The facts of Dr. Rice's table show that there is a positive relation between the general standing of a school system in the tests and the amount of time devoted to arithmetic by its program. The relation is not close, however, being that expressed by a correlation coefficient of .36½. Within any one school system there is no relation between the standing of a particular school and the amount of time devoted to arithmetic in that school's program. It must be kept in mind that the amount of time given in the school program may be counterbalanced by emphasizing work at home and during study periods, or, on the other hand, may be a symptom of correspondingly small or great emphasis on arithmetic in work set for the study periods at home.

A still more elaborate investigation of this same topic was made by Stone ['08]. I quote somewhat fully from it, since it is an instructive sample of the sort of studies that will doubtless soon be made in the case of every elementary school subject. He found that school systems differed notably in the achievements made by their sixth-grade pupils in his tests of computation (the so-called 'fundamentals') and of the solution of verbally described problems (the so-called 'reasoning'). The facts were as shown in Table 11.

TABLE 10

Averages for Individual Schools in Arithmetic

CitySchool6th Year7th Year8th YearSchool Average
ResultPrincipleResultPrincipleResultPrincipleResultPrinciplePercent of
Mechanical Errors
Minutes
Daily
III 179.380.381.182.391.793.980.083.1 3.7 53
I 180.481.564.267.280.982.876.680.3 4.6 60
I 280.983.443.550.972.779.169.375.1 7.7 25
I 372.274.063.566.274.576.667.872.2 6.1 45
I 469.972.254.657.866.569.164.370.3 8.5 45
II 171.275.333.635.736.840.060.264.8 7.1 60
III 243.745.053.956.751.153.154.558.9 7.4 60
IV 158.960.431.234.141.643.555.158.4 5.6 60
IV 259.863.122.522.553.958.8 8.3
IV 354.958.135.238.643.545.051.557.610.5 60
IV 442.345.116.119.248.748.742.848.211.2
V 144.148.729.232.551.158.345.951.310.5 40
VI 168.371.333.536.626.930.739.042.9 9.0 33
VI 246.149.519.524.230.240.636.543.616.2 30
VI 334.536.430.535.123.324.136.042.515.2 48
VII 135.237.729.132.525.127.240.545.911.7 42
VII 235.238.715.016.419.621.236.540.610.1 75
VII 327.633.7 8.910.111.311.325.331.519.6 45

High achievement by a system in computation went with high achievement in solving the problems, the correlation being about .50; and the system that scored high in addition or subtraction or multiplication or division usually showed closely similar excellence in the other three, the correlations being about .90.

TABLE 11

Scores Made by the Sixth-Grade Pupils of Each of Twenty-Six School Systems

SystemScore in Tests
with Problems
Score in Tests
in Computing
23 356 1841
24 429 3513
17 444 3042
4 464 3563
25 464 2167
22 468 2311
16 469 3707
20 491 2168
18 509 3758
15 532 2779
3 533 2845
8 538 2747
6 550 3173
1 552 2935
10 601 2749
2 615 2958
21 627 2951
13 636 3049
14 661 3561
9 691 3404
7 734 3782
12 736 3410
11 759 3261
26 791 3682
19 848 4099
5 914 3569

Of the conditions under which arithmetical learning took place, the one most elaborately studied was the amount of time devoted to arithmetic. On the basis of replies by principals of schools to certain questions, he gave each of the twenty-six school systems a measure for the probable time spent on arithmetic up through grade 6. Leaving home study out of account, there seems to be little or no correlation between the amount of time a system devotes to arithmetic and its score in problem-solving, and not much more between time expenditure and score in computation. With home study included there is little relation to the achievement of the system in solving problems, but there is a clear effect on achievement in computation. The facts as given by Stone are:—