RECORDS OF MEETIXGS 415 



Distribution of time. — Several different schedules of time distribution 

 were carried out, which can not be described here in detail, but suffice it 

 to sa}" that the main comparison was made between a regular distribu- 

 tion — i. e., 15, 15, 15, etc., or 10, 10, 10, etc. — per day, or per every other 

 day, as the case might be, with an irregular distribution — i. e., 15, 15, 10, 

 10, 5, etc. — or a decreasing number of minutes and an increasing time- 

 interval between the periods of practise. 



Time of day. — ^o special effort was made to keep the time of day con- 

 stant. Some classes worked in the afternoon and some in the forenoon. 

 A check was kept on this and no appreciable difference seems to be mani- 

 fest by this factor. 



Stress. — In the beginning it was aimed to stress accuracy at least 

 enough to keep up to the standard set in the initial practise and if possi- 

 ble to show a consistent gain in same. This was done. 



Conductors of the experiment. — The writer personally conducted all 

 the initial and final practises, the entire experiment with the sixth and 

 seventh grades and practically all the work with the lower grades. In a 

 few instances it was impossible because of lack of time to get around to 

 all the sections at the time appointed, and in such cases the teacher took 

 charge of the practise. All the teachers were in the rooms during all the 

 work and consequently were able to take charge without any change in 

 either the stress or the method. 



Conclusion. — The results show an improvement of about 50 per cent, 

 in addition, 100 per cent, in division and from 50 to 150 per cent, in 

 geogTaphy and history, with a noticeably larger increase in case of the 

 groups working according to the irregular distribution of time. In point 

 of accuracy there was an increase of 1 to 9 per cent, in all cases. The 

 experiment seems to warrant the conclusion that an irregular distribu- 

 tion of time is more advantageous in the case of such school subjects as 

 are here considered. 



Mr. Trabue said in abstract: After testing over six thousand public- 

 school children with a series of fifty-six mutilated sentences, twenty-four 

 of the sentences were selected to serve as a language scale. ^ 



Sentences were used rather than paragraphs, because it was believed 

 that the paragraph was too large a unit of thought for the child in the 

 lower grades to handle successfully. The mental labor required of the 

 teacher in evaluating paragraph completions has heretofore kept teachers 

 from making much use of completion tests. The short sentence units 

 here presented to the child for completion and to the teacher for evalua- 



■5 Cf. M. R. Trabue : "Some ResuUs of a Graded Series of Completion Tests." School 

 and Society, Vol. I, pp. 537-540. 



