These figures show again the reduction of problem behavior on the part of the original problem group and its increase on the part of the original nonproblem group. Also, by comparing the above data with those given on pages 50 and 53, it may be seen that the comparison of the first and last records of the 68 paired groups results in figures almost identical with those obtained from the comparison of the first and last records of 81 problem children and 85 nonproblem children, respectively, who included some unequated individuals. The fact may thus be stressed that the findings secured by studying various groups and combinations of data all consistently support one another by pointing to the same conclusions.

cases.

CHAPTER V. THE PROBLEM CONTROL GROUP

In Chapter I, Part II, it was explained that the canvass of serious problem children in the schools resulted in the location of 250 such One hundred and thirteen of these, which constituted, in the estimation of principals and teachers, the most serious of the group and for which home cooperation seemed forthcoming, were scheduled for intensive clinical treatment. At the close of the first year of the experiment and each semester thereafter, a School Record B was filled out for each of the remaining children reported who were still in the Berkeley schools. Of these records, 50 were selected which yielded the highest problem scores on the first record submitted. There are thus for these 50 children three consecutive records which give some basis for an analysis of their development and for a comparison of their progress with that of the problem group which did receive clinical help.

General description of the group.-This group of 50 is made up of 45 boys and 5 girls. The age range extends from 5 to 14 years. The intelligence quotients range from 74 to 130, with a median of 99. Their average educational achievement is not significantly different from what would be expected of any children of the same intelligence. In physical condition they resemble the other two groups studied. In the percentage of broken homes they resemble the experimental problem group.

Behavior records.-In considering the development of the behavior of these 50 children, it is important to note that they were given no special clinical treatment during the time of the experiment. They did receive, however, all the help which the regular procedure of the schools afforded, and it would be expected that much attention was centered upon each one of them by principal, counselor, and teachers in the effort to adjust their difficulties.

It should also be kept in mind that these 50 cases, while they were considered serious, did not represent the most urgent types of maladjustment for which teachers sought assistance. In the first place, the method of selection of cases for clinical treatment, as described in Chapter I, Part II, insured the choice of those which were causing the school the greatest amount of difficulty. In the second place, 23 of the original problem group had police records on one or more counts of juvenile offenses, and 19 of these were among the 81

58

cases which remained to the end of the study; while in this second group of 50 children, only 3 had police records, and in 2 of these cases the difficulty developed during the period of the study. These two facts are sufficient evidence to justify the statement that, at the beginning of the experiment, the records of these 50 children (had they been taken at that time) would have yielded a mean behavior problem score lower than that of the 109 children in the original experimental group or of the 81 children in this group for whom complete data are available both at the beginning and at the end of the investigation.

At the close of the term (August-December, 1929) for which the first record was taken for these 50 children, their mean behavior problem score was 213.3. For the experimental problem group of 81 children the mean behavior problem score at this time was 213.9. The two groups are, therefore, practically together at this point; but the EP group had already decreased its mean score since the beginning of the study (a year previous) by 26.3 points. Even though there is for the 50 PC cases no behavior record corresponding in time to the first one of the EP group to give statistical proof to the fact, yet from the evidence cited in the previous paragraph it is apparent that any such decrease in their records would have been highly improbable, since their initial problem behavior would not have scored as high as that of the EP group.

How the development of the two groups compares beyond this point (December, 1929) can be traced through subsequent behavior records. In Table 12 are given the distributions of the problemscores of the 50 children who did not receive clinical attention, as of the first and last records available. The means of these distributions show a decrease of 8.3 points from December, 1929, to December, 1930; but the standard error of this difference is 15, almost twice as great as the difference itself. The chances are only 71 in 100 that the true difference is greater than zero. There is little to indicate,. therefore, that this group of children, during the three semesters in which behavior records were submitted for them, have shown any change in overt problem behavior which is statistically reliable.

During the same time the mean behavior problem score of the EP group decreased from 213.9 to 190.7. The difference is 23.2-almost three times as great as that which is obtained for the 50 cases in the PC group. The standard error of this difference is 12.8. The chances are 96 in 100 that a significant change has taken place during this one year in the overt problem behavior of the group.

1 From 240.2 to 213.9.

TABLE 12.-Behavior problem scores of 50 problem children who were not given clinical attention (problem control group)

Read the table as follows: Of 50 problem children not given clinical attention, 2 had behavior problem scores for August to December, 1929, between 400 and 449; 5 had scores between 350 and 399, etc. Read similarly for scores for August to December, 1930. For meanings of statistical symbols, see Table 5 ani text.

When the time is extended to include the year previous to this one, the change in the EP group is doubled and becomes statistically certain, as has already been demonstrated in Chapter IV, Part II. No positive statement can be made regarding the total change (in these two years) which has taken place in the PC group, owing to the lack of behavior problem records for the first year of the study. But even if the amount of change during the first year equaled that which occurred during the second year (as it did for the EP group), the total amount of change for the two years would not constitute a reliable difference. The statistical evidence of the records for both groups during the second year of the study plus the deductions which may be made for the first year of the study would point to the conclusion that the change which took place in the PC group during the two years was not a significant one.

Uncooperative cases in the experimental problem group.-In attempting to carry out a program of behavior adjustment, every child guidance clinic experiences at times a lack of understanding, insufficient cooperation, waning interest, or all of these together, on the part of parents or guardians, which blocks the realization of its plans for treatment. This obstacle was encountered in the Berkeley program also. During the two years of the study there were 20 of the experimental problem cases which in reality had little actual clinical treatment, because those who were responsible for the child in question became indifferent or even refused to follow the clinical recommendations. This occurred despite the fact that at the beginning of the program they had expressed their desire to cooperate.

All these 20 cases were continued in the EP group to the close of the investigation. This was done on the basis, first, that each of them had at the initiation of the study been matched with a nonproblem control mate, and the unit character of the study demanded that they be retained; second, that clinical contacts had at least been begun and some influence might have carried over to subsequent attitudes or actions of the family; third, that principals and teachers had, through these initial clinical contacts, been given insight into the nature of the child's difficulties and might well have made some change in their own treatment of him by reason of such knowledge; fourth, that the retention of these children in the group would be in the interests of conservatism in evaluating the development of the clinical group as a whole.

When these 20 cases are isolated, the significant fact appears that their mean behavior problem scores in the first and last records submitted (two years apart) differ by only three points. The first mean is 212.2 and the last one is 209.1. While the sample is too small for detailed statistical treatment, this finding lends corroboration to that reported for the 50 children in the PC group, namely, that the chances are small, that the true difference between the first and last problem scores is greater than zero.

An even more significant point is noted in the development of the EP group when these 20 cases are subtracted from the 81 for whom complete records are available from beginning to end of the study. For the remaining 61 children who were actually under clinical care during the entire period of the investigation, the mean behavior problem score on the first record is 248.7 (instead of 240.2 as it was for the total group of 81 children). The mean behavior problem score on the last record is 185.1 (instead of 190.7 for 81 children).

Two facts are evident from these comparisons: First, the reliable difference already reported as existing between the initial and final problem scores of the EP group becomes even more certain when the 20 uncooperative cases are subtracted; second, this is true despite the fact that the 61 children remaining include more serious cases. with high initial behavior problem scores, as may be seen by comparing the means of the respective groups.

There are thus consistent indications that the problem children who received continued clinical attention during the two years of the study improved significantly in behavior, as measured by behavior problem score, while the problem children who received no clinical attention or who were deprived of its continuance due to lack of cooperation remained altogether or almost stationary in the develop