The effect of attendance upon school achievement

THE UNIVERSITY

OF ILLINOIS

LIBRARY

370

F«tt

>. I G

UNIVERSITY OF ILLINOIS BULLETIN

Issued Weekly
Vol. XX April 2, 1923 No. 31

[Entered as second-class matter December 11, 1912, at the post office at Urbana, Illinois, under the
Act of August 24, 1912. Accepted for mailing at the special rate of postage provided for in
Section 1103, Act of October 3, 1917, authorized July 31, 1918.]

EDUCATIONAL RESEARCH CIRCULAR NO. 16

BUREAU OF EDUCATIONAL RESEARCH
COLLEGE OF EDUCATION

THE EFFECT OF ATTENDANCE UPON
SCHOOL ACHIEVEMENT

By

Charles W. Odell

Associate, Bureau of Educational Research

PUBLISHED BY THE UNIVERSITY OF ILLINOIS
URBANA

370

tl o . I&

The Effect of Attendance Upon School Achievement

In connection with a recent experiment1 in the promotion and
classification of pupils the writer had to determine the placement of
a large number of pupils. As a basis for so doing various items of
information concerning each pupil were collected. Among these items
was a record of the scores made upon several intelligence and achieve-
ment tests. These tests were given near the close of each of the sev-
eral semesters that the experiment continued. The attendance record
showed the percent of time that each pupil was present from the first
of each semester until the time the tests were given. In the case of
pupils who were not enrolled at the first of the semester their attend-
ance records were computed on the basis of the time elapsing after
they entered. Another item of information secured was the average
school mark. This mark was given by the classroom teacher and was
the average of the marks made in the following seven subjects, or in
as many of them as the pupil happened to be carrying: Arithmetic,
English, Geography, History, Reading, Spelling and Writing. It was
computed for the same period of time as the percent of attendance,
that is, from the first of the semester until the time that the tests
were given.

In carrying out the placement of pupils it was in many cases
necessary to confer with the various teachers taking part in the ex-
periment. In discussing what effect the attendance record of a pupil
should have upon his placement these teachers expressed two opin-
ions that were radically different. Many of them believed that if a
pupil had been absent any considerable portion of a semester he
should not be promoted at the end of the semester. They argued
that no matter how bright he might be he had missed a certain
amount of work that he should have covered and should not go ahead
until he had completed it. Furthermore, if it was suspected that
attendance had been poor largely as the result of disinclination to
attend school these teachers believed that it was only just that the

^This experiment is described in Bulletin No. 12 of the Bureau of Educational
Research.

[3]

pupil should be penalized by being failed. They said that if he were
allowed to go ahead it would produce the impression that attendance
was not an important matter and thus encourage further truancy.
On the other hand many teachers took the opposite viewpoint. If a
pupil had done rather poor work during the semester because of a
large amount of absence they said that if his ability justified it he
should not be denied promotion. This was supported by the argu-
ment that the pupil would have been able to do the work if he had
been present and therefore should not be held back. This argument
was of course applied only in the cases of pupils of average or supe-
rior ability. Furthermore, it was said that in the case of a pupil who
stayed out of school because he disliked to attend the condition would
be made worse by failing him as this would result in still greater
dislike of school. Because of these two radically divergent viewpoints
the writer was interested in making a study of the question named
above in order to see what effect attendance or non-attendance
actually produces upon the achievement of pupils.

As was stated, certain achievement tests were given near the
end of each of the several semesters. These tests covered reading
and arithmetic and also some spelling in the lower grades. The scores
made thereon were converted into achievement ages and quotients.
By computing the gain from one testing period to another it was
possible to find the approximate increase in achievement age during
each semester. Table I shows the percent of pupils included in each
attendance group, their median achievement age at the end of the
semester, their median achievement a-Uotient at the same time, their
median increase in achievement age during the semester and their
median average school mark for this time. It is to be interpreted as
follows: approximately 75.3 percent of all pupils attended from
91-100 percent of the time. This group had a median achievement
age at the end of the semester of 11.1 years, a median achievement
quotient of 107.8, a median increase in achievement age of 1.4 years
and a median average school mark of 3.I.1 From the third column
it appears that there is a definite tendency for higher achievement

xThe average school marks were computed on a scale of 5 points, 1 being the
poorest and 5 the best. These numerical marks were equivalent to the following ex-
pressions: l=poor; 2=fair; 3=good; 4=excellent; 5=superior.

[4]

TABLE I. RELATION OF ATTENDANCE TO ACHIEVEMENT
DURING THE SAME SEMESTER

Percent

Percent

Median

Median

Median

Median

of

of Pupils

Achievement

Achievement

Increase

Average

Attendance

Included

Age

Quotient

in A. A.

School Mark

91-100

75.3

11.1

107.8

1.4

3.1

81-

15.9

10.6

107.2

1.3

2.9

71-

4.7

10.4

106.8

1.2

2.7

61-

1.7

10.4

105.4

1.5

2.6

51-

.8

10.2

109.2

1.6

2.4

41-

.7

10.4

113.3

1.6

2.5

31-

.3

9.4

103.3

.0

2.4

21-

.2

9.2

95.0

.3

2.4

11-

.2

9.0

103.3

.5

2.5

1-

.2

9.9

90.0

1.5

2.4

All

100.0

11.0

107.5

1.4

3.0

No. of Cases

23,036

6,874

13,746

22,803

ages to be associated with better attendance. The median achieve-
ment quotients and increases in achievement age show this tendency
very slightly if at all. There is a small increase in the medians of
the three highest attendance groups which of course contain the bulk
of the pupils. But on the other hand there are several groups lower
than these whose medians are larger. The lowest medians are, how-
ever, found among the groups that have poor attendance records.
The column of average school marks tends to agree with column
three. Although there is no appreciable change one way or another
in the lower half, in the upper the medians rather steadily increase
along with better attendance. It is noteworthy, however, that even
in the group with the lowest percent of attendance the median school
mark is 2.4, or almost half-way between "fair" and "good."

The median percents of attendance for all pupils of each given
achievement age, quotient, increase, and school mark were also com-
puted. A study of these medians shows the same tendencies that
have been mentioned; the pupils who had the higher achievement
ages and school marks tended to be those with the higher percents
of attendance, whereas this was very slightly if at all true in the case
of the achievement quotients and increases in achievement age. The
evidence presented by Table I and the medians just referred to is a

[5]

little conflicting. A common-sense view of the situation would appear
to be that attendance during a given semester is more closely asso-
ciated with the work done during that semester, that is the increase
in achievement, than it is with the achievement age or quotient at
the end of the semester, which represents total achievement since
entering school. This is, however, apparently not the case in so far
as the figures given above show. The evidence of the school marks
tends to corroborate what we should expect, that the pupils who at-
tend better secure higher marks.

Partly because of this unexpected result and partly for other
reasons it was thought desirable to study the relation of attendance
during one semester to achievement during the next. Table II pre-
sents the same data as certain of the columns of Table I except that
they are computed for the semester following that for which the at-
tendance was taken. The data in this table do not show strikingly dif-
ferent tendencies from those in Table I. Perhaps the most noteworthy
difference is that the median increases in achievement age for the three
lowest attendance groups are very high. A possible explanation for this
is that the pupils in these groups tended to make small increases during
the same semester for which the attendance record was taken and
therefore during the following semester they balanced this by mak-
ing unusually large ones. This might have been caused by pressure
on the part of the teacher or hard effort and diligent application on
the part of the pupils. The tendency for the median average school
mark to increase along with attendance is somewhat less noticeable
than was shown by the data in Table I but is still present.

Summing up the evidence it appears (1) that the percent of
time which a pupil attends school has a rather definite effect upon
his achievement age at the end of the period under consideration,
(2) that it has practically no effect upon his actual increase in achieve-
ment age unless his attendance record is very poor in which case it
lessens it for the same semester and increases it for the following
one, and (3) that it is fairly closely related to his average school
mark for the same semester and somewhat less so to that for the
following one. On the whole the attendance appears to be a factor
conditioning achievement but not so weighty a factor as many have
believed.

[6]

TABLE II. RELATION OF ATTENDANCE TO ACHIEVEMENT
DURING THE FOLLOWING SEMESTER

Percent

Percent

Median

Median

of

of Pupils

Increase

Average

Attendance

Included

in A. A.

School Mark

91-100

75.3

1.3

3.1

81-

15.9

1.5

3.0

71-

4.7

1.0

2.9

61-

1.7

1.1

2.9

51-

.8

.8

2.6

41-

.7

.8

2.9

31-

.3

1.9

2.7

21-

.2

3.0

3.0

11-

.2

2.5

2.9

1-

.2

3.3

3.1

All

100.0

1.3

3.1

No. of Cases

7,074

12,773

In connection with this study of attendance it was thought worth
while to determine whether or not pupils tend to be present at school
about the same percent of the time semester after semester. In other
words, are most of the pupils who attend rather regularly any one
given semester likely to do so during the succeeding one and vice
versa? In order to answer this question the attendance records dur-
ing the first and second semester of the same school year were cor-
related. The coefficient of correlation was found to be .19 ± .01,
which indicates that there is a small but definite amount of correla-
tion. This is further shown by computing the median percent of
attendance for the second semester of each of the ten-percentile
groups for the first. Although these medians do not run in regular
order they show on the whole an increase from the lower groups to
the higher. This is especially marked in the six highest groups. A
similar condition is found when the attendance groups for the second
semester are compared with the corresponding medians for the first.
These data would seem to point to the fact that poor attendance is
not caused by persistent truancy so often as has been supposed. It
must be remembered, however, that these data were secured from a
portion of the Chicago school system and that in this system the

[7]

compulsory attendance department is rather efficient, enough so that
it is difficult for pupils to be absent without a good cause very long
at a time.

Another matter of interest is whether or not there is any relation
between the intelligence of pupils and their attendance at school. In
order to investigate this the records of those pupils who were in
school during the four semesters for which data were obtained were
studied. The average intelligence quotient resulting from four mental
tests1 was correlated with the average attendance during four semes-
ters. These data were complete for almost 4000 pupils. The coeffi-
cient of correlation was found to be .06 ± .01, which shows that
there was practically no correlation. A study of the medians supports
this. The median I. Q. for the pupils who attended from 91 to 100
percent of the time was the highest of that of any of the groups, but
except for this the medians do not show any general tendency in
either way. There is also a very slight tendency for the median per-
cents of attendance of the groups having different intelligence quo-
tients to increase. We are not justified, however, in concluding from
this study that there is any relation between intelligence and attend-
ance in school.

Those who wish to compare the attendance records of their
pupils with those of the pupils in this experiment may do so by
means of the second column in each of the two tables. This shows
that approximately three-fourths of the pupils attended from 91 to
100 percent of the time. The median attendance for all pupils con-
cerned was 94.4 percent. It was also found that, although consid-
ering the semesters separately about 2.2 percent of the pupils attended
less than 61 percent of the time, when the average attendance of
those pupils in school during the whole four semesters was computed
only about .2 percent had attended less than 61 percent of the total
time. This is what might be expected from the fact that the corre-
lation of attendance in succeeding semesters was found to be rather
low.

*In some cases the pupils were tested with two different mental tests, one of
which was used three times, whereas in others three different tests were used, one
of them being used twice.

[8]

A

X-