These percentages are computed from the age totals in Table VI, just as the ones preceding are computed from Table V. It seems worthy of note here that close to 50 per cent of the non-failing drop-outs occur under 16 years of age, for both the boys and the girls; but that the number of the failing pupils who drop out does not reach 20 per cent for the boys or the girls in these same years. It is likewise remarkable in these distributions that the percentages for boys and for girls show such slight differences in either of the two groupings.

---

SUMMARY OF CHAPTER II

If to the recorded failures the virtual but unrecorded ones are added, the percentage of failing pupils is 66 per cent. This percentage is higher for the boys than for the girls by a difference of 6 per cent.

Of the graduating pupils, 58.1 per cent fail one or more times.

Of the non-failing non-graduates 78 per cent are lost from school by the end of their first year. But the failing non-graduates have not lost such a percentage before the end of the third year.

The percentage of pupils failing increases for the first four semesters, and lowers but little for two more semesters. One third to one half of the pupils fail in each semester to seventh.

In the distribution of failures by ages and semesters, 86 per cent are found from ages 15 to 18 inclusive. Thirty-four per cent of the failures occur after the end of the second year, when 52.2 per cent of the pupils have been lost and others are leaving continuously.

Mathematics, Latin, and English head the list in the percentages of total failures, and together provide nearly 60 per cent of the failures; but English has a large subject-enrollment to balance its count in failures.

Mathematics, Latin, and German fail the highest percentages on the number of pupils taking the subjects.

In several subjects the percentages of failure based on the total failures are higher for the graduates than for the non-graduates.

For the pupils dropping out without failure the median age is at 16, with the mode at 15. For the failing drop-outs both the median and the mode are at the age of 17. Nearly 50 per cent of the non-failing drop-outs occur under age 16, but not 20 per cent of the failing non-graduates are gone by that age. The percentage of drop-outs is higher for older pupils.

---

CHAPTER III

WHAT BASIS IS DISCOVERABLE FOR PROGNOSTICATING THE OCCURRENCE OF OR THE NUMBER OF FAILURES?

1. ATTENDANCE, MENTAL OR PHYSICAL DEFECTS, AND SIZE OF CLASSES ARE POSSIBLE FACTORS

Any definite factors available for the school that have a prognostic value in reference to school failures will help to perform a function quite comparable to the science of preventive medicine in its field, and in contrast with the older art of doctoring the malady after it has been permitted to develop. Such prognostication of failure, however, need not imply a complete knowledge of the causes of the failures. It may simply signify that in certain situations the causes are less active or are partly overcome by other factors.

Perhaps one of the simplest factors with a prognostic value on failure may be found in the facts of attendance. Persistent or repeated absence from school may reach a point where it tends to affect the number of failures. It happened, unfortunately, that the reports for attendance were incomplete or lacking in a considerable portion of the records employed in this study. Consequently the influence of attendance is given no especial consideration in these pages, except as explained in Chapter I, that the pupil must have been present enough of any semester to secure his subject grades, else no failure is counted and no time is charged to his period in school. In this connection, Dr. C.H. Keyes[14] found in a study of elementary school pupils that of 1,649 pupils losing four weeks or more in a single year 459 belonged to the accelerate pupils, 647 to those arrested, and 543 to pupils normal in their school work. He accredits such large loss of time as almost invariably the result of illness and of contagious disease. He also says, "Prolonged absence from school is appreciable in producing arrest especially when it amounts to more than 25 days in one school year." But the diseases of childhood, with the resultant absence, are less prevalent in the high school years than earlier. Furthermore, the losses due to change of residence will not be met with here, for, as explained in Chapter I, no transferred pupils are included subsequent to the time of the transference either to or from the school.

The influence of physical or mental defects also deserves recognition here as a possible factor relative to school failures, although this study has no data to offer of any statistical value in that regard. A few pupils in high school may actually reach the limits prescribed by their 'intelligence quotient'[15] or general mental ability, or perhaps, as Bronner[16a] so interestingly points out, be handicapped by some special mental disability. If such be true, they will doubtless be found in the number of school drop-outs later referred to as failing in 50 per cent or more of their work; but we have no measurement of intelligence recorded for them to serve our purposes of prognostication. In the matter of physical defects alone, the report of Dr. L.P. Ayres[17] on a study of 3,304 pupils, ten to fourteen years old, in New York City, states that "In every case except in that of vision the children rated as 'dull' are found to be suffering from physical defects to a greater degree than 'normal' or 'bright' children." The defects of vision, which is the exception noted, may be even partly the result of the studious habits of the pupils. Bronner[16b] remarks on the "relationships between mental and physical conditions," and also on how "the findings on tests were altogether different after the child had been built up physically." But Gulick and Ayres[18] conclude that it is evident from the facts at hand that if vision were omitted the percentage of defects would dwindle and become comparatively small among the upper grades. This would probably be still more true for the high school; but this whole field has not yet been completely and thoroughly investigated.

It would be very desirable to have ascertained the size of the classes in which the failures were most frequent, as well as the relative success of the pupils repeating subjects in larger or in smaller classes. But, as such facts were unobtainable, it is permitted here simply to recognize the possible influence of this factor. It seems deserving in itself of careful and special study. From the standpoint of the pupil, the kind of subject, the kind of teacher, and the sort of discipline employed will tend to influence the size of class to be called normal, and to make it a sort of variable. Thirty pupils is regarded by the North Central Association as the maximum size of class in high school.[19] Surely the size of class will react on the pupil by affecting the teacher's spirit and energy. Reference is made by Hall-Quest[20] to an experiment, whose author is not named, in which 829 pupils stated that their "most helpful teachers were pleasant, cheerful, optimistic, enthusiastic, and young." If such be true then the very large size of classes will tend to reduce the teacher's helpfulness.

---

2. THE EMPLOYMENT OF THE SCHOOL ENTERING AGE FOR PROGNOSIS

A promising but less emphasized basis of prognosticating the school success or failure of the pupils is found in the employment of the school entering ages for this purpose. The distribution of all the pupils (except 30 undistributed ones, for whom the records were incomplete), according to entering age, is here presented, independently for the boys and for the girls.

DISTRIBUTION OF PUPILS BY THEIR ENTRANCE AGES TO HIGH SCHOOL

	AGES									Undis-
Total	12	13	14	15	16	17	18	19	20	tributed
2646 B.	16	211	820	900	497	148	23	10	7	14
3495 G.	8	259	1124	1217	614	194	51	10	8	16

The entering ages of these 6,141 pupils are distributed from 12 to 20, with 30 of them for whom the age records were not given. The median age for all the entrants is 15.3. But in order to compare this with the median entering age (14.9) of the 1,033 pupils reported by King[21] for the Iowa City high school, or with the median entering age (14.5) of 1000 high school pupils in New York City, as reported by Van Denburg,[22a] it is necessary to reduce these medians to the same basis of age classification. Since age 15 for this study starts at 141/2, then 15.3 would be only 14.8 (15.3-.5) as by their classification. The percentages of the total number of pupils for each age are given below.

PERCENTAGES OF PUPILS FOR EACH ENTERING AGE

	AGES
	12	13	14	15	16	17	18	19        20
								Undistributed
Total	0.4	7.6	31.6	34.4	18.1	5.5	1.2	1.0
Boys	0.6	8.0	31.0	37.8	18.8	5.6	0.8	1.1
Girls	0.2	7.4	32.4	34.8	17.5	5.5	1.4	1.0

We see that 84 per cent of the pupils enter at age 14, 15, and 16, or, what is perhaps more important, that nearly 40 per cent enter under 15 years of age. The similarity of percentages for boys and for girls is pronounced. The slight advantage of the boys for ages 12 and 13 may be due to home influence in restricting the early entrance of the girls, thus causing a corresponding superiority for the girls at age 14. The mode of this percentage distribution is at 15 for both boys and girls.

What portion of each entering-age group has no failures? This question and the answer presented below direct our attention to the superiority of the pupils of the earlier entering ages. That these groups of earlier ages of entrance are comprised of pupils selected for their capabilities is shown by the successive decrease in the percentages of the non-failing as the ages of their entrance increases, up to age 18.

DISTRIBUTION OF THE PUPILS WHO DO NOT FAIL,
FOR EACH ENTERING-AGE-GROUP

	AGES
Totals	12	13	14	15	16	17	18	19	20
1061 B	11	102	320	309	186	56	9	4	4
1575 G	3	133	522	545	256	73	29	7	6
% of
Entrants	58.0	50.0	43.4	40.0	39.8	37.7	55.0

Here is definite evidence that the pupils of the earlier entering ages are less likely to fail in any of their school subjects than are the older ones. Those entering at ages 12 or 13 escape school failures altogether for 50 per cent or more of their numbers. Those entering at age 14 are somewhat less successful but still seem superior to those of later entrance ages. It is encouraging, then, that these three ages of entrance include nearly 40 per cent of the 6,141 pupils. There is, of course, nothing in this situation to justify any deduction of the sort that pupils entering at the age of 17 would have been more successful had they been sent to high school earlier, except that had they been able to enter high school earlier they would have represented a different selection of ability by that fact alone. There is also a sort of selection operative for the pupils entering at ages 18, 19, or 20, which tends to account at least partly for the rise in the percentage of the non-failing for these years. It is safe to believe that for the most part only the more able, ambitious, and purposeful individuals are likely to display the energy required or to discern the need of their entering high school when they have reached the age of 18 or later. The appeal of school athletics will in this case seem very inadequate to explain their entrance so late, since the girls predominate so strongly for these years. Then it may be contended further that the added maturity and experience of those later entrants may partly compensate for a lack of native ability, if such be the case, and thereby result in a relatively high percentage of non-failing pupils for this group.

It is readily conceded that the avoidance of failure in school work serves as only one criterion for gauging the pupils' accomplishment. It is accordingly important to inquire how the different age-groups of school entrants compare with reference to the persistence and ability which is represented by school graduation. A truly striking array of percentages follows in reference to the question of how many of the entering pupils in each age-group do graduate.

DISTRIBUTION OF THE PUPILS GRADUATING FOR EACH ENTERING-AGE GROUP

	AGES
Totals	12	13	14	15	16	17	18	19	20
796 B	14	115	290	253	99	20	2	1	2
1140 G	5	151	465	363	121	26	5	1	0
% of Entrants	79.1	56.6	38.8	29.9	20.0	13.4	9.1	10.0	13.3

These percentages bear convincing testimony in support of the previous evidence that the pupils of the earlier entering years are highly selected in ability. Of all the high school entrants they are the 'most fit,' the least likely to fail, and the most certain to graduate. The percentage of pupils graduating who entered at the age of 12 is approximately four times that of pupils who entered at the age of 16. Thirteen is more than four times as fruitful of graduates as age 17; fourteen bears a similar relationship to age 18; and the percentage for fifteen is three times that for age 19, as is apparent from the above figures. The fact that the decline of these percentages ceases at age 19 is probably due to the greater maturity of such later entrants.

When we make inquiry as to what portion of the graduates in each of the above groups 'goes through' in four years or less, we get the series of percentages indicated below.

PERCENTAGE OF THE GRADUATES WHO FINISH IN FOUR YEARS OR LESS,
FOR EACH OF THE ENTERING-AGE GROUPS

Ages	12	13	14	15	16	17	18
% of Each Group	84.3	85.7	75.8	79.5	84.3	80.4	100

It appears that the ones in the older age-groups who do graduate are not so handicapped in reference to the time requirement for graduation as we might have expected them to be from the facts of the preceding pages. Perhaps that fact is partly accounted for by the not unusual tendency to restrain the more rapid progress of the younger pupils or to promote the older ones partly by age, so that by our school procedure the younger and the brighter pupils may at times actually be more retarded, according to mental age, than are the older and slower ones.

Since the same teachers, the same schools, and the same administrative policy were involved for the different entrance-age groups, the prognostic value of the factor of age at entrance will seem to be unimpaired, whether it operates independently as a gauge of rank in mental ability, or conjointly with and indicative of the varying influence on these pupils of other concomitant factors, such as the difference of economic demands, the difference of social interests, the difference in permanence of conflicting habits of the individual, or the difference in effectiveness of the school's appeal as adapted for the several ages. One may contend, and with some success, that the high school régime is better adjusted to the younger pupils, with the consequent result that they are more successful in its requirements. The distractions of more numerous social interests may actually accompany the later years of school age. In reference to the social distractions of girls, Margaret Slattery says,[23] "This mania for 'going' seizes many of our girls just when they need rest and natural pleasures, the great out-of-doors, and early hours of retiring." But surely such distractions are not peculiar to the girls alone. The economic needs that arise at the age of sixteen and later are often considered to constitute a pressing factor regarding the continuance in school. But VanDenburg[22b] was convinced by the investigation, in New York City, of 420 rentals for the families of pupils that "on the whole the economic status of these pupils seems to be only a slight factor in their continuance in school." A similar conclusion was reached by Wooley,[24] in Cincinnati, after investigating 600 families, in which it was estimated that 73 per cent of the families did not need the earnings of the children who left school to go to work. The corresponding report by a commission[25] in Massachusetts shows 76 per cent. The same facts for New York City[26] indicate that 80 per cent of such families are independent of the child's wages. But Holley concludes,[27] from a study of certain towns in Illinois, that "there is a high correlation between the economic, educational, and social advantages of a home and the number of years of school which its children receive." It will hardly be denied that even aside from the relation of the family means to the school persistence, the economic needs may have a direct influence on the failing of the children in their school work, either because home conditions may be decidedly unfavorable for required home study, or because of the larger portion of time that must be given to outside employment, with its consequent reduction of the normal vitality of the individual or of his readiness to study. But, in spite of the possible interrelationship of these factors, it still appears that the school entrance age of pupils will serve as a valuable sort of educational compass to foretell in part the probable direction of their later accomplishment.

---

3. THE AMOUNT OF FAILURE AT EACH AGE AND ITS RELATION TO THE POSSIBILITY OF FAILING FOR THAT AGE

We have considered at some length the prognostic value of the age at entrance. Here we shall briefly consider the prognostic value of age in reference to the time when failures occur and the amount of failure for such age. If we were to total all the failures for a given age, as shown in Table I, what part will that form of the total subjects taken by these pupils at the time the failures occur? In other words, what are the percentages formed by the total failures on the possibility of failing, for the same pupils and the same semesters, considered by age groups? The summary line of Table I gives the total failures according to the ages at which they occurred. The number of pupils sharing in each group of these failures is also known by a separate tabulation. Then the full number of subjects per pupil is taken as 41/2, since approximately 50 per cent of the pupils take five or more subjects each semester and the other 50 per cent take four or less (see p. 61). With the number of pupils given, and with a schedule of 41/2 subjects per pupil, we are able to compute the percentages which the failures form of the total subjects for these failing pupils at the time. These percentages are given below.

THE PERCENTAGES FORMED BY FAILURES AT EACH AGE ON THE POSSIBILITIES OF FAILING AT THAT AGE AND TIME, FOR THE SAME PUPILS

Ages	13	14	15	16	17	18	19	20	21
%	36.6	38.0	37.9	40.9	40.8	41.2	41.3	42.0	42.7

There is an almost unbroken rise in these percentages from 36.6 for age 13 to 42.7 for age 21. Not only do a greater number of the older pupils fail, as was previously indicated, but they also have a greater percentage of failure for the subjects which they are taking. It seems appropriate here to offer a caution that, in reading the above percentages, one must not conclude that all of age 14 fail in 38 per cent of their work, but rather that those who do fail at age 14 fail in 38 per cent of their work for that semester. The evidence does not seem to indicate that the maturity of later years operates to secure any general reduction of these percentages. The prognostic value of such facts seems to consist in leading us to expect a greater percentage of failures (on the total subjects) from the older pupils who fail than from the younger ones who fail. If it were possible to translate the above percentages to a basis of the possibility of failure for all pupils, instead of the possibility for failing pupils only, the disparity for the different ages would become more pronounced, as the earlier ages have more non-failing pupils. But this we are not able to do, as our data are not adequate for that purpose.

---

4. THE INITIAL RECORD IN HIGH SCHOOL FOR PROGNOSIS OF FAILURE

For this purpose the pupil record for the first year, in reference to failures, is deemed more adequate and dependable than the record for the first semester only. Accordingly, the pupils have been classified on their first year's record into those who had 0, 1, 2, 3, and up to 7 or more failures. Then these groups were further distributed into those who failed 0, 1, 2, 3, and up to 7 or more times after the first year. From such a double distribution we may get some indication of what assurance the first year's record offers on the expectation of later failures. Table VII presents these facts.

Table VII is read in this manner: Of all the pupils who have failures the first year (805 boys, and 1,129 girls) 397 boys and 672 girls have failures later, 105 boys and 130 girls have 1 failure later, 77 boys and 98 girls have 2 failures later, while 68 boys and 63 girls have seven or more failures later. The column of totals to the right gives the pupils for each number of failures for the first year. The line of totals at the bottom gives the pupils for each number of failures subsequent to the first year.

The table includes 3,508 pupils, since those who did not remain in school more than three semesters are not included (1,120 boys, 1,513 girls). Obviously, those who do not stay more than one year would have no subsequent school record, and those remaining only a brief time beyond one year would not have a record of comparable length. It seems quite significant, too, for the purposes of our prognosis, that of the 2,633 pupils dropping out in three semesters or less only about 43 per cent have ever failed (boys—46 per cent, girls—41 per cent). In contrast to this, nearly 70 per cent (69.6) of those continuing in school more than three semesters fail one or more times. Those who drop out without failure, in the three semesters or less, constitute nearly 60 per cent of the total non-failing pupils (2,568), but the failing pupils who drop out in that same period constitute less than 32 per cent of the total who fail (3,573). This situation received some emphasis in Chapter II and will be further treated in Chapter IV, under the comparison of the failing and non-failing groups.

TABLE VII

SUBSEQUENT RECORD OF FAILURES FOR PUPILS
FAILING 1, 2, 3, ETC., TIMES THE FIRST YEAR

FAILURES
OF 1ST	FAILURES SUBSEQUENT TO FIRST YEAR
YEAR	0	1	2	3	4	5	6	7+	TOTALS
0 B.	397	105	77	50	47	37	24	68	805
G.	672	130	98	60	53	27	26	63	1129
	1069	235	175	110	100	64	50	131	1934
1 B.	46	43	34	33	35	21	15	46	273
G.	65	43	53	33	33	19	17	67	330
	111	86	87	66	68	40	32	113	603
2 B.	22	24	23	23	30	21	13	57	213
G.	42	32	27	21	22	13	15	83	255
	64	56	50	44	52	34	28	140	468
3 B.	7	5	16	10	10	13	10	30	101
G.	8	9	7	10	17	6	7	41	105
	15	14	23	20	27	19	17	71	206
4 B.	6	8	5	7	7	11	7	23	74
G.	8	7	5	6	10	8	4	27	75
	14	15	10	13	17	19	11	50	149
5 B.	3	1	0	2	1	5	3	11	26
G.	5	9	5	6	5	4	2	14	50
	8	10	5	8	6	9	5	25	76
6 B.	0	1	4	2	1	1	1	10	20
G.	2	1	2	2	6	2	0	6	21
	2	2	6	4	7	3	1	16	41
7+ B.	3	2	1	0	1	0	2	5	14
G.	1	2	1	1	5	2	0	5	17
	4	4	2	1	6	2	2	10	31
Tot. B.	484	189	160	127	132	109	75	250	1526
G.	803	233	198	139	151	81	71	306	1982
	1287	422	358	266	283	190	146	556	3508

Referring directly now to Table VII, we find that 44.7 per cent of those not failing the first year do fail later. Of all those who fail the first year, 13.8 per cent escape any later failures. Of all the pupils included in this table 15.8 per cent have 7 or more failures, while of those failing in the first year 27 per cent later have 7 or more failures. For the number included in this table 30.4 per cent have no failures assigned to them.

PERCENTAGE OF FIRST YEAR FAILING GROUPS, WHO LATER HAVE NO FAILURES

No. of F's. in First Year	1	2	3	4	5	6	7+
Per Cent of Groups Having
No Failures Later	18.4	13.7	7.2	9.4	10.5	5.0	12.9

About the same percentage of the boys and of the girls (near 60 per cent) is represented in Table VII. The girls have an advantage over the boys of about 8 per cent for those belonging to the group with no failures, and of about 1 per cent for the group with seven or more failures.

No unconditional conclusion seems justified by this table. In the first year's record of failures there are good grounds for the promise of later performance. We may safely say that those who do not fail the first year are much less likely to fail later, and that if they do fail later, they have less accumulation of failures. Yet some of this group have many failures after the first year, and others who have several failures the first year have none subsequently. Generally, however, the later accumulations are in almost direct ratio to the earlier record, and the later non-failures are in inverse ratio to the debits of the first year.

---

5. THE PROGNOSIS OF FAILURES BY THE SUBJECT SELECTION

From the distribution of failures by school subjects as presented in Chapter II, this will seem to be the easiest and almost the surest of all the factors thus far considered to employ for a prognosis of failure. For of all pupils taking Latin we may confidently expect an average of a little less than one pupil in every five to fail each semester. For the entire number taking mathematics, the expectation of failure is an average of about one in six for each semester. German comes next, and for each semester it claims for failure on the average nearly one pupil in every seven taking it. Similarly French claims for failure one in every nine; history, one in every ten; English and business subjects, less than one in every twelve. It will be noted that the average on a semester basis is employed in this part of the computation. Consequently, it is not the same as saying that such a percentage of pupils fail at some time, in the subject. The pupil who fails four times in first year mathematics is intentionally regarded here as representing four failures. Likewise, the pupil who completes four years of Latin without failure represents eight successes for the subject in calculating these percentages. Every recorded failure for each pupil is thus accounted for.

It was also noted in Chapter II that the percentages of the total failures run higher in mathematics, Latin, history, and science, for the graduates than for the non-graduates. This fact is not due to the greater number of failures of graduates in the earlier semesters, when most of the non-graduate failures occur, but to the increase of failures for the graduates in the later years, as is disclosed in Tables II and IV. Accordingly, we may say that those two subjects which are most productive of school failures are increasingly fruitful of such results in the upper years. This does not seem to be the usual or accepted conviction. Certain of the school principals have expressed the assurance that it would be found otherwise. Such deception is easily explainable, for the number of failures show a marked reduction, and the rise of percentages is consequently easily overlooked. It is quite possible, too, that in some individual schools there is not such a rise of the percentages of failure for the graduates in any of the school subjects. In a single one of the eight schools reported here neither Latin nor mathematics showed a higher percentage of failure for the graduate pupils over the non-graduates. In the other seven schools the graduates had the higher percentage in one or both of these subjects.

---

6. THE TIME PERIOD AND THE NUMBER OF FAILURES

The statement that the number of failures will be greater for the failing pupils who remain in school the longer time may seem rather commonplace. But it will not seem trite to state that the percentage of the total failures on the total subject enrollments increases by school semesters up to the seventh; that the percentage of possible failures for all graduating pupils increases likewise; or that the failures per pupil in each single semester tend to increase as the time period extends to the later semesters. Yet radical as these statements may sound, they are actually substantiated by the facts to be presented.

PERCENTAGE OF THE TOTAL FAILURES
ON THE TOTAL SUBJECT ENROLLMENT, BY SEMESTERS

Semester	1	2	3	4	5	6	7	8	9	10
Per Cent	11.5	13.9	14.5	15.1	14.5	15.3	12.1	9.9	10.9	6.2

The 808 pupils who received no marks, and many of whom dropped out early in the first semester, are not included in the subject enrollment for the above percentages. Otherwise the enrollments taken are for the beginning of each semester and inclusive of all the pupils. These percentages rise from 11.5 in the first semester to 15.3 in the sixth semester. Then the percentages drop off, doubtless due to the increasing effect by this time of the non-failing graduates on the total enrollment. The graduates alone are next considered in this respect.

PERCENTAGES OF THE TOTAL FAILURES FOR THE GRADUATES
ON THE TOTAL SUBJECT ENROLLMENT FOR GRADUATES, BY SEMESTERS

Semester	1	2	3	4	5	6	7	8	9	10
Per Cent	5.9	6.6	7.8	9.1	9.2	10.5	9.1	7.3	8.8	5.2

These percentages are based on the total possibility of failure, and reach their highest point in the sixth semester, where the percentage of failure is nearly twice that for the first semester. These same facts may be effectively presented also by the percentages of such failures for the graduates on the total subject enrollment for only the failing graduates in each semester.

PERCENTAGES OF THE TOTAL FAILURES FOR THE GRADUATES
ON THE TOTAL SUBJECT ENROLLMENT FOR FAILING GRADUATES, BY SEMESTERS

Semester	1	2	3	4	5	6	7	8	9	10
Per Cent	31.4	31.2	31.8	32.7	32.3	36.6	37.5	37.4	38.0	36.0

The percentages here are limited to the total possibilities of failure for those graduates who do fail in each semester. They reach the highest point in the ninth semester, with a gradual increase from the first. The high point is reached later in this series than in the one immediately preceding, because while the percentage of pupils failing decreases in the final semesters (p. 14), there is an increase in the number of failures per failing pupil (Table IV).

This increase of percentages by semesters for the graduates on the total possibility of failure, as just noted, is due to an actual increase in the number of failures for the later semesters. By the distribution of failures in Table II more than 56 per cent of the failures are found after the completion of the second year, in spite of the fact that about 10 per cent of the pupils who graduate do so in three or three and a half years. The failures of the graduates are simply the more numerous after the first two years in school. That this situation is no accident due to the superior weight of any single school in the composite group, is readily disclosed by turning to the units which form the composite. For these schools the percentages of the graduates' failures that are found after the second year range from 40 per cent to 66 per cent. In only three of the schools are such percentages under 50 per cent, while in three others they are above 60 per cent.

Further confirmation of how the increase of failures accompanies the pupils who stay longer in school is offered in the facts of Table IV. Here are indicated the number of pupils who before graduating fail 1, 2, 3, etc., times, in semesters 1, 2, 3, etc., up to 10. Of all the occurrences of only one failure per pupil in a semester, 50 per cent are distributed after the fourth semester. In this same period (after the fourth semester) are found 53.2 per cent of those with two failures in a semester; 67.6 per cent of those with three failures in a semester; 71.6 per cent of those having four; 78.6 per cent of those having five; and all of those having six failures in a single semester. One could almost say that the longer they stay the more they fail.

The statements presented herein regarding the relative increase of failures for at least the first three years in school are likely to arouse some surprise among that portion of the people in the profession, with whom the converse of this situation has been quite generally accepted as true. Such an impression has indeed not seemed unwarranted according to some reports, but the responsibility for it must be due in part to the manner of presenting the data, so that at times it actually serves to misstate or to conceal certain important features of the situation. Since the dropping out is heaviest in the early semesters, and since the school undertakes the expense of providing for all who enter, it does not seem to be a correct presentation of the facts to compute the percentage of failure on only the pupils who finish the whole semester. Such a practice tends to assign an undue percentage of failures to the earlier semesters, one that is considerably too high in comparison with that of the later semesters where the dropping out becomes relatively light. It is not sufficient to report merely what part of our final product is imperfect, instead of reporting, as do most institutions outside of the educational field, what part of all that is taken in becomes waste product. This situation is sufficiently grievous to demand further comment.

In his study of the New Jersey high schools, Bliss states[28] that one of the striking facts found is the "steady decrease of failure from the freshman to the senior year." If we bear in mind that Bliss used only the promotion sheets for his data, and took no account of the drop-outs preceding promotion, and if we then estimate that an average of 10 per cent may drop out before the end of the first semester (the percentage is 13.2 for our eight schools), then the percentages of failure recorded for the first year will be reduced by one-eleventh of their own respective amounts for each school reported by Bliss, as we translate the percentages to the total enrollment basis. As a consequence of such a procedure, Bliss' percentages, as reported for the second year, will be as high as or higher than those for the first year in six of the ten schools concerned, and nearly equal in two more of the schools. It is also evident that his percentages of failure as reported for the junior and senior years are not very different from each other in six of the ten schools, although there is no inclusion of the drop-outs in the percentages stated. The only pronounced or actual decrease in the percentages of failures as Bliss reports them, occurs between the sophomore and junior years, and it is doubtless a significant fact that this decided drop appears at the time and place where the opportunity for elective subjects is first offered in many schools. Yet apparently it has not seemed worth while to most persons who report the facts of failure to compute separately from the other subjects the percentages for the 3- and 4-year required subjects.

A rather small decline is shown in the percentages of failure for the successive semesters, as quoted below for 2,481 high school pupils of Paterson[29] (the average of two semesters), although these percentages are based upon the number of pupils examined at the completion of the semester. It may further be noted that these percentages do not follow the same pupils by semesters, but state the facts for successive classes of pupils. The same criticisms may be offered for the percentages as quoted from Wood[30] for 435 pupils.

PERCENTAGES OF PUPILS FAILING, BY SEMESTERS

	SEMESTERS
	1	2	3	4	5	6	7	8
Paterson	17.8	18.4	16.7	15.0	15.6	11.6	9.4	7.4
Wood	24.5	14.5	29.5	30.0	31.0	7.9	16.2	. .
OBrien (p. 41)	11.5	13.9	14.5	15.1	14.5	15.3	12.1	9.9

The above series of percentages tend to agree at least in showing little or no decline in the percentages of failure for the first five or six semesters in school.

Another tendency to conceal important features in relation to the facts of school failures may be found in the grouping together of non-continuous and continuous subjects, the latter of which are generally required. F.W. Johnson found in the University of Chicago High School[31] that the percentage of failures by successive years indicated little or no decrease for mathematics and for English (which were 3- and 4-year subjects respectively). The figures were based on the records for a period of two years. In regard to St. Paul, it was possible to compute similar information from the data which were available.[32] The percentages of failure are presented separately in each case for Latin, German, and French, not more than two years of which are required in the schools referred to above. A contrast is thus presented that is both interesting and suggestive.

PERCENTAGES OF PUPILS FAILING, BY YEARS. (Johnson, F.W.)

	YEARS
	1	2	3	4
English	18.1	9.5	18.4	14.4
Math	12.9	12.9	13.6	5.6
Latin	14.1	9.0	2.9	. .
German	12.4	7.4	. .	. .
French	14.3	9.6	3.1	. .

PERCENTAGES OF PUPILS FAILING, BY SEMESTERS. (St. Paul)

	SEMESTERS
	1	2	3	4	5	6	7	8
English and Math	17.8	18.0	16.3	16.9	8.1	14.0	. .	. .
Latin, German, French	17.6	17.5	15.1	7.6	3.0	. .	. .	. .

Apparently the full story has by no means been told when we simply say that there is a general decline in the percentages of failure by years or semesters. First, the failures of the drop-outs should be included, so far as it is at all feasible; second, the percentage should be based on the total enrollment in the subject, not on the final product, if we wish to disclose the real situation; third, the continuous or required subjects should be distinguished in order to give a full statement of the facts. On page 41 are presented the percentages of failure for the 1,125 failing graduates alone, as found in this study, the greater portion of whose work, as it actually happened, consisted of 3- and 4-year subjects continuous from the time of entrance, and for whom the percentages of failure increase to the ninth semester.

---

7. SIMILARITY OF FACTS FOR BOYS AND GIRLS

Nowhere is there any definite indication that any of these factors of prognosis operates more distinctly or more pronouncedly on either boys or girls. Some variations do occur, but differences between the sexes in personal attitudes, social interests, or conventional standards may account for slight differences such as have been already noted. To simplify the statement of facts, no comparison of facts for boys and girls has, in general, been attempted where there was only similarity to be shown.

---

A SUMMARY OF CHAPTER III

The influence of non-attendance as a factor in school failure is partly provided for here, but no statistical data were secured.

The percentage of physical and mental defects are doubtless comparatively small for high school pupils except in the case of vision.

The facts regarding size of classes were unobtainable.

The pupils are distributed by their ages of entrance from 12 to 20, with the mode of the distribution at 15. The younger entering pupils are distinctly more successful in escaping failure. They are also strikingly more successful in their ability to graduate.

The older pupils who fail have a higher percentage of failure on the subjects taken.

The first year's record has real prognostic value for pupils persisting more than three semesters. But 57 per cent of those leaving earlier have no failures. This includes nearly 60 per cent of all the non-failing pupils, but less than 32 per cent of the failing ones have gone that early.

Prediction of failure by subjects is relatively easy and sure, and the later years seem more productive of this result.

The percentage of failure on the total possibility of failure increases with the time period up to the seventh semester. The same facts are true for the graduates when considered alone. Fifty-six per cent of the failures for the graduates occur after the second year. The longer stay in school actually begets an increase of failures. The boys and girls are similarly affected by these factors of prognosis.