388 MILDRED WEST LORING 



six curves, a one, two, and three syllable curve each for adjectives 

 and nouns. Full lines represent the adjectives and dotted lines 

 the nouns. The number of syllables is indicated on the curve 

 itself. Smoother curves might have been obtained if larger 

 steps in the reaction time had been chosen, and if the number of 

 cases for the various groups of words had been more uniform. It 

 is worth while emphasizing the fact that reaction times do not 

 follow the error curve of mathematics. Whereas the error curve 

 is perfectly symmetrical, the reaction time curve is very much 

 skewed, with the maximum of the curve lying at a much lower 

 value than the mean of the extreme reaction times. The reason 

 for this is obvious. There is a physiological limit for the lowest 

 value of a series of reaction times, below which no reaction time 

 can fall, while on the other hand there is no such limit put upon 

 high reaction times. Reaction times may be expected to be of 

 any increasingly greater value beyond the minimum, up to 

 infinity (which we call failure) depending on the ability of the 

 subject. The curve then takes the form indicated here. It 

 should be noticed therefore that only such mathematical formulae 

 should be applied to the data of such curves as have been 

 developed for this type of measurements. For this reason 

 only the number of cases, averages and mean variations are given 

 in the data presented here. There has been an unfortunate 

 tendency to apply formulae that refer exclusively to the mathe- 

 matical curve of error to such curves as are given here. 



The reaction times may at first sight appear abnormally long. 

 Likewise the mean variations may seem high. But this is not so 

 however. In the first place no conclusion is valid unless based 

 on a sufficiently large number of cases. In a great many 

 investigations only a short list of words has been used and 

 sweeping conclusions are made on the basis of their results. 

 In this experiment over 8000 adjectives and nouns have been 

 used for stimulus words, so that a fairer indication is given of 

 the true nature of the reaction tune under the conditions laid 

 down in the experiment. In the second place no attempt has 

 been made in this work to choose "easy" words for stimulus words. 

 In much of the association work easy words have been chosen 



