Questions? Call 800-624-6242 About | Ordering | New Menu | Items in cart [0] SEARCH

Rights & Permissions Free PDF Access Sign in to download PDF book and chapters Free Resources Display this book on your site! Related Titles Using the American Community Survey: Benefits and Challenges Using American Community Survey Data to Expand Access to the School Meals Programs Other Related Titles NCHRP Report 588: A Guidebook for Using American Community Survey Data for Transportation Planning (2008) National Cooperative Highway Research Program (NCHRP) Citation Manager Export the bibliographic data for this book in your chosen format Web Search Builder Use this book's key terms to search within this book, across our collection, or across the Web. Skim This Chapter Skim this chapter and use this chapter's key terms to search within this book. Reference Finder Paste in your own text to find books that relate to your topic. Citation Manager Ruiter, Earl, McGuckin, Nancy, Transportation Research Board. "Appendix I - Comparison of ACS and Decennial Census Transportation Planning Estimates." NCHRP Report 588: A Guidebook for Using American Community Survey Data for Transportation Planning. Washington, DC: The National Academies Press, 2008. Please select a format: Page 253   E-mail Share on facebook Facebook Share on delicious Delicious Share on stumbleupon Stumble Share on twitter Twitter More Sharing ServicesMore Page 253 Front Matter (R1-R9) 1.1 Overview of the American Community Survey (1-1) 1.2 Some Important Implications of ACS for Data Users (2-2) 1.3 Purpose and Organization of this Guidebook (3-3) 1.4 Additional Information Sources for an Introduction to ACS (4-5) 2.1 ACS Implementation (6-17) 2.2 Additional Information Sources on ACS Implementation (18-21) 3.1 ACS Data Products (22-29) 3.2 Additional Information Sources for Obtaining ACS Data (30-30) 4.1 Accuracy of ACS Data (31-36) 4.2 Data Accumulation over Time and Geography (37-43) 4.3 Data Disclosure Limitations (44-47) 4.4 Understanding, Working with, and Reporting Sample Data (48-52) 4.5 Comparison of ACS Estimates to Census (53-64) 4.6 Implications of ACS Data Release Frequency (65-72) 5.1 Descriptive Analyses (73-75) 5.3 Descriptive Analysis Case Studies (76-100) 5.4 Other Specific Uses of Census Data for Descriptive Analyses (101-101) 6.1 Trend Analysis (102-103) 6.3 Trend Analysis Case Study (104-110) 6.4 Conclusions from the Case Study (111-113) 6.5 Specific Uses of Census Data for Trend Analyses (114-116) 7.1 Transportation Market Analysis (117-118) 7.3 Transportation Market Analysis: Environmental Justice Case Study (119-125) 7.4 Conclusions from the Case Study (126-126) 7.5 Specific Uses of Census Data for Market Analyses (127-130) 8.1 Survey Development and Analysis (131-131) 8.2 Benefits and Limitations of ACS for Survey Development and Analysis (132-132) 8.3 Case Study (133-140) 9.1 Travel Demand Modeling (141-143) 9.3 Travel Demand Modeling Case Studies (144-148) 9.4 Specific Uses of Census Data for Travel Demand Modeling (149-150) Appendix A - Housing and Population Questions From ACS and Census Long Form (151-161) Appendix B - ACS Base Tables (162-184) Appendix C - ACS Data Profiles (185-206) Appendix D - ACS Multiyear Profiles (207-226) Appendix E - ACS Ranking Tables (227-233) Appendix F - ACS Thematic Maps (234-235) Appendix G - ACS Subject Tables (236-245) Appendix H - ACS Selected Population Profiles (246-252) Appendix I - Comparison of ACS and Decennial Census Transportation Planning Estimates (253-262) Appendix J - Seasonality Analyses Using ACS (263-275) Abbreviations used without definitions in TRB publications (276-276)

Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 253
APPENDIX I Comparison of ACS and Decennial Census Transportation Planning Estimates We evaluated the general quality and validity of three-year accumulations (1999­2001) of ACS residence-, workplace-, and flow-based transportation-related data for nine test counties by comparing them to Census 2000 data that correspond to CTPP Part 1, Part 2, and Part 3 data. The ACS and Census data tables were provided to the project team by FHWA, who had received them for evaluation from the Census Bureau. This appendix summarizes the analyses that were conducted. The database included tract-level tabulations for nine counties and transportation analysis zone (TAZ)-level tabulations for the five counties for which TAZ data had been specified. Table I.1 shows the geographic areas covered. The comparison tables that were provided included those CTPP-type tables that are listed in Tables I.2, I.3, and I.4 for the residence-, workplace-, and flow-based estimates. Residence-Based Evaluation For the residence-based estimates, we computed the differences in estimates between the ACS and CTPP, tried to identify statistically significant differences, and looked for factors that might contribute to those differences through regression analysis. The estimates available in the Part 1 datasets are counts (e.g., number of people/house- holds/housing units with a certain characteristic). To compare the estimates, we converted the individual table cell estimates to percentages of the table totals. This means that differences due to slightly different weighted populations were accounted for. We then graphically examined the differences in the percentages between the CTPP data and ACS data at the tract and TAZ levels. For the most part, the two datasets appear to show the same patterns for the transportation- related tables. Only a small number of tracts and TAZs show significant variance between the two datasets. However, it should be noted that for some tracts and for many TAZs, ACS sample sizes were too small to show values. Next, we tested the significance of the observed differences. The standard errors of the ACS esti- mates were calculated by the Census Bureau and provided in the datasets used for this analysis. Since the CTPP standard errors were unavailable, we computed them using the methods described in the SF3 documentation, Section 8, "Accuracy of the Data." For the purpose of statistical signif- icance testing, the difference in estimates is defined as the ACS count minus the CTPP count (note that for several of the tables provided, the ACS and CTPP estimates are means and medians). It is difficult to find any statistically significant differences at the 95 percent level of confi- dence, especially for smaller geographies. Examining those standard errors further, we note that 253

OCR for page 254
254 A Guidebook for Using American Community Survey Data for Transportation Planning Table I.1. Geographic areas represented in evaluation dataset. Census Census County Tract-Level Data TAZ-Level Data Pima County, Arizona San Francisco County, California Broward County, Florida Lake County, Illinois Hampden County, Massachusetts Douglas County, Nebraska Franklin County, Ohio Multnomah County, Oregon Bronx County, New York Source: Federal Highway Administration, 2004. in most cases the ACS standard errors are much higher than the CTPP standard errors, ren- dering the t-statistic of the difference small enough as to be statistically insignificant. Figures I.1 through I.9 show the ACS and CTPP 95 percent confidence intervals for randomly selected tracts within each test county for an example variable. Note that the ACS confidence intervals are in general much wider than the CTPP confidence intervals because of the larger ACS stan- dard errors. The estimates for the variables appear to be largely in line with one another, but the sampling error for the three-year ACS data is too large to allow us to statistically confirm that this is the case or to determine if certain variables are more prone to be different. Although it was difficult to draw any meaningful conclusions about the statistical significance of the differences in estimates, we could examine the data to see whether the estimate differences were correlated with other tract characteristics. If this were the case, future comparisons between ACS and year 2000 CTPP data would be biased. We modeled these differences as a function of various population and household character- istics to check whether any particular variable is likely to bias the ACS estimates. We tested for the presence of systematic biases through regression analysis. For population estimates, we regressed the difference between the ACS and CTPP percentages as a function of the following population characteristics: · Total tract population, · Percent of population that are non-Hispanic white, · Percent of population that are Hispanic, · Percent of population that are 75 years or older, · Percent of population without a high school diploma, · Percent of population with an income below poverty line, and · Percent of workers in households with a disability. For household estimates, we regressed the difference between the ACS and CTPP percentages as a function of the following household characteristics: · Total households, · Percent of householders that are non-Hispanic, · Percent of householders that are Hispanic, · Percent of householders that are 75 years or older, · Percent of households with six or more people, · Percent of households with income below poverty line, and · Percent of households with single female head and kids.

OCR for page 255
Comparison of ACS and Decennial Census Transportation Planning Estimates 255 Table I.2. Residence-based evaluation dataset tables. Table Universe Content 1 All persons Total Population 2 Number of People Sampled 3 Sex by Age 4 Hispanic Origin by Race 5 Persons 16 and over Employment Status 6 Workers in Total Workers 7 households Mode to Work 8 Travel Time to Work 9 Time Leaving Home for Work 11 Household Income 12 Disability Status 13 Poverty Status 14 Disability Status by Mode to Work 15 Industry 16 Mode to Work by Time Leaving Home for Work 17 Mode to work by Travel Time to Work 18 Hispanic Origin 19 Hispanic Origin by Race 20 Hispanic Origin by Race by Mode to Work 21 Number of Workers in Household 22 Means of Transportation 23 Median Travel Time by Mode to Work 24 Mean travel Time by Mode to Work 26 Average Number of Workers per Vehicles 28 Households Vehicles Available 29 Tenure 30 Number of Persons in Household 31 Number of Persons in Household by Number of Workers in Household 32 Number of Persons in Household by Vehicles Available 33 Number of Persons in Household by Household Income 34 Number of Workers in Household by Vehicles Available 35 Number of Workers in Household by Household Income 36 Telephone Availability Number of Workers in Household by Vehicles Available by Household 37 Income 38 Mean Household Income 39 Median Household Income 40 Housing units Aggregate Number of Vehicles 41 Total Housing Units 42 Number of Housing Units Sampled 43 Percent of Housing Units Sampled 44 Quarter 45 Occupancy Status Source: Federal Highway Administration, 2004.

OCR for page 256
256 A Guidebook for Using American Community Survey Data for Transportation Planning Table I.3. Workplace-based evaluation dataset tables. Table Universe Content 1 Workers in Time of arrival at work 2 Households Worker earnings by Mode to work 3 Mode to work by Time of arrival at work 4 Mode to work by Travel time to work 5 Hispanic origin by Race by mode to work 6 Household income 7 Household income by Mode to work 8 Vehicles available by Mode to work 9 Median earnings by Mode to work 10 Mean earnings by Mode to work 12 Aggregate number of vehicles by Time Leaving Home for Work 13 Average Number of workers per vehicles by Time Leaving Home for Work 14 Number of workers per carpool by Time Leaving Home for Work 15 Median travel time by Mode to work 16 Mean travel time by Mode to work 17 Mean travel time by Mode to work 18 Mode to Work 19 Median travel time by Mode to work by Time arriving at work 20 Mean travel time by Mode to work by Time arriving at work 22 Aggregate number of carpools by Time Leaving Home for Work Source: Federal Highway Administration, 2004. The results of this analysis would answer questions such as: does the presence of minorities, low-income populations, or hard-to-reach communities in a certain area systematically bias the ACS estimates for that area? Would the ACS estimates be systematically larger or smaller than the CTPP estimates in seasonal areas? The regressions are ordinary least squares regressions that were estimated for most of the population and household variables listed in Table I.2. The analy- sis was done at the tract level. Systematic biases were measured to varying degrees in each of the Part 1 tables. For cer- tain tables, the bias is structural and is likely to be related to the differences in the survey instruments. For other tables, the measured differences seem to be related to sample size and would decrease as sample size increases. The residence-based tables with relatively large biases were · Disability status, · Disability status by mode to work, Table I.4. Flow-based evaluation dataset tables. Table Universe Content 1 Workers in Total Workers 2 Households Vehicles Available per Household by Mode to work 3 Means of Transportation 4 Household Income by Means of Transportation 5 Mean Travel Time by mode to work by Time Leaving Home 6 Median Travel Time by mode to work and Time Leaving Home 7 Aggregate Number of Vehicles by Time Leaving Home for Work 8 Average Number of Workers per Vehicle by Time Leaving Home 9 Aggregate Number of Carpools by Time Leaving Home for Work 10 Number of Workers per Carpool by Time Leaving Home for Work 11 Aggregate travel time by mode to work and Time Leaving Home Source: Federal Highway Administration, 2004.

OCR for page 257
Number of Workers Number of Workers 1,000 2,000 3,000 4,000 5,000 6,000 0 1,000 2,000 3,000 4,000 5,000 6,000 0 Figure I.2. Figure I.1. ACS-010100 ACS-000200 CTPP-010100 CTPP-000200 ACS-010900 ACS-001200 CTPP-010900 CTPP-001200 ACS-011800 ACS-002100 CTPP-011800 CTPP-002100 ACS-012600 ACS-002602 CTPP-012600 CTPP-002602 ACS-013400 ACS-003002 CTPP-013400 CTPP-003002 ACS-015700 ACS-003504 CTPP-015700 CTPP-003504 ACS-016500 ACS-003903 CTPP-016500 CTPP-003903 ACS-017602 ACS-004030 CTPP-017602 CTPP-004030 ACS-020300 ACS-004040 CTPP-020300 CTPP-004040 ACS-021100 ACS-004050 CTPP-021100 CTPP-004050 ACS-022600 ACS-004060 CTPP-022600 CTPP-004060 Tract Number Tract Number ACS-022902 ACS-004112 CTPP-022902 CTPP-004112 ACS-023200 ACS-004318 CTPP-023200 CTPP-004318 ACS-025403 ACS-004412 ACS versus CTPP for Selected Tracts ACS versus CTPP for Selected Tracts CTPP-025403 CTPP-004412 ACS-026003 ACS-004505 CTPP-026003 CTPP-004505 ACS-026402 ACS-004616 CTPP-026402 CTPP-004616 ACS-031100 ACS-004626 CTPP-031100 CTPP-004626 ACS-033000 ACS-004636 CTPP-033000 CTPP-004636 ACS-035400 ACS-004715 CTPP-035400 CTPP-004715 ACS-061000 ACS-004900 CTPP-061000 CTPP-004900 Estimate of Workers Driving Alone-95 Percent Confidence Intervals (Pima County, Arizona). Comparison of ACS and Decennial Census Transportation Planning Estimates Low Low High High Estimate Estimate Estimate of Workers Driving Alone-95 Percent Confidence Intervals (San Francisco County, California). 257

OCR for page 258
Number of Workers Number of Workers 258 0 1,000 2,000 3,000 4,000 5,000 6,000 1,000 2,000 3,000 4,000 5,000 6,000 0 Figure I.4. Figure I.3. ACS-860302 ACS-010301 CTPP-860302 CTPP-010301 ACS-860806 ACS-010605 CTPP-860806 CTPP-010605 ACS-860904 ACS-020206 CTPP-860904 CTPP-020206 ACS-861011 ACS-020317 CTPP-861011 CTPP-020317 ACS-861201 ACS-020501 CTPP-861201 CTPP-020501 ACS-861404 ACS-030801 CTPP-861404 CTPP-030801 ACS-861510 ACS-040502 CTPP-861510 CTPP-040502 ACS-861702 ACS-041700 CTPP-861702 CTPP-041700 ACS-862000 ACS-043100 CTPP-862000 CTPP-043100 ACS-862502 ACS-050500 CTPP-862502 CTPP-050500 ACS-862902 ACS-060112 CTPP-862902 CTPP-060112 Tract Number Tract Number ACS-863400 ACS-060206 CTPP-863400 CTPP-060206 ACS-863801 ACS-060505 CTPP-863801 CTPP-060505 ACS-864101 ACS-070204 ACS versus CTPP for Selected Tracts ACS versus CTPP for Selected Tracts A Guidebook for Using American Community Survey Data for Transportation Planning CTPP-864101 CTPP-070204 ACS-864205 ACS-070314 CTPP-864205 CTPP-070314 ACS-864402 ACS-080402 CTPP-864402 CTPP-080402 ACS-864412 ACS-090900 CTPP-864412 CTPP-090900 ACS-864514 ACS-100102 CTPP-864514 CTPP-100102 ACS-864521 ACS-110304 CTPP-864521 CTPP-110304 ACS-865501 ACS-110318 CTPP-865501 CTPP-110318 Estimate of Workers Driving Alone-95 Percent Confidence Intervals (Lake County, Illinois). Estimate of Workers Driving Alone-95 Percent Confidence Intervals (Broward County, Florida). Low Low High High Estimate Estimate

OCR for page 259
Number of Workers Number of Workers 1,000 2,000 3,000 4,000 5,000 6,000 1,000 2,000 3,000 4,000 5,000 6,000 0 0 Figure I.5. Figure I.6. ACS-000600 ACS-800100 CTPP-000600 CTPP-800100 ACS-001900 ACS-800500 CTPP-001900 CTPP-800500 ACS-002600 ACS-801101 CTPP-002600 CTPP-801101 ACS-003300 ACS-801402 CTPP-003300 CTPP-801402 ACS-003900 ACS-801602 CTPP-003900 CTPP-801602 ACS-004700 ACS-801800 CTPP-004700 CTPP-801800 ACS-005400 ACS-802300 CTPP-005400 CTPP-802300 ACS-006000 ACS-810100 CTPP-006000 CTPP-810100 ACS-006400 ACS-810412 CTPP-006400 CTPP-810412 ACS-006604 ACS-810602 CTPP-006604 CTPP-810602 ACS-006806 ACS-810800 CTPP-006806 CTPP-810800 Tract Number Tract Number ACS-007003 ACS-811102 CTPP-007003 CTPP-811102 ACS-007309 ACS-811500 CTPP-007309 CTPP-811500 ACS-007407 ACS-812000 ACS versus CTPP for Selected Tracts ACS versus CTPP for Selected Tracts CTPP-007407 CTPP-812000 ACS-007432 ACS-812401 CTPP-007432 CTPP-812401 ACS-007439 ACS-812702 CTPP-007439 CTPP-812702 ACS-007446 ACS-813000 CTPP-007446 CTPP-813000 ACS-007453 ACS-813206 CTPP-007453 CTPP-813206 ACS-007460 ACS-813403 CTPP-007460 CTPP-813403 ACS-007509 ACS-813700 CTPP-007509 CTPP-813700 Comparison of ACS and Decennial Census Transportation Planning Estimates Estimate of Workers Driving Alone-95 Percent Confidence Intervals (Douglas County, Nebraska). Low Low High High Estimate Estimate Estimate of Workers Driving Alone-95 Percent Confidence Intervals (Hampden County, Massachusetts). 259

OCR for page 260
Number of Workers Number of Workers 260 1,000 2,000 3,000 4,000 5,000 6,000 1,000 2,000 3,000 4,000 5,000 6,000 0 0 Figure I.8. Figure I.7. ACS-000502 ACS-000220 CTPP-000502 CTPP-000220 ACS-000902 ACS-000910 CTPP-000902 CTPP-000910 ACS-001400 ACS-002000 CTPP-001400 CTPP-002000 ACS-001900 ACS-002800 CTPP-001900 CTPP-002800 ACS-002402 ACS-004810 CTPP-002402 CTPP-004810 ACS-002901 ACS-005820 CTPP-002901 CTPP-005820 ACS-003401 ACS-006352 CTPP-003401 CTPP-006352 ACS-003702 ACS-006500 CTPP-003702 CTPP-006500 ACS-004101 ACS-006931 CTPP-004101 CTPP-006931 ACS-004700 ACS-007112 CTPP-004700 CTPP-007112 ACS-005500 ACS-007424 CTPP-005500 CTPP-007424 Tract Number Tract Number ACS-006200 ACS-007540 CTPP-006200 CTPP-007540 ACS-006701 ACS-007932 CTPP-006701 CTPP-007932 ACS-007202 ACS-008162 ACS versus CTPP for Selected Tracts ACS versus CTPP for Selected Tracts A Guidebook for Using American Community Survey Data for Transportation Planning CTPP-007202 CTPP-008162 ACS-008001 ACS-008380 CTPP-008001 CTPP-008380 ACS-008500 ACS-009100 CTPP-008500 CTPP-009100 ACS-009102 ACS-009332 CTPP-009102 CTPP-009332 ACS-009804 ACS-009382 CTPP-009804 CTPP-009382 ACS-010303 ACS-009520 CTPP-010303 CTPP-009520 ACS-010408 ACS-009800 CTPP-010408 CTPP-009800 Estimate of Workers Driving Alone-95 Percent Confidence Intervals (Franklin County, Ohio). Low Low High High Estimate of Workers Driving Alone-95 Percent Confidence Intervals (Multnomah County, Oregon). Estimate Estimate

OCR for page 261
Comparison of ACS and Decennial Census Transportation Planning Estimates 261 ACS versus CTPP for Selected Tracts 6,000 5,000 4,000 Number of Workers High 3,000 Estimate Low 2,000 1,000 0 ACS-000200 CTPP-000200 ACS-005200 CTPP-005200 ACS-006700 CTPP-006700 ACS-008500 CTPP-008500 ACS-011501 CTPP-011501 ACS-013200 CTPP-013200 ACS-015400 CTPP-015400 ACS-017500 CTPP-017500 ACS-020100 CTPP-020100 ACS-021502 CTPP-021502 ACS-022702 CTPP-022702 ACS-025100 CTPP-025100 ACS-028900 CTPP-028900 ACS-031800 CTPP-031800 ACS-034000 CTPP-034000 ACS-036100 CTPP-036100 ACS-037400 CTPP-037400 ACS-038600 CTPP-038600 ACS-039902 CTPP-039902 ACS-041400 CTPP-041400 Tract Number Figure I.9. Estimate of Workers Driving Alone-95 Percent Confidence Intervals (Bronx County, New York). · Tenure (owned with mortgage category), · Number of workers in household by vehicles available by household income, · Poverty status (category for incomes between 100 and less than 150 percent of poverty), and · Telephone availability. Workplace-Based Evaluation For the workplace-based estimates, the evaluation of the differences between ACS and Cen- sus place of work tables was complicated by: · The absence of the extended place of work allocation system for ACS, and · The difficulty of calculation of standard errors of place of work for Census 2000. Therefore, the evaluation of the workplace-based estimates was based on descriptive analysis of the difference between the ACS and Census estimates. We evaluated the following variables at the county and tract levels: mode to work, vehicles available by mode to work, mean travel time by mode to work, worker earnings, and mean earnings by mode to work. The general conclusions from the workplace-based evaluation are: · The differences in the estimates between ACS and Census tend to be larger as the geographic level becomes smaller due to the larger variances in the ACS estimates; and · Overall, the ACS estimates do not seem to be biased; the distributions of the differences between ACS and Census estimates are not skewed in a certain direction. However, the ACS estimates of the percentage of workers who carpooled to work and the ACS estimates of travel time to work seem to be consistently lower than the corresponding Census estimates on average.

OCR for page 262
262 A Guidebook for Using American Community Survey Data for Transportation Planning Worker Flow-Based Evaluation The worker flow evaluation datasets did not allow for a meaningful comparison of the differ- ences between ACS and CTPP, because: · The test site data were only for isolated counties, so only the worker flows with both trip ends in the counties were available; and · The disclosure limitations placed on the ACS test data were probably different than will be employed for actual future releases of the data, so conclusions about the test data are not likely to be valid. Figure I.10 is representative of the comparisons that were conducted on the worker flow-based tables. In general, the pattern of differences for travel times by mode among the comparable origin-destination flows did not reveal systematic bias in one direction or the other. 05X1: Mean Travel Time 35 Percentage of Tract O-D Pairs 30 25 20 15 10 5 0 50 0 0 0 10 0 5 5 15 25 5 -5 10 -5 -1 -2 -1 to to to to to to to to to to to to -5 0 0 25 5 10 15 00 5 0 5 -1 50 -1 -5 -2 -1 ACS estimate - Census estimate (minutes) Figure I.10. Comparison of ACS and Census 2000 worker flow data for tracts in San Francisco County.