July CTPP Status Report (http://www.mcs.com/~berwyned/census/newsltr/sr0700.html), we mentioned the Census Bureau proposal for PUMS 2000 to round travel time to work to five minute intervals, and round departure time to work to 15 minute intervals. Dr. Ashish Sen, Director, Bureau of Transportation Statistics and Mary E. Peters, Chair, AASHTO Standing Committee on Planning sent a letter to the Census Bureau expressing our concerns. Since then, the Census Bureau assigned Phil Salopek to work with the Department of Transportation and AASHTO to address our concerns. The following are the suggestions of the CTPP Working Group for reporting Travel time and Departure time in PUMS 2000. We would like your comments on our proposal. Please write to Nanda Srinivasan at ctpp(a)fhwa.dot.gov (phone: 202-366-5021) by September 1, 2000. Thank you! Nanda Srinivasan Proposal for PUMS: To protect individual confidentiality, the Census Bureau is generally using two approaches to reduce the chance of disclosure. For variables that are considered to form a continuous distribution, top coding is being used. At the national level, the top category must contain at least: a) 0.5% of the total US population (about 1.4 million persons), or b) 3% of the subpopulation or universe of the variable. The smaller result of a) and b) may be used. For variables that are categorical in nature, the Census Bureau is requiring that there be at least 10,000 persons (nationwide) in any specified category or grouping of values. There is no restriction on how the categories are formed, as long as they each contain enough people. If a category that has been specified beforehand turns out to have too few cases in it, the category will be combined with another one. Travel time: Travel time is a continuous variable. We would like Travel time to be reported in individual minutes between 1 and 89 minutes, and 90 and over as a top code. Explanation: According to the 1990 census 1.7 million people took 90+ minutes to get to work (but only about 850,000 took 91+ minutes). For criterion b) we would use 3% of the number of workers not working at home (111,664,249), or 3.3 million people. Since the result of criterion a) is smaller, for 2000 PUMS we could expect to get individual minutes of travel time from 1 to 89 minutes, with the top category being 90+ minutes. Departure time: We believe that time leaving home to go to work (departure time) should be treated as a categorical variable. We recommend the following scheme: 1. Midnight to 3 a.m. - 30 minute intevals. 2. 3 a.m. to 5 a.m. - 10 minute intervals. 3. 5 a.m. to 11 a.m. - 5 minute intervals. 4. 11 a.m. to 7 p.m. - 10 minute intervals. 5. 7 p.m. to midnight - 15 minute intervals. The attached adobe acrobat (.pdf) file provides the proposed specification of categories.