MAKING IT USEABLE
   Points to Remember		      

- For the NPTS, a conservative approach to editing is used. Data is changed in the edits only when it is very clear that the data reported could not have happened (e.g., walk trip averaging 65 miles per hour.)

- For some applications, particularly in travel demand models, the user may want to develop and use stricter edits.

- WEIGHTED DATA MUST BE USED. If the unweighted or sample data is used, the results will look like travel in New York, Massachusetts and Oklahoma, because of the oversampling in add-on areas.

- There are four weighting factors:

  1. Household Weight (WTHHFIN) - used for households or vehicles
  2. Person Weight (WTPERFIN) - used for persons
  3. Travel Day Weight (WTTRDFIN) - used for travel day trips and segmented trips
  4. Travel Period Weight (WTTRPFIN) - used for travel period (75+ miles) trips

- Use the weight that is on the file you are tabulating - for example, Workers by Age, Sex and Main Means to Work would be run from the Person File, thus the Person Weight would be used.

- When running a table using data from more than one file, use the weight for the variable that is in the cells. Examples:

  1. Workers with Bus Available by Age, Sex, and Main Means to Work - workers are the value in the cells of this table, so the Person Weight would be used
  2. Travel Day Trips by Worker Status and Means of Transportation - travel day trips are in the cells, so the Travel day trip weight would be used

- When computing a rate, like Travel Day Trips per Person, you need to weight the two elements of the rate then divide the two weighted values. Here weighted travel day trips would be divided by weighted persons.

- The NPTS data are not valid at the individual state or metro area, unless an add-on sample was conducted.

- The data are valid by the sample strata, which are the Census Division, the metro area status, and the presence of a subway system. For example, the NPTS data for Columbus, Ohio would not accurately reflect travel in Columbus. However, data tabulated by the East North Central Division- areas of 1 million or more without rail, gives the user some parameters of what travel is like for that area of the country and that size city.

- Remember that codes for legitimate skip (999997), refused (999998), and not-reported (999999) must be dealt with appropriately when the user writes their own code. These special codes are dealt with appropriately with the on-line analysis module on the web site.

This is the end of the Making It Useable module. The next module, Analysis, provides examples showing how to mine the dataset using SAS, the Web Analysis Engine, or database analysis languages.