Box A - The New National Freight Composite Estimates

Box A - The New National Freight Composite Estimates

The largest single data source for estimating U.S. freight activity is the Commodity Flow Survey (CFS). This survey covers a large proportion of the nations domestic and export freight movements associated with manufacturing, mining, and wholesale trade. However, it does not capture all of the freight that moves on the U.S. freight system because many economic activities are not covered. In order to report on the current state of freight shipments in America and to describe the major changes that are taking place in U.S. freight transportation, this report makes use of several other data sources to provide a more complete snapshot of the nations freight activities in 2002 (the year of the most recent CFS). Where 2003 and 2004 data are available, such as goods imports and exports and overall ton-miles, this report also uses these data to describe recent freight activity.

The composite estimates presented in this report are the result of a joint effort by the Research and Innovative Technology Administration (RITA), Bureau of Transportation Statistics (BTS) and the Federal Highway Administration (FHWA), Office of Freight Management and Operations, to develop a more complete picture of the nations commercial freight shipments, including all economic sectors that handle freight in one way or another (see exhibit A). This fuller composite picture draws on the CFS data and non-CFS freight data from several sources for economic sectors not covered in the CFS, such as retail, services, construction, and households goods movements — that traditionally are not perceived as freight producers but that do handle freight in their daily operations. It also includes shipments of agricultural products from farms to processing plants, logs and rough wood, fishery products, crude petroleum, and municipal solid waste. The new composite estimate is different and larger than the BTS estimates published in the Freight Shipments in America report, which used preliminary 2002 CFS data, not the final data used here, and does not include the non-CFS shipments discussed above.

Some data gaps still exist in the national freight picture. The new estimates do not include transshipments, such as shipments from Canada that are transported on U.S. highways to Mexico but do not officially enter into the U.S. economy. With the exception of municipal solid waste, the estimates do not include government shipments. The composite estimates, using the current methods, were developed for 2002 only. Because the additional complementary data have not been assembled for 1993 and 1997, the other CFS years, the report uses only the CFS data when discussing changes in freight movements by type of transportation modes used in moving the nations freight, the kinds of commodities moved, and the distances traveled.

In order to make comparisons by mode, commodity, and other freight characteristics and determine, for example, which commodities are shipped the most, BTS performed statistical significance testing on CFS data for freight value, weight, and ton-miles.1 It was not practical to do this testing for the non-CFS estimates because they were drawn from different sources and some of the estimates are modeled data with many assumptions that cannot be tested statistically.

Throughout this report, a statistically significant difference between two different entities in the CFS (numbers, groups, classifications, categories, etc. developed from a sample) is measured at the 10 percent level. This provides a 90 percent level of certainty about CFS estimates. That is, if we were to repeatedly make new estimates using exactly the same procedure (by drawing a new sample, conducting new interviews, calculating new estimates and new confidence intervals), the confidence intervals would contain the average of all the estimates 90 percent of the time.

1 The BTS Office of Survey Programs performed the statistical significance testing for this report, using statistical reliability methods based on Raj and Chandhok 1998.