MAKING IT USEABLE
Weighting
How the Weights Were Computed for NPTS![]()
Weighting is a procedure of adjusting the distribution of units in the sample so that the frequency attached to each unit reflects the frequency in the total population, rather than in the sample. Weighting is done on the basis of control totals from such sources as the U.S. Decennial Census of Population and the Current Population Survey. In general, weighting of any survey data can be done on selected variables only, such as population or number of households. For the NPTS, weighting is done on population estimates for the following:
Households by:
- Region
- MSA Size
- Household Size
- Travel Month
- Ethnicity
- Race
Persons by:
- Age
- Gender
- Race
- Ethnicity
- Region
- Travel Month
The weighting scheme is based on the method of sampling, which was a stratified list-assisted random sample. There is a separate weight calculated for each stratum of the stratified sample. Four weights are calculated, which are applied to the different files:
- Household weight
- Person weight
- Travel-day trip weight
- Travel-period trip weight
There are different weights for different files, because the sample changes from file to file:
- The Household File contains all usable households that responded to the survey.
- The Person File contains data for all persons in usable households who provided valid responses -- in some households there will be some people who did not respond, while all people in some households responded. The Person Weight is adjusted to account for people who did not respond in the household.
- The Travel-Day File contains data for all travel-day trips reported for complete person interviews within usable households.
- The Travel-Period File contains data for all those travel-period trips reported in complete person interviews within usable households.
![]()
![]()