TABLE XI (Continued)

Showing percentile ratings of the two groups at the beginning and end of the six months' interval

N[16]I.Q. Healy Total Possible Average
Gain
1·25·27·10·29·30415·05
1C·21·15·01·12·53462·089
2·84·89·60·66·38416·063
2C·45·51·29·18-·07358 -·011
3·49·44·49·51·40328·066
3C·25·32·44·45·30287·05
4·59·59·38·22168·073
4C·27·32·71·42195·14
5·53·71·10·42213·14
5C·89·90·77·24133·08
6·53·77·02·55229·183
6C·37·38·04·27210·09
7·45·71·22·07·37478·062
7C·53·71·34·59·74393·148
8·45·34·70-·50303 -·125
8C·67·81·33·52·10241·025
9·27·46·23·39·71375·118
9C·35·40·73·89·72217·12
10·88·86·15·35·34400·056
10C·81·73·57·971·39392·218
11·79·73·20·34·21379·035
11C·77·79·05·23-·44379 -·073
12·67·53·47·49-·11161 -·036
12C·67·77·19·15·19208·032
13·09·19·79·78·02129·003
13C·06·05·24·67·06178·01
14·59·84·41·84·69212·115
14C·97·96·21·56·41104·068
15·62·49·38·55·12106·04
15C·19·20·26·79·66193·11
16·05·12·86·82·15205·03
16C·13·25·66·76·24313·04
17·09·17·63·40·66223·220
17C·59·71·27·43·21174·053
18·67·67·13·15·08175·027
18C·40·49·04·07·18247·03
19·59·69·91·85·1667·032
19C·67·67·11·18·16287·027
20·82·79·25·18-·06177 -·02
20C·91·94·66·75·2065·066
21·05·07·47·69-·16209 -·032
21C·34·62·99·92·46157·077
22·45·73·84·81·87120·174
22C·03·02·74·72·74275·123
23·32·49·86·98·65168·163
23C·23·59·33·501·26253·21
24·94·92·88·95- ·0689 -·012
24C·38·46·68·93·54179·09
25·17·15·37·55·49280·098
25C1·00·97·63·60·53134·088
26·21·15·51·36226 0
26C·40·37·90·94·04184·007
27·88·86·10·43·05268·013
27C·11·59·29·591·49320·25
28·23·29·63 1·00·59153·097
28C·53·67·96·87·2483·048

Table XI expresses the results of Table II, with the scores given in percentile values. In each test, the group was taken as composed of the two scores of every individual—the total number of scores in tests and retests, eliminating those scores where the other member of the pair was lacking, or where no retest was given. Thus case number 1 was just within the lowest 27% of the group in weight at the first weighing, but had advanced to the 44 percentile at the second. In height he gained from the 25 percentile to the 40 percentile. His total gain in all tests is 30 percentile out of a possible 415, and the average gain is ·05. The reader may see by scanning the table that the gains in the test group are practically equaled by those in the control group. There seems to be no consistent relationship between a low score in the first test and a large gain. This is true even though the method of calculation tends to minimize gains at the high end of the group, and losses at the low end. In table XII this may be seen more clearly in respect to I.Q. and the results for all the tests taken together with the I.Q. weighted by being counted twice. A large possible gain indicates that the score at the first testing was low, and vice versa. Considering I.Q. values, the largest possible gain in the test group was 95 per cent of the group. This occurred twice, in one case the actual gain being 7% of the group and in the other 2%. In the control group, the largest possible gain was 97% of the group, but actually this case fell 1% of the group. If we correlate possible gain with actual gain for each group, using the formula r = 2sin((Π/6)ρ)) when ρ = 1 - ((6ΣD²)/(n(n²-1))) we get a coefficient of correlation ·36 in the test group, and ·19

TABLE XII

Showing gains in percentile rating for I.Q., and for a total of all the tests with I.Q. weighted by being counted twice.

I.Q. Total
A B
1st P.R. 2d P.R.possible gainactual gainpossible gainactual gain Av. Gain
12527752415305
1C211579-6462538·9
28489165416386·3
2C4551556358- 7-1·1
3494451-5328406·6
3C2532757287305
4595941168227·3
4C27327351954214
5537147182134214
5C8990111133248
6537747242295518·3
6C3738631210279
745715526478376·2
7C537147183937414·8
8453455-11203-50-12·5
8C67813314241102·5
9274673193757111·8
9C35406552177212
10888612-2400345·6
10C817319-839213921·8
11797321-6379213·5
11C7779232379-44-7·3
12675333- 6161- 11- 3·6
12C67773310208193·2
1391991101292·3
13C6594- 117861
14598441252126911·5
14C97963- 1104416·8
15624938-13106124
15C19208111936611
16512957205153
16C13258712213244
179178182236622
17C59714112174215·3
1867673317582·7
18C4049609247183
195969411067163·2
19C676733287162·7
20827918- 3177-6- 2
20C91949365206·6
2157952209- 16- 3·2
21C34626628157467·7
22457355281208717·4
22C3297- 12757412·3
23324968171686516·3
23C2359773625312621
2494926- 289-6- 1·2
24C3846628179549
25171583- 2280499·8
25C10097- 3134538·8
26211579- 6226
26C403760- 31844·7
27888612- 226851·3
27C1159894832014825
282329776153599·7
28C5367471483244·8

in the control group. With the small number of cases involved the probable error is too great to allow either of these measures as indicative of relationship. We may say, then, that there is no definite tendency for those of low I.Q. to improve in six months after operation to a greater degree than those of higher I.Q.

Finally, in order to compare the results of the various tests, the measures of the gains of the test group in excess of the control were, for each test, expressed in terms of P. E. The averages and medians of these measures are collected in Table XIII. They show a very slight tendency toward gain in weight, height, and weight-height-age relationship; neither improvement nor loss in grip, tapping fatigueability and I.Q., and a rather curious tendency to loss in the Healy scores. This latter is very probably not a true measure since performance in the Healy Picture Completion test shows a rather high variability, and the cases are so few as to make the influence of single very high or low scores unduly great.

TABLE XIII