255 



misuse of the data for the NPS's purpose. All knowledgeable statisticians, after even a 

 cursory analysis of this particular process, would agree with this conclusion. 



Because, as indicated above, the fundamental inference upon which the NPS based 

 its FFA is statistically invalid, the conclusion based thereon, that a twelve (12%) percent 

 franchise fee will cause FST's profitability to equal or approach the average profitability of 

 all other businesses nationwide classified under the same SIC code classification, is also 

 statistically invalid and completely inaccurate and unacceptable procedurally. In fact, no 

 valid inferences regarding the general population of businesses, or any particular business, 

 including any concession operation classified according to any SIC code, may be drawn 

 from the D&B Industry Norms or the RMA Statement Studies discussed above. 



Additionalh'. the D&B Industry Norms data simply has well known deficiencies 

 which are both irrefutable and insurmountable statistically, including the following: 



a. The samples of businesses from which the D&B infonnation is drawn are not 

 random samples. According to the most basic statistical principles, if one is 

 to draw any inference about a national population of businesses classified 

 under the same SIC code, based upon samples recei\ed only from a relativelx 

 small number of businesses who \oluntarily submit their financial data, any 

 such inference is basicaih fiawed because the sample businesses are not 

 randomly selected. The voluntary convenience samples so used are 

 tlindamentally not considered representative of all other businesses classified 

 under the same SIC code number, and information gathered in this manner is 

 therefore considered invalid and unreliable for such purposes as those used by 

 NPS. This well known D&B' Industrv Norms data deficiencv is either 



