509 



Attaching economic values to the results of future basic 

 research ; 



Equating dollar values with social or esthetic values ; 



Equating the cost/eif ectiveness of basic and applied researches ; 



Identifying the total cost/benefit factors of a new or future tech- 

 nological application ; 



Identifying all impacts of a given technology on the environ- 

 ment ; 



Justifying a particular ceiling on level of scientific effort : and 



Eliminating all possibility of error, so as to achieve a 100-percent 

 probability. 



Much effort can be consumed unfruitfully in the quest for answers 

 to the questions implied in this list, or others of like nature. In 

 particular, the last item on the list presents frequent difficulties. Science 

 does not deal in certainties but in probabilities. Scientific relationships 

 are relative, and are usually accompanied by a range of probability or 

 a margin of error. Thus, when Dr. Astin, Director of the National 

 Bureau of Standards, was asked, in the AD-X2 case, whether the NBS 

 analysis of the battery additive had been sufficiently accurate to con- 

 firm the nonexistence of a beneficial mystery ingredient, he replied in 

 probabilistic terms that he thought it was. The chance of such an 

 ingredient existing at all was small ; the chance that NBS had failed 

 to detect it in an analysis was small ; the chance that even if such an 

 ingredient existed and was beneficial, that its existence in undetectable 

 quantity would be significantly beneficial, was small; and the chance 

 that if it did exist and was beneficial even in unmeasurable quantities, 

 that NBS would be unable to detect the beneficial effect, was small. The 

 four improbabilities, multiplied together, made for an extremely small, 

 ultimate possibility. But not a certainty. 



Technical Differences of Opinion 



A recurring problem is the situation in which witnesses with out- 

 standing technical qualifications take opposite sides on a technical 

 issue. Members of Congress experience an understandable sense of 

 frustration when they find themselves obliged, as in the Test Ban 

 Treaty case, to decide on a complex technical matter that ranged out- 

 standing scientists against each other. The problem in that case was 

 that the two sets of scientists favored two conflicting hypotheses. Those 

 opposed to the treaty supported the hypothesis that further scientific 

 investigation would reveal phenomena that would enable development 

 of a workable defense against ballistic missiles. Those favoring the 

 treaty supported the hypothesis that the teclmical problem of over- 

 coming a defensive technology was inherently much simpler and less 

 costly than designing a defense — and that therefore the offense would 

 always keep well ahead of the defense. While there may be many non- 

 scientific reasons for a bias in a technical witness, there are many 

 occasions on which the witnesses disagree over unproved — and some- 

 times unprovable — scientific judgment. In such cases, the disagree- 

 ment itself is illuminating. 



Administration Versus Congress 



Mention was made of the particular difficulty of screening out bias 

 of technical civil servants who come before congressional committees 



