2001 NHTS Version 4.0 Release Notes
July 2005

Version 4.0 of the 2001 NHTS is being released to accomplish several objectives, including:

• Adding an abridged version of the Long Trip File to the dataset
• Correcting errors found on the file
• Making like data from different files agree with each other
• Making slight revisions to confidentiality filters
• Revising WHYTRP90 trip purpose variable
• Cleaning up discrepancies between the variables regarding the first trip of the day (WHYFROM, AWAYHOME, FRSTHOME)
• Adding new variables for trip purpose (TRIPPURP), population size of CMSAs/MSAs of 1 million or more (MSAPOP), and vehicle age (VEHAGE)

In addition to adding the abridged Long Trip file, there were a number of changes made to the Household, Person, Vehicle, and Day Trip files.  To help dataset users determine whether these changes affect data they typically tabulate, the listing below shows all changed variables, the files they appear on, and where the change is described in this documentation:

1. Long Trip File

The long trip file (also referred to as the travel period file) contains long distance trips of 100 miles or more roundtrip. The Version 4 long trip file contains core variables from the data set found on the BTS CD-ROM. To obtain the full long trip file from BTS go to  www.bts.gov/pdc and, once opened, search for NHTS.

In addition to reducing the amount and detail of each long trip, the household and person level variables in the long trip file were updated to reflect values in their respective files.

The definition of a valid NHTS long trip is “a trip where the furthest destination was 50 miles or more from home with the return home trip during the 4-week period preceding and including the household’s assigned Travel Day.” Unlike the Travel Day file, which contains one-way trips, each long trip on the Travel Period file is a roundtrip in the sense that the respondent had to go to a destination and return home in order for the trip to an eligible long trip.

On the file, the long trips are split into two parts, with FARxxxx or TPxxxxxx  designating the variables that apply to the trip to the destination and the time at the destination, and RETxxxxx designating variables for the return home trip. Splitting the two halves of the trip was necessary to account for trip characteristics that differed between the two, such a mode used, number of people on the trip, etc.  

Other naming conventions used in long trip variables include:

ACCxxxx – to describe access to an airport, bus station or bus stop, train station, or pier.  ACCxxxx variables provide data on the trip between home and the place where you got on the airplane, bus, train, or ship.

EGRxxxx – to describe egress from an airport, bus station or bus stop, train station, or pier.  EGRxxxx variables provide data on the trip between the place where you got off the airplane, bus, train, or ship and your actual destination.

Any one trip can have an access and egress on the way to the destination and on the return home.

FARxxxx – variables describe characteristics of the place you traveled to.  If there were multiple places you went, the one furthest from your home location was considered to be the “destination”.   

2. Updates to Other Files related to Long Trips

The four other files (Household, Person, Vehicle, and Travel Day) have been updated to include BEGTRAV and ENDTRAV, which contain date information on the beginning and end of the household’s assigned  travel period.  In addition, the variable CNTTPHH (Sum of all travel period person trips) on the Household file has been updated to reflect the long trip file.  In the Person file, CNTTPTR (Sum of travel period person trips) and CNTTPUNQ (Number of unique travel period trips) were also updated in accordance with the long trip file.

3. Access and Egress Variables

For a small number of trips where ACCDIV, ACCREG, EGRDIV, or EGRREG were =0,  their respective variables were changed to -9 (not reported.) This impact 4 access trips and 18 egress trips.

• ACCST or EGRST with values of PR, VI, or AP were set to -9
• Some of the values of  ACCST or ACCMSA were originally XX or XXXX when the correct value should have been –1, and these were changed to –1 to agree with  ACCDIV.   Similarly, ACCST or ACCMSA was set to -9 if they equaled -1, -8, XX or ZZ and ACCDIV=-9.

4. MSACAT, CDIVMSAR, and RAIL*

A correction was made to MSACAT, CDIVMSAR and RAIL.  It affects only those households located in the 17 CMSAs (Consolidated Statistical Metropolitan Areas).  Background– MSACAT is composed of four categories:

• MSA of 1 million or more, with rail (subway)
• MSA of 1 million or more, without rail
• MSA less than a million
• Not in MSA (CMSA).

If the household is in an area that is a CMSA, the area must be at least a million in population, and the CMSA is composed of PMSAs (Primary Statistical Metropolitan Areas). For example, if you live in the Chicago CMSA, you could live in one of four PMSAs: Chicago IL, Gary IN, Kankakee IL, or Kenosha WI. 

Originally the coding of MSACAT, CDIVMSAR, and RAIL was done at the PMSA level.  In the case of a Chicago household, three of the four component PMSAs do not have rail (subway or heavy rail). However, the coding was later corrected to be at the CMSA level, so all households within the Chicago CMSA are now classified as in an ‘MSA of a million or more with rail’.  This was done because a resident of Kenosha WI may decide to go or not go into Chicago based on the availability of a subway (actually in this case, an elevated rail system) in Chicago.

5. PRMACT - Primary Activity last week*

This variable represents the Primary Activity of the respondent in the week preceding travel day (e.g., working, going to school, retired).  In reviewing the data file we found that, for a certain number of respondents, the PRMACT values had been changed to –1 (appropriate skip) on their Person file record, but there was a different category on their Travel Day Trip record.  Persons with R_AGE values that were reserve codes (-1, -7, -8, -9) were incorrectly given values of -1 in the Person file.  This has been corrected.

6. HHR_RACE

Because an estimate from the national data set of the number of blacks in the US differed from the control totals used for weighting the dataset by about +500,000 (control total is 33,368,322), a problem was discovered in the way that HHR_RACE was determined.  The problem affected 364 households across the final NHTS dataset. The main problem occurred in the logic used to identify whether the household respondent was black (differentiating people who identified themselves as black-only from people with a combination of race codes that included black). This problem was how the computer-assisted telephone interviewing system logic handled cases where, during the screener interview, the household respondent answered "no" to some of the race classifications in Question C7 but then refused to answer the remaining race categories.  As a result, too many households were processed as black.  The problem did not affect weights in any way, they are all correct. IMPTRACE was also modified in response to this issue.

7. Confidentiality Changes

Values of the variables HHSTATE, HHSTFIPS, WKSTFIPS and WORKSTAT that reflected Hawaii had confidentiality filters removed because it was an add-on state. Note that Hawaii was the only add-on area identified that is below our standard confidentiality thresholds (i.e., CMSA/MSA of 1 million or more and state of 2 million or more).  Identifying other add-on areas that were below the threshold would have potentially compromised respondent confidentiality. 

8. Inter-file Consistency

Several changes were made to ensure that data in one file matched data from all other files. The following variables were updated to match the Household file: LIF_CYC and R_RELAT in the Person and Travel Day files; DRVRCNT Person, Travel Day, and Vehicle files; HHVEHCNT, NUMADLT, and WRKCOUNT Travel Day and Vehicle files; and HHFAMINC in the Vehicle file.  In addition, OCCAT and R_AGE in the Travel Day file were changed to match the Person file.

TRPTRANS for trips by POV was changed in the Travel Day file to match VEHTYPE in the Vehicle file.  After researching discrepancies between TRPTRANS and VEHTYPE, it was determined that VEHTYPE was more often correct, so TRPTRANS was changed to reflect that, and make the two files consistent.  Note that this change to TRPTRANS will impact the relative amount of trips and miles made by car, van, SUV, pickup, other truck, RV, and motorcycle.

Comparison of Vehicle Trips and VMT by Vehicle Type, Version 3 versus Version 4 (ALL NUMBERS IN 000,000s)

NOTE: The slight changes in the total vehicle trips and VMT was due to the corrections made for underage drivers, a description of which follows this table. 

9. Trips by Underage Drivers

In Version 3 (January 2004 release) of the 2001 NHTS data, there are 62 trips driven by people under the age of 15.  The designation of who drove on these trips is believed to be in error, and the trip data has been adjusted using the following variables: DRVR_FLG, PSGR_FLG, WHODROVE, VEHUSED, ONTD_P1-ONTD_P5, HH_ONTD, NONHHCNT, and NUMONTRP.  In addition, R_AGE set to -9 changed for three individuals.

10. Vehicles with Primary Drivers under 15

Fifteen vehicles were identified with primary drivers less than 15 years of age.  For these vehicles, the value of WHOMAIN was set to -9, and the value of PRMDRVR1 for these persons was set to -1.

11. Summary Variables in the Household File

The summary variables in the Household file were also updated to match values in the Person file.  These variables include AGE_P1 through AGE_P14, DRV_P1 through DRV_P7, INCM_P1 through-INCM_P7, REL_P2, SEX_P1 and SEX_P6.

12. R_AGEWGT

In the person file, persons surveyed by Morpace (SMPLFIRM=02) were incorrectly assigned an R_AGEWGT value equivalent to the value of R_AGE for the Household Respondent.  This value did not affect weighting in any way, and has been corrected to reflect the value in each individual respondent’s R_AGE value.  Also, a handful of persons with R_AGE= -9 were previously weighted using an R_AGEWGT of 14, and the Version 4 data has been updated to reflect such values.

13. WHYTO and WHYTRP01 = -1

Version 3 (January 2004 release) of the 2001 NHTS has 261 travel day trips with values of -1 (Legitimate Skip) for WHYTRP01 and WHYTO.  We were unable to determine a situation where these variable were legitimately skipped, so they have been corrected using other information from the survey contractor.  Most of these trips were “To Home,” so their values of WHYTO and WHYTRP01, as well as values of WHYTRP1S were changed accordingly.  For non-home trips, the value of WHYTRP90 was also corrected to reflect the new WHYTO and WHYTRP01 values.

14. WHYFROM, AWAYHOME, and FRSTHM

AWAYHOME (Reason start travel day away from home) should only have a value for the first trip of the travel day.  However, 242 records that are not the first trip of the travel day had a positive value- for AWAYHOME. These records now have a correct AWAYHOME value of -1.

The variable AWAYHOME in the Travel Day file, and FRSTHM (At home at start of travel day) in the Person file, were determined to be inconsistent in a handful of cases.  The following fixes were made:

• If FRSTHM=1 (Yes) and AWAYHOME NOT = -1 (Appropriate Skip), then one of two changes was made.  If WHYTO of the first trip of the travel day was 01 (Home), the value of FRSTHM in the Person file was changed to 2 (No).  Otherwise, AWAYHOME was set to –1 (Appropriate Skip). If FRSTHM=2 (No) and AWAYHOME = -1 (Appropriate Skip) and WHYTO=01 (First trip of the day was to home), then AWAYHOME was set to –8 (Don’t Know).

In addition, the value of WHYFROM was corrected to 01 (Home) for the first trip of the travel day, unless an actual AWAYHOME value existed.  In that case, the value of AWAYHOME was assigned to WHYFROM for the first trip of the day.

15. WHTRP90

WHYTRP90 was recomputed using the original specifications.  This resulted in the changes documented below:

Comparison of Person Trips by WHYTRP90, Version 3 versus Version 4 (ALL NUMBERS IN 000,000s)

16. New Variables

TRIPPURP - Trip Purpose
Using Home-based and non-home based categorization, the variable TRIPPURP was created.  This variable is intended to reflect the purpose categories most often used in transportation planning by states or metropolitan areas.  The coding reflects the 1990 values of the variable of the same name. The TRIPPURP codes are:

1 = Home-based Work
2 = Home-based Shopping
3 = Home-based social/recreational
4 = Home-based other
5 = Not home-based

MSAPOP – 2000 Census population of CMSA or MSA
Population of the MSA/CMSA was added for areas with populations over 1 million.

VEHAGE – Vehicle Age
Vehicle age was computed using the variable VEHYEAR.  If VEHYEAR had a value of 2001-2003, VEHAGE was assigned a value of 1.  If VEHYEAR had a non-reserve code value (i.e., the value was NOT -7, -8 or –9), VEHAGE was computed using 2002- VEHYEAR.

17. Miscellaneous Changes/Fixes

• 4 persons in the Person file had values of VI for WORKSTAT, while their HHSTATE was VA.  We believe the WORKSTAT values are typos and thus they have been changed to VA.
• For records where WORKSTAT had reserve code values (i.e., the value was -1, -7, -8 or –9) and WKSTFIPS did not have a reserve code, the value of WKSTFIPS was translated into character form and assigned to WORKSTAT.  The reverse was also done for records where WKSTFIPS had reserve code values and WORKSTAT did not have a reserve code value.
• VEHID and VEHTYPE from the Vehicle file were added to the Travel Day file, for trips using a household vehicle. This corrects differences found in the vehicle type between the two files.  After extensive checking, it was determined that the Vehicle Type in the Vehicle file was more accurate than the corresponding TRPTRANS of 01-07 (private vehicle codes) in the Travel Day file, so the Travel Day data on TRPTRANS was changed to reflect the better description in the Vehicle file.
• WHOMAIN in the Vehicle file was updated to add a leading zero where necessary to make for easier merging with the Person file.
• Claritas variables (those starting with HB or HT)  were switched to numeric instead of character format in order to lessen confusion on alignment issues.  Also, the label for HTEEMPDN was changed from “Jobs per square mile” to “Workers per square mile”
• Changes were made to ONTD_P1 through ONTD_P9 where a person previously identified as on a given trip was determined not to be on that trip.
• Households in the Oahu add-on area (SMPLAREA=07) were given a value of 3320 (Honolulu) for HHC_MSA in all files.
• There were some vehicles that showed a BESTMILE value of 0, but ANNMILES was not 0.  After researching these cases it was decided to change the BESTMILE value  to the value of ANNMILES. Note that the BESTMILE value was also assigned to the value of ANNMILES for cases where VEHTYPE was Other Truck, RV, Motorcycle, or Other.   Flags and other related variables (BEST_FLG, BEST_OUT, BEST_EDT, ANN_FLG, ANNUALZD, ANULZDSE) were also edited to agree with BESTMILE and ANNMILES for these vehicles.
• For a small number of MSAs, the values of MSASIZE, MSACAT, and CDIVMSAR did not match,  Those variables were changed as appropriate.
• Selected values of HHC_MSA, HHSTFIPS, WKSTFIPS, R_AGEWGT, MAKECODE, MODLCODE, FARMSA, ACCMSA, ACCST, EGRMSA, and EGRST were set to missing to ensure consistency with NHTS data collection and preparation procedures.