256 



overlooked or disregarded by NPS, but this data deficiency is clearly based 

 upon well known and basic principles in the field of statistics. 



b. Additionally, I call into question the method of data collection employed by 

 the compilers of D&B Industry Norms. According to reliable sources inside 

 D&B, much of the data is collected through telephone interviews by 

 transcribers, rather than through more standard acceptable procedures such as 

 thorough financial statement analysis, detailed corporate personal 

 interviewing, and/or rigorous data collection through financial audit. I 

 therefore question the reliability of the D&B Industry Norms data because it 

 is improperly collected and potentially inaccurate on its face for this purpose, 

 or any purpose in which a rigorous statistical methodology is employed and 

 the D&B Industry Norms data is utilized as the underlying foundation for the 

 analysis. 



c. Moreover, based upon information I consider reliable, the entire population of 

 businesses which could be properly classified under SIC number 4489, is 

 approximately 3.000 nationwide. The sample sizes reported by D&B Industry' 

 Norms under SIC 4489. for FST, were 16 in 1988 and 56 in 1989. and the 

 sample size employed in the RMA Statement Studies used by the NPS was 10. 



Clearly, even if these samples had been randomly and properly selected as 

 required, they are too small to draw any valid inference about the larger 

 population of other businesses under the same SIC code, because the sample 

 sizes are too small relative to the vast number of other businesses classified 

 under the same SIC code. 



