6 



Indiana Unive'^sity Studies 



as in the above arrangement. All 27 systems reported in the 

 seventh and eighth grades in the four tests, the total number 

 of the children being 1,017 and 919 respectively; 26 systems 

 reported the fifth and sixth grades in all four tests with a to- 

 tal of 1,326 and 1,194 children respectively, whereas only 18 

 systems reported the fourth grade in addition, subtraction, 

 and multiplication, and only 14 in division, all having a total 

 of 1,016. Thus the grand total of children reported in the 

 investigation is 5,472. The scarcity and irregularity of the 

 data for the fourth grade has caused the omission of this 

 grade from some parts of the study. 



Checking the Class Record Sheets. As in the two earlier 

 studies, 1914^ and 1915-, the tests were given and scored by 

 the local school officers, the results tabulated on the Class Rec- 

 ord Sheet No. 1, and sent to the Bureau of Cooperative Re- 

 search. 



When this material was examined considerable corrective 

 work was found necessary. 



Student assistants who were assigned to this work were 

 given the following instructions : 



1. Count the individual test sheets for each class, e.g. 7B, Central 

 School, Bluff ton, 



2. Check number of pupils indicated for each class on the Class 

 Eecord Sheets. See if this number tallies with number of individual test 

 sheets. If it does, O.K. with your initials and date; if not, go thru the 

 individual sheets comparing scores with those recorded on the Class 

 Record Sheet and correct all errors. Then O.K. with initials and date. 



3. Total the number of pupils and throw together the records for 

 all classes within a given grade, for each city; e.g. Bluff ton may have 

 three 7B's and four 7A's. Throw together all these records to get the 

 record in speed and accuracy for the seventh grade in Blufiton. 



Calculation of City Scores. Following the correction of the 

 class tabulations, median scores for both rate and accuracy 

 were calculated for each city. Class Record Sheet No. 1 for 

 use with Series B yields distributions of the scores for rate or 

 number of examples attempted and for accuracy or per cent of 

 examples done correctly. For cities having only one class 

 for a grade the city medians for attempts and accuracy were 

 calculated directly from the class record sheets. In the case 

 of cities having two or more classes to a grade, the corres- 



1 M. E. Haggerty : "Arithmetic : A Cooperative Study in Educational Measurements" 

 (Indiana University Studies, No. 27). 



2 M. E. Haggerty: "Studies in Arithmetic" (Indiana University Studies, No. 32), 



