32 FREDM. UBER 



Often in routine determinations, as for example a quantitative chemi- 

 cal analysis, it is customarj^ to run merely a duplicate set of obser- 

 vations. The variables are fewer in this case and more nearly under 

 control. The purpose of the duplication is primarily to avoid gross 

 mistakes or blunders and not to attain that increase in the reliability 

 of the observations that would accrue from a decrease in the standard 

 error. Where the time consumed in repeating a set of measurements 

 may be unreasonal)ly long owing to the complexity of an experiment, 

 it may not be worth while to proceed beyond a single repetition. 

 More can be gained oftentimes by taking increased precautions to 

 eliminate possible sources of systematic and constant errors and by 

 exercising meticulous care in making and recording the observations. 

 The records usually should be complete even though the observations 

 may not be fully understood at the time. The importance of schedul- 

 ing critical check experiments as a safeguard against unwitting blun- 

 ders probably cannot be stressed too highly. 



I. SECURE EFFECTIVE PUBLICATION 



To be meaningful, according to the viewpoint adopted for this 

 chapter, experimental research should contribute factual information 

 that will be of value in correlating, explaining, or predicting behavior. 

 But "facts," in the words of Midgley {21), "are still, and probably 

 always will be, determined by vote." No matter how completely 

 convinced a research worker himself is that he has found a new fact, 

 it does not become generally accepted as such unless and until it re- 

 ceives the favorable vote of a "scientific jury." The more complete 

 and the better established the experimental evidence, the better are 

 the chances of having the results of research readily accepted, but this 

 in itself is not sufficient. It must be considered also an essential 

 duty, inescapably obligatory on the part of the research scientist, to 

 present the evidence forcefully to the "jury." 



It is helpful to write in a language that can be read with ease by 

 the majority of the specialists in a field. For this reason chemical 

 research was published primarily in German during the past century. 

 Chemists who felt they had made an important contribution which 

 might be concealed by their native language, perhaps Russian o)' 

 Swedish, would not only write their articles in German but also pub- 

 lish them in a German periodical. It is entirely possible that a scien- 

 tific article published in Chinese might well be lost for a number of 



