EDI integrated into accounting software
Friday, April 6, 2012, 8:00 AM
EDI, or Electronic Data Interchange, has been integrated into accounting software nation-wide. In order to carry out EDI integration properly, it is crucial to understand the methods of EDI Data Collection.
The data that we will examine today was collected from studies of the US shipping industry and TL general freight motor carriers. Samples were taken by querying the National Motor Carrier Directory and the Official Directory of Industrial and Commercial Traffic Executives. Every single firm identified as a TL general freight carrier in the National Motor Carrier Directory was sent a mail survey. This was more than three hundred and sixty firms. At the same time mailers were sent to a comparably-sized group of companies identified as carriers or shippers in the most recent Official Directory of Industrial and Commercial Traffic Executives. Since there are many, many more carriers and shippers than TL general freight carriers, every twenty-fifth corporation was selected. This ensured that the 350 companies chosen were evenly distributed from A to Z within the sample frame.
Each subject received a survey in the mail, complete with a cover letter and a self-addressed stamped envelope. Three weeks later they were sent another mailing, nearly identical to the firstw as a reminder. Twenty-five of the surveys sent to carrier companies were returned as undeliverable or inapplicable. For example, if the company does not use the truck mode, there is no reason to include them in the survey. Therefore the effective sample size was 335 companies. 57 of those companies gave usable responses to the survey package, for a 17% response rate. With shippers another 25 companies were returned or ineligible, which leaves an effective sample group of 325. 65 survey packages were returned with usable responses, for a response rate of 20% A higher response rate would have been preferred, but is impractical. The carrier sampling frame encompasses all known TL general freight motor carriers, so the sample size is solidly fixed. Because the carrier sampling frame is unchangeable, it was reasonable to work with a sample of similar size from the shippers. The statistical gulf between the two groups means that relatively tiny differences for one group can appear disproportionately large in the other.
Non-response bias was tested in both groups by comparing respondents from the first mailing to the second. It is speculated that late respondents have more similarities to non-respondents than the early respondents do. Surveys that returned within three weeks were tested against those that took longer. Calculated EDI variables indicate that there is only one statistically significant difference between early-responding shippers and late-responding shippers, and none whatsoever between carriers. Therefore non-response bias may be reasonably surmised to be minor.
The methodology of EDI data collection is thorough and sound. Excellent data collection techniques are the foundation of EDI integration.