are not well graded in language, if our scales are giving us a fair measure of language ability. The next to the last row in the table gives the median score for deaf children, and this represents at present the standard scores on this scale for the deaf. Below this, in the last row, are given the median scores for hearing children, based on the results obtained by the writer.+

The medians for the hearing children. are grade medians. A score of 8 is the median for third-grade children, and so on for the other scores. The striking difference between deaf and hearing children is well brought out by a comparison

757

of these medians. Third-grade hearing children are able to accomplish as much on these language tests as the highest grades in these schools for the deaf. The medians for fourth-grade hearing children and above are higher than any of the medians found in these schools. This once again illustrates the tremendous handicap under which deaf children labor in the acquirement of language. table of distribution allows us to see that there are individual deaf children who are well up to and above the medians for the hearing children, but these, of course, are only exceptional cases.

Texas

Kentucky

Our

Feb., 1917 Mar., 1917

Total

8

ΙΟ

022456 4466

2400677gaa

9 356

255

Scale B.

Year of instruction

3

4

6

65

28

30 17

2

3

10 15 51 40

ΙΟ 14

2

3 2

6

25 47 49

51 23 14

192342

I

13322

123162227

I

134 88

Median hearing

30

2∞

4

ΙΟ

1883

363

52

14

SCALE C

Six months after Scale B had been given, Scale C was given to two of the schools-Ohio and Indiana. Unfortunately, it was impossible to give this scale to the other two schools. The median scores for these two schools are shown in Table III. Although, on the whole, the scores tend to increase from the lower to the higher grades, we note, as in Scale B, some fluctuation of the scores.

The distribution of all the cases in both schools is shown in Table IV. It is to be interpreted in the same way as Table II. As before, the wide variation in language ability in children in the same grade is the striking feature. In the ninth year of instruction, for example, we have children scoring all the way from 0 to 16; in the tenth year all the way from 2 to 18, and so on.

The next to the last line gives the standard scores for deaf children on Scale C. The number of cases upon which the standardization is based is 622, and therefore we cannot regard this as good a standardization as in the case of Scale B, where we have 1,098 cases. We note, however, that the score increases slowly as we progress up through the

Score

Year of instruction

Class

759

II

ΙΟ

II

12

2

12

4

12

2

II

16

15 13

17

12

15

18

II

12

Loboo

4.5

10

II

12

15

9

II II

3 Inter., M. 4 Inter., O.. 4 Inter., M. 5 Inter., O.. H. S. Jun. H. S. Sen..

Average class loss-0.73.

hearing children of the corresponding grade. This comparison of the hearing and deaf children on Scale C bears out what we have noticed in the comparison of the deaf and hearing on Scale B.

PROGRESS IN LANGUAGE ABILITY

The giving of Scale C six months after Scale B was undertaken for the purpose of measuring the progress in language during the six months of school. The result of the experiment was surprising to the writer; and although it may be characterized as a negative result in general, it is nevertheless instructive and of great interest. In the comparison of Scales B and C only the two schools in which both of the tests were given can be taken into account.

A first comparison of the combined. norms for the two schools for Scales B and C can be made, disregarding the individuals making up the total group.

the whole, of course, the individuals are the same; but owing to absence, etc., there are many children who took only one of the tests. This comparison is shown below.

Six grades show no change; three show one point less after six months, and two show a gain of one point. On the whole, this would seem to indicate a slightly worse performance in language after a period of six months. If, however, we call to mind that Trabue says that Scale C is slightly harder than Scale B, we may conclude that practically no progress in language ability has been measured by our scales. Two conclusions are possible: (1) That the children have actually made no progress, or (2) that our scales are not fine enough to measure the slight progress that has been made. Which of these two conclusions is justified we may see later after a study of the individual gains and losses. A point to be noted be

5

4

3

I

∞998

7000

0220

I

༡༦༤༽

906

Class

IA

IB

IC

13

ONNONON 2326326 +0000 TO TO

II

9

9

ΙΟ

7

13

Senior

Average class loss -0.25.

fore passing on is that the two grades which record an increase of one in the median are both higher grades.

We will now turn to a study of the two schools separately and see if they show the same tendency or whether there is any marked difference between them. Table V shows the class medians for Ohio. The first column gives the name of the class ("o" signifies oral method of instruction and "m" manual); the second column the number of children in the class when Scale B was given; the third the median score for Scale B; the fourth the median score for Scale C; the fifth the number in the class when C was given, and the sixth the difference between the scores of Scale B and Scale C. The minus quantities in this last column. indicate that the class did poorer on Scale C than on Scale B. It is this last column that interests us. We note that of the 28 classes, 13 classes showed no change in score; 13 classes showed a lower score on Scale C than on Scale B, and 2 classes a better score. These two classes are among the highest classes in the school. Adding the plus values and the minus values and subtracting, we have a negative value, and dividing this by the number of classes we get -0.73; that is to say, the average loss by class medians is 0.73 of a point. The classes on the aver

age have done poorer on Scale C than on Scale B.

Table VI shows the same results for Indiana. Of the 22 classes, 5 show no gain; II show a loss, and 6 a gain. The gains this time are not all made by the higher classes, although the largest gains are made by them. The average for all the 22 classes is a loss of 0.25 of a point, a somewhat smaller loss than in the case of Ohio. On the whole, however, the results from both of the schools are very much alike, so that neither school alone is responsible for the loss shown in the combined results.

A further inspection of our data is now necessary. So far we have compared schools and classes irrespective of the individuals making up the groups. We must now examine those individuals who took both the tests, throwing out all cases that for some reason or other were not present when either of the tests were given. In Ohio there were 331 children and in Indiana 226 who took both of the tests; altogether a total of 557 children. The individual gains and losses in score obtained by subtracting the score made on Scale B from the score made on Scale C are shown pictorially in Figure 4. The base line represents the number of points gained (plus) or lost (minus). The vertical line shows the number of cases. The