1. The National Household Travel Survey (NHTS) Add-On Update
Huiwei Shen, FDOT Systems Planning

2. Florida’s Participation in the NHTS Add-On Program
Florida Model Task Force recommended participation in the National Household Travel Survey (NHTS) Add-On program
MPOAC agreed to use 2.6 million PL fund for 12,000 samples at $175 per sample plus $500K for data analysis
FDOT funded $350K for rural samples across the state
The NHTS data are used primarily for gaining a better understanding of travel behavior. The data enable DOT officials to assess program initiatives, review programs and policies, study current mobility issues, and plan for the future.
The NHTS is a tool in the urban transportation planning process; it provides data on personal travel behavior, trends in travel over time, trip generation rates, national data to use as a benchmark in reviewing local data, and data for various other planning and modeling applications. The transportation research community, including academics, consultants and government, use the NHTS extensively to examine:
travel behavior at the individual and household level;
the characteristics of travel, such as trip chaining, use of the various modes, amount and purpose of travel by time of day & day of week, vehicle occupancy, and a host of other attributes;
the relationship between demographics and travel; and
the public’s perceptions of the transportation system.
People in various fields outside of transportation use the NHTS data to connect the role of transportation with other aspects of our lives. Medical researchers use the data to determine crash exposure rates of drivers and passengers, including the elderly, who have heightened morbidity and mortality rates. Safety specialists study the accident risk of school-age children, particularly when they are traveling on their own by walking or biking. Social service agencies need to know more about how low-income households currently meet their travel needs.
The Nationwide Personal Transportation Survey (NPTS) and American Travel Survey (ATS) have served as the nation’s inventory of daily and long distance travel. The NHTS combines and replaces these two surveys. Data is collected on daily trips taken in a 24-hour period, and includes:
purpose of the trip (work, shopping, etc.);
means of transportation used (car, bus, subway, walk, etc.);
how long the trip took, i.e., travel time;
time of day when the trip took place;
day of week when the trip took place, and
if a private vehicle trip:
number of people in the vehicle , i.e., vehicle occupancy;
driver characteristics (age, sex, worker status, education level, etc.); and
vehicle attributes (make, model, model year, amount of miles driven in a year)
These data are collected for:
all trips, all modes, all purposes, all trip lengths, and all areas of the country, urban and rural.
In addition to the data above, the 2001 NHTS also collected additional data on trips to a destination 50 miles or more from home (long-distance travel) that started from home and ended at home during a four-week travel period. Data collected on long-distance trips included all the data mentioned above for daily trips with the exception of travel time and the time of day the trip took place. Data on the location of overnight stops and access/egress to an airport, train station, bus station or boat pier were also collected. If no long-distance trips were taken during the four-week period, a series of questions were asked regarding the most recent long-distance trip prior to that four-week travel period.
The 2001 NHTS data set includes, but is not limited to:
household data on the relationship of household members, education level, income, housing characteristics, and other demographic information;
information on each household vehicle, including year, make, model, and estimates of annual miles traveled;
data about drivers, including information on travel as part of work;
data about one-way trips taken during a designated 24-hour period (the household's travel day) including the time the trip began and ended, length of the trip, composition of the travel party, mode of transportation, purpose of the trip, and the specific vehicle used (if a household vehicle);
data describing round-trips taken during a four-week period (the household’s travel period) where the farthest point of the trip was at least 50 miles from home, including the farthest destination, access and egress stops and overnight stays on the way to and from the farthest destination, mode, purpose, and travel party information;
if no long-distance trips were made during the four-week travel period, data on the most recent long-distance trip by any mode and the most recent long-distance train trip;
information to describe characteristics of the geographic area in which the sample household and workplace of sample persons are located;
data on telecommuting;
public perceptions of the transportation system;
data on Internet usage; and
the typical number of transit, walk and bike trips made over a period longer than the 24-hour travel day.

3. Sampling Plan Design Developed in consultation with FHWA NHTS team
Reviewed and approved by MTF Data Committee
Conforms to existing FDOT boundaries
Meets urban and regional modeling needs
Original plan based on 2000 Census Households
Revised plan guarantees that minimum sample size per region is 1,200
Allows for meaningful comparisons
A primary goal of most travel surveys is to collect data that allow analysts to estimate a range of travel-related and socioeconomic parameters that are often used to develop travel demand models. Ideally, the survey output would be error-free, so that the estimates derived from the sample would reflect the true value of the parameter in the population and could be used confidently in later analyses.
In collecting data from a sample of the population, measurement error can be divided into two types:
Sampling errors; and
Non-sampling errors/biases.
The “total error” of a parameter estimate derived from a survey sample reflects both the sampling and non-sampling errors in the collected data. Sampling errors are the random errors that are introduced into the survey simply because not every member of the survey population is included in the drawn sample.
Sampling errors reflect the potential variability between the estimate of a parameter in the sample and its true value in the population. These errors affect the precision of the survey results.
Non-sampling errors are the assortment of problems that can occur during the survey design and data collection stages which may cause survey measures and parameter estimates to be systematically incorrect. Nonsampling errors reflect how well the information is collected, and include non-response biases often caused by refusals and response biases which reflect a systematic distortion of survey responses. These errors affect the accuracy of the survey results. As a result, it can be stated that the goal of sampling is to:
Reduce sampling errors that can cause the parameter estimates and other measures to be imprecise; and
Reduce non-sampling errors or survey biases that can cause the measurements to be inaccurate.
Reliability which is a common social science term, refers to the level of precision and sampling error in the collected data. The term validity refers to the accuracy level and bias of the data. Figure 3 shows the concepts involved in arriving at a sample from the population.
Justification for 1,200 vs 1,000
Allows for more samples when segmenting by HH size or auto ownership
Ensures that weights derived are appropriate and not based on single or two observations per cell. The goal is to have a sufficient number of observations to support the number reported in any important cell in the expanded tabulations. An estimate of 1,000 single person household in a cell is probably not helpful if it is derived from only two underlying observations with sample-expansion weights that average 500!
Donor Regions gave approximately 5% of their samples to D1 and 3

4. Sampling Plan
Region 1 – Collier, Lee, Charlotte, Sarasota, and Manatee (Census 2000 Households : 545,158 )
Region 2 – Duval, St. Johns, Clay, Nassau, and Alachua (Census Households : 613,890 )
Region 3 – Leon, Gadsen, Wakulla, Walton, Okaloosa, Santa Rosa, Escambia, and Bay (Census 2000 Households : 424,149 )
The NHTS Add-on for Florida will be a survey of the civilian, non-institutionalized population of
Florida. Sampling will be done by creating a random-digit dialing (RDD) list of telephone numbers. An eligible household excludes telephones in motels, hotels, group quarters, such as nursing homes, prisons, barracks, convents, or monasteries and any living quarters with 10 or more unrelated roommates.
Telephones in dorm rooms, fraternity and sorority houses will be eligible provided the residence had less than 11 household members. Therefore, students who normally reside at school but were living at home for the summer will not be considered household members at their parent's home.
Household members include people who think of the sampled household as their primary place of residence. It includes persons who usually stay in the household but were temporarily away on business, vacation, or in a hospital. It will not include people just visiting, such as a college student who normally has been living away at school.
This survey will be designed as a list-assisted random digit dialing survey, to yield an equal probability sample of households with telephones. Special procedures will be developed to reach households where cell phones are the primary phones. The target sample size is 12,000 completed households for the Florida add-on.
The sample frame will be consisted of all telephone numbers in 100-banks of numbers in which there was at least one listed residential number. A 100-bank is a set of 100 telephone numbers with the same first eight digits, that is, the same area code, exchange, and the next two digits. Each quarter, a new sampling frame will be constructed and sample will be selected for use until a new sample was drawn.
While a systematic or stratified sampling method is suggested, the final sample selection method will be determined in consultation with DOT staff and survey contractor. Weights to expand the survey to the population will also be developed in consultation with DOT staff and survey contractor.

5. Sampling Plan
Region 4 – Broward, Miami-Dade, Palm Beach, Indian River, St. Lucie, and Martin (Census 2000 Households : 2,150,666 )
Region 5 – Orange, Osceola, Seminole, Lake, Sumter, Brevard, Volusia, and Marion (Census 2000 Households : 1,040,095 )
Region 7 – District 7 (except Citrus) and Polk (Census 2000 Households : 1,196,954 )
Region 8 – Rural counties not in the first six (Census 2000 Households : 370,209 )

6. Confidence Interval (95% CI)
Sampling Plan
Sample Size
Confidence Interval (95% CI)
Region
Number of Households
Original
Revised 1
545,158
1,096
1,200
± 3.0%
± 2.8% 2
613,890
1,234 3
424,149
852
± 3.4% 4
2,150,666
4,322
4,116
± 1.5% 5
1,040,095
2,090
2,000
± 2.1%
± 2.2% 7
1,196,954
2,406
2,250
± 2.0% 8
370,209
Total
6,341,121
14,000
± 0.8%

7. Sampling Plan Region Sample Size Region 1 1,200 Region 2 1,224
4,116
Region 5
2,000
Region 7
2,250
Region 8

8. Sampling Plan MPO Sample Size
Each MPO will be sampled such that at least 80% of the sampled households from the MPO are included in the region sample
Why do this?
Ensures smaller MPOs are sufficiently sampled
Makes data collection more cost effective

9. Rural Samples FDOT District Total Households Sample Size
Confidence Interval (95% CL) 1
83,828
453
± 4.6% 2
125,436
678
± 3.8% 3
51,894
280
± 5.9% 5
21,284
115
± 9.2% 6
35,106
190
± 7.1% 7
52,661
284
± 5.8%
Rural Total
370,209
2,000
± 2.2%

10. 2001 NHTS Sample for Florida
Region
Sample Size
Region 1
136
Region 2
158
Region 3
111
Region 4
401
Region 5
240
Region 7
295
Region 8 86
1427 Total

11. Additional Questions General Questionnaire Improvements
Elder Travel
Transit Access
Transportation Funding Preference
Additional Questions for Florida
Residential Location Decision
Seasonal Residents
Transit Preference to Supplement General Questionnaire

12. NHTS Status Awarded contract to Westat Completed Florida sampling plan
Finalized Florida questionnaire
Still need $$$ for the national samples
Pretest December/January 2008 and full data collection March 2008
Setup reporting system on internet