Box E - Source and Reliability of the CFS Data and Other Sources Used in this Report
The CFS data presented in this report are estimates derived from a sample survey. Because the CFS is a sample survey and uses estimation methods and procedures to estimate total freight activity for the entire United States, the data are subject to sampling and nonsampling errors. Sampling errors occur because one set of businesses sampled in a survey may have different characteristics from another sample set of companies selected from the entire universe of businesses. Nonsampling errors occur because of all other factors that may contribute to the total errors in a sample survey, such as nonresponse, response errors, and differences in interpretation of questions by respondents. See the CFS source cited below for detailed discussion of reliability of the CFS data and estimates of standard errors.
The supplementary data on crude petroleum, logging, imports, and farm shipments used in this report are not based on surveys and therefore are not subject to sampling errors. However, the data are subject to nonsampling errors that may also occur in censuses and other forms of data collection. The impact of these nonsampling errors on the supplementary data is not fully quantifiable. Clearly, because the supplementary data did not cover all out-of-scope industries and government shipments, the estimates of overall national freight activity are incomplete.
In this report, the BTS estimates of the value, tons, and ton-miles of crude petroleum and petroleum products shipments by pipelines are based on pipeline data from the Federal Energy Regulatory Commission (FERC). BTS converted FERC information on barrels of petroleum and petroleum products transported into tons and ton-miles. Estimates of waterborne shipments not captured in the CFS are based on information from the U.S. Army Corps of Engineers (USACE) and the Census Bureau's Foreign Trade Division (FTD). The USACE reports data on the tons and ton-miles of domestic and foreign waterborne commerce. The FTD reports information on the value and tons of waterborne U.S. international trade. BTS used data from both the USACE and Census Bureau to develop a set of values, tons, and ton-miles for all domestic and foreign waterborne freight.
This report also compares the CFS data with data from the Surface Transportation Board's Carload Rail Waybill Sample. The preliminary analysis of rail freight activities reported by the CFS and the Waybill shows some differences in the overall change in volume of freight moved between 1993, 1997, and 2002. While the CFS shows a slight increase in rail tonnage, the Waybill shows a larger increase in rail tonnage over the same period. These differences for rail single mode and intermodal truck and rail combination are being analyzed, and BTS will present any findings after additional processing of the CFS during 2004.
Additional information on (1) comparability of 2002 CFS with the 1993 and 1997 CFS, (2) reliability of the CFS estimates, and (3) sample design, data collection, and estimation is available at http://emedjimurje.info/cfs/prod.html.
SOURCE: U.S. Department of Transportation, Bureau of Transportation Statistics and U.S. Department of Commerce, U.S. Census Bureau, 2002 Economic Census: Transportation Commodity Flow Survey, Preliminary Report, December 2003, EC02TCF-US(P).