124 



THE ACQUISITION OF SKILL IN ARCHERY. 



Table 42. 



THE PROBLEM OF "DISTRIBUTION-RATE" RELATION. 



These results agree very well with the general finding.s for the effect of the 

 distribution of practice in the formation of other types of habits. Thorndike 

 has summarized the more important work done before 1910, and it is unneces- 

 sary to review this work here in detail. In general, studies upon human 

 learning indicate that the amount of improvement for a given amount of 

 practice is directly proportional to the time interval between the practice 

 periods and inversely proportional to the length of the jM-actice periods, but 

 there is some conflict in the results of experiments upon different types of 

 habits. The results of a few experiments fail to show these relations, but do 

 not seem to be conclusive. 



For students learning to write English words in German script Leuba and 

 Hyde give the results shown in table 42 for four groups practicing for periods 

 of equal length but at different intervals. The group practicing every third 

 day alone fails to show the advantage of 

 distributed practice, and not enough sub- 

 jects were used to eliminate the possibility 

 that this is only a chance variation. 



Lyon finds that short practice periods 

 give better results in the formation of lan- 

 guage habits, except in the case of short 

 verse and prose selections whose meaning 

 can be grasped as a unit; these are learned 

 most readily by concentrated practice. In 

 the case of such material, involving preex- 

 isting language habits of almost inconceivable complexity, it seems very ques- 

 tionable whether the learning process is comparable to the formation of sim- 

 pler habits. 



Hahn and Thorndike found no advantage in favor of either of two distribu- 

 tions of practice, but their groups of subjects are not strictly comparable. 

 The habits studied were those involved in computation and in every case the 

 group with concentrated practice had the greater number of preexisting 

 habits, which may have ol)scured any disadvantages due to the concentration 

 of practice. 



With the exception of these cases the evidence favors the belief that the 

 rate of learning varies inversely as the concentration of practice. The major- 

 ity of investigators seem to believe that this holds true only within certain 

 time limits, but with the possible exception of the results of Leuba and Hyde, 

 no such limits have been established experimentally. The tj^Des of habits 

 subject to the "distribution-rate" relation and the time-limits of the relation 

 are not yet settled and the solution of the problem will demand a closer inquiry 

 into the physiological causes of the relation than has yet been made. 



The following quotation from Starch summarizes practically all the explana- 

 tions which have been advanced to account for the effect of the distribution 

 of practice upon the rate of learning : 



Why are shorter and more numerous periods economical? The main reason, no doubt, 

 is the well-known fact that a period of rest after newly formed associations gives them a 

 chance to become settled and fixed. The slower rate of improvement of the third and 

 fourth groups is due in part to fatigue. The forty-minute group shows no gain in the last 



