;:::=_,__: g:éfgifg_ m

~i 5-

T“
u

.
V

‘ ‘ "
5” Aéioko.

 

R5

v.11...
r .u m

 

Thisistoeertfigthatthe

m
m.
w
.m.

 

Pursuit Rotor Porter-nee under Alternate Condition:

of mutilated and lend! Practice.

presented by

’00
..
’Q/

I
.' P-

9 t

Y'.

n v J70.

.M' .

n
f

of the requirements lor

h‘! degree in ?.IChOlgg
M

has been accepted towards fulfillment

 

r professor

 

2):.

 

 

 

 

PURSUIT ROTOR PERFORMANCE UNDER ALTERNATE
CONDITIONS OF DISTRIBUTED AND MASSED PRACTICE

BY

Norman Frisbey

A THESIS

Submitted to the School of Graduate Studies of Michigan
State College of Agriculture and Applied Science
in partial fulfillment of the requirements

for the degree of

MASTER OF ARTS

Department of Psychology

Year 1952

/a - IV’b’Lfa'}
6.

ACKNOWLEDGMENT

The author sincerely thanks Dr. M. Ray Denny
for his advice and the patient assistance rendered
in this research and the preparation of the manu-
script. He also appreciates the advice of Drs.

L. Katz and H. L. Harter of the Department of
Mathematics. The writer is grateful to his
wife for her constant encouragement and assist-

ance.

(”HEW‘ 1.151
rh' H. “JG 1“)

TABLE OF CONTENTS
Page

INTRODUCTION . . . . . . . . . . . . . . . . . . . . . 1
STATEMENT OF THE PROBLEM . . . . . . . . . . . . . . .
EXPERIMENTAL PROCEDURE . . . . . . . . . . . . . . . 0

Subjects. . . ... . . . . . . . . . . . . . . . . .

Apparatuflocoooococoooccocoo...

\OGJOD®UI

EXperimental Design . . . . . . . . . . . . . . . .
Instructions to Subjects. . . . . . . . . . . . . . 11
RESULTS AND DISCUSSION . . . . . . . . . . . . . . o .’ 1h
BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . . . . 38

Figure

II

III

IV

LIST OF FIGURES
Page

Performance curves for four groups for the
first and second practice sessions (6
and 12 minutes, respectively) and first
trial of third practice session as based
on the means of 30-second trials. . . . . . . 15

Performance curves for four groups for the
first (pro-rest) practice session and 6
minutes of the second (post-rest) practice
session as based on the means of lO-second
units 0 o o o c o o o o o o o o o o o o o o o 16

Theoretical deve10pmenta1 curve of I in
S R
pro-rest practice as based on
lO-BBCODd units 0 o o o o c o o o o o o o o o 26

Theoretical develOpmental curve of SIR in
post-rest practice as based on
30-second trials. 0 o o o o o o o o o o o o o 27

Theoretical extinction curve of SIR in
post-rest practice as based
on 30-second trials 0 o o o c o o o o o o o o 29

Performance curves for eight groups for
the third (3 minute) practice session
as based on the means of 30-second trials . . 33

LIST OF TABLES
Table Page
1 Schema of the eXperimental design . . . . . . . ll

2 Analysis of variance of the first 30-second
work period for 8 groups with data
converted to angles 0 o o o o o c o o o o o o 17

3 Analysis of variance of the last 30-second
work period of 6-minute practice for A
groups with data converted to angles. . . . . 18

A Mean reminiscence scores in .01 second over
the 5-minute rest period as based on
30-second trials. 0 o o c o o o o c o o o o o 19

5 Mean reminiscence scores in .01 second over
the 3-minute rest period as based on
30-second trialSo o o e o o o o o c o 0 o c o 19

6 Analysis of variance of reminiscence gain in
.01 second for 4 groups over the 5-minute
rest period as based on 30-second trials. . . 21

7 Analysis of variance of reminiscence gain in
.01 second for h groups over the 3-minute
rest period as based on 30-second trials. . . 21

8 Mean reminiscence scores in .01 second over
the 5-minute rest period as based on
lOOSGCODd units 0 o o c c o o o o o o o o c o 22

9 Mean reminiscence scores in .01 second over
the 3-minute rest period as based on
lO-second units 0 o o o o o c o o o o o o o o 22

10 Analysis of variance of reminiscence gain in
.01 second for h groups over the 5-minute
rest period as based on lO-second units . . . 23

11 Analysis of variance of reminiscence gain in
.01 second for u groups over the 3-minute
rest period as based on lO-second units . . . 23

12 Analysis of variance of the last 30-second
work period of 12-minute practice for h
groups with data converted to angles. . . . . 2S

LIST OF TABLES CONT.
Table Page

13 Mean difference between first and third
lO-second periods for first post-rest
30-second period for all groups- in
001 second. 0 o o o o c o o o o o o o c o o o 30

In Analysis of variance of reminiscence scores
over the 30-second and 5-minute rest
periods for the distributed groups. . . . . . 31

INTRODUCTION

EXperimentation over the last fifty years has resulted
in the accumulation of a great deal of data on motor learn-
ing. Two basic variables which were noted early and still
remain significant are the length of the practice interval
and the rest interval. Certain phenomena which seem to be
specifically related to these variables are the following:
(1) the advantage of spaced practice (interpolated rest
periods) over massed practice (continuous practice);

(2) reminiscence, or the gain in performance over a rest
period without additional practice; and (3) the relatively
rapid increase in performance, early in performance after
a rest, as compared with initial learning. The latter
phenomenon has been called regaining of set, or warm-up

by various eXperimenters. .Another characteristic of the
_post-rest curve is a gradual decline after this abrupt
rise before a resumption of gradual improvement such as

is observed in the pro-rest curve.

An adequate theory of motor learning should be able
to account for these performance phenomena, apply to all
other available data, and stimulate further research.

Early theoretical attempts to account for the super-
iority of distributed over massed practice and reminiscence

have been reviewed by McGeoch and Irion (9), and for one

reason or another are deemed inadequate. More recently
we have the application of the inhibitory concept of
Hull's (3) behavior theory to the area of motor learning.

Hull divides inhibition into two components, reactive
inhibition (IR) and conditioned inhibition (SIR). Reactive
inhibition is a negative drive state produced by response
which reduces the tendency to repeat that response.

IR accumulates with continued response being a positive
function of the number of reactions and the amount of work
involved in response. IR, however, is a temporary state,
dissipating rapidly with the passage of time. The other
major concept, conditioned inhibition, is a learned

resting response which reduces the drive of reactive
inhibition and is reinforced by its dissipation. More-
over, IR and SIR are assumed to have all the characteristics
respectively of other drives and habits.

The superiority of distributed practice over massed
practice in.motor learning may be attributed to the total
inhibition present in the massed practice situation which
consists of the permanent SIR and the temporary IR'
Reminiscence is due to the dissipation of IR over a rest
period. Any more or less permanent difference between
massed.and distributed practice is attributed to the
develOpment of SIR under massed conditions. There are no
concepts orplicit in Hull's system which can be used to
explain the initial sharp rise in the post-rest performance

curve e

Ammons (l), on the other hand, has a concept to
handle this phenomenon. His comprehensive system for
explaining rotary pursuit performance contains three
variables: (1) a temporary work decrement which dissipates
over rest and is similar to Hull's reactive inhibition;

(2) a permanent work decrement similar to Hull's con-
ditioned inhibition; and (3) a decrement due to necessity
to ”warm-up" after rest which is used to explain the in-
itial sharp rise in post-rest performance through regaining
of set.

Kimble (A, 5) has done much to extend and apply Hull's
concepts to motor learning phenomena. He reasons that
since IR is a drive the accumulation of a certain critical
amount will produce resting and once the IR is reduced
below the critical level the organism will resume working
until this level of IR is again attained. The critical
level of IE will probably not be reached if the inter-trial
rest periods are of more than a few seconds duration. In
fact, Kimble has shown in one meter learning situation
that no conditioned inhibition can develOp unless the
inter-trial rest is less than 5 seconds.

In relation to Kimble's hypothesis, the work of
Weaver (11) has indicated that a near maximal value of
IR is reached in approximately 30 seconds of work and is

almost completely dissipated in a rest period of the same

duration.

Both Hull and Ammons postulate that the index of
their respective permanent work decrements will increase
as a negatively accelerated (habit) function of the amount
of practice. Kimble and Weaver both have shown that SIR
shows some tendency to develOp as a negatively accelerated
(habit) function of the number of trials. In addition,
Weaver has shown a curve for the extinction of SIR which
is a decay function similar to curves found for the
extinction of other habits.

The theories of Hull, Kimble and Ammons are basically
similar and constitute the theoretical foundation for the
present investigation. However, because Hull's system
is intended for wider application, the terminology of
Hull;plus the concept of warm-up will be employed.

STATEMENT OF THE PROBLEM

This study is an attempt to isolate and specify the
three constructs which have been postulated to account for
motor learning phenomena. These are conditioned inhibition,
reactive inhibition, and warm-up or set. To accomplish
this purpose the effect of alternate conditions of dis-
tributed and massed practice on pursuit rotor performance
was investigated, and a measure of performance at lO-second
intervals within the work periods was obtained. The
design of this experiment and the analysis of the data,
but for a few exceptions, closely paralleled an investi-
gation conducted by Weaver (11), on a pursuit rotor
rotating in a counter-clockwise direction.

Reactive inhibition dissipates rapidly during the
inter-trial rest periods. Therefore, due to the lack of
inhibitory potential in distributed practice, we predict:

l. The pro-rest performance of groups under distributed

conditions will be superior to that of groups under
massed conditions.

Due to dissipation of IR during rest, we predict:

2. More reminiscence will be present in the massed

groups than in the distributed groups after an

equal amount of rest.

6

Inhibitory potential will develOp readily during the
post-rest trial in the distributed-massed group and SIR
will be extinguished in the massed-distributed group.
Therefore:

3. After the variable of distribution of practice has
had a chance to Operate in post-rest performance,
groups practicing under like conditions will con-
verge and remain together; and at the end of the
period the distributed groups (distributed-dis-
tributed; massed-distributed) will be superior
to the massed groups (massed-massed; distributed-
massed).

On the basis of the previous discussion we also pre-

dict that:

h. The index of conditioned inhibition will increase
as a negatively accelerated (habit) function of
the number of trials, and

5. The extinction curve of conditioned inhibition
will be a decay function similar to extinction
curves for other habits.

Formulation of the following hypothesis is made
'possible through using the lO-second unit of measurement
‘within a 30-second trial of distributed practice. Because
of the variable of set or warm-up we predict:

6. On the first post-rest 30-second trial the perform-

ance on the third lO-second unit will be higher
than on the first lO-second unit. The opposite

7

relationship will be true in the pro-rest 30-second
trials.

Due to the fact that IR builds up rapidly to a critical
value and dissipates rapidly with rest, reminiscence will be
evident over the 30-second rest periods in the distributed
groups soon after the first pro-rest 30-second trial. On
the basis of findings by Weaver we predict:

7. The amount of reminiscence displayed in the dis-
tributed groups over the last 30-second pro-rest
interval will not differ significantly from that
found over the 5-minute rest period for the same
group.

Conditioned inhibition has been previously built up

in the first massed practice session and extinguished in

the subsequent distributed session in the massed-distributed-
massed group. Since reconditioning is typically a rapid
process, SIR should build up very rapidly in the third

work period and as a result we expect:

8. The massod-distributedpmassed group will show the
most immediate and greatest decline in performance
of any group working under massed practice in the
third work period, particularly when compared with
the massed-massed-massed and distributed-massed-
massed groups.

Any statement regarding the final 3-minute work period

must be recognized as tentative due to the size of the

groups and the length of the period.

EXPERIMENTAL PROCEDURE

Subjects

A total of 76 college students was used as subjects.
Four were graduate students and the remainder were from
introductory psychology classes. No subject had had prior
pursuit rotor experience. Twelve of the subjects' records
were omitted in order to match the groups. The present
data were based on 6h subjects (8 groups of 8 subjects
each), 36 of whom were women. The sexes were distributed
as evenly as possible throughout the groups. Mean age
was 20.6 years.

The subjects worked singly in a quiet room. The
first R8 subjects were assigned randomly to one of the
eight groups, while the remainder were assigned to the
various groups on the basis of the first 30 seconds of
performance. This was done without interruption of the

practice session.

Apparatus
The apparatus consisted of a.Koerthntype circular
pursuit rotor, two Standard Electric timers which.measured
to the nearest 0.01 second the subject's time on target,
a hinged stylus, a stepwatch, and a double-throw four pole

toggle switch. The equipment was mounted on a wooden

table 30 inches high. The rotor disk was wood finished
with black paint and varnish. The rotor disk was 28.5 cm
in diameter with a circular brass target 1.9 cm in diameter
set flush with the larger disc 8.5 cm from its center.

The rotor turned in a clockwise direction at 60 rpm.

Experimental Design

 

The total time On target was recorded for each subject
every 10 seconds. Thirty second intervals were obtained
by simply adding 3 successive lO-second intervals. The
recording was accomplished by manual Operation of the
toggle switch every 10 seconds which simultaneously stOpped
one timer and started the other. The eXperimenter recorded
the time on target and reset the timer to zero every 10
seconds while the other timer was in the circuit. With a
little practice this was easily done in the allotted 10
seconds.

The stepwatch was used to indicate the length of the
10-second intervals, to measure the 30-second trials and
30-second rest intervals, and to measure the over-all time
for the practice and rest sessions.

The measurement of pursuit rotor performance by means
of lO-second intervals within 30-second performance trials
in the distributed practice groups is a new technique
used first by Weaver (11). It was introduced with the
hepe that it would permit a more detailed analysis of the
characteristic phenomena of the learning curve.

During the entire experimental session all subjects

10
worked a total of 21 minutes. The practice time was
divided into three periods of 6, 12 and 3 minutes separated
by two rest periods of S and 3 minutes, respectively. The
length of the various work and rest periods was determined
on the basis of previous research and eXpedience of experi-
mentation. The length of the 6-minute work period was
chosen to provide maximal reminiscence which Ammons (2)
showed occurred after approximately 8 minutes of pro-rest
practice. The 5-minute rest period was also used to
provide maximal reminiscence from pre-rest to post-rest
performance, as indicated in studies by Kimble and
Horenstein (8), Ammons (2) and‘Weaver (11). The second
work period was extended four minutes beyond that used by
Weaver to allow time for groups practicing under like
conditions to converge. It might have been more desirable
to have extended the 3-minute rest period and final 3-minute
work period but it was necessary to keep the length of the
entire practice session of the distributed groups within
50 minutes in order to obtain subjects during hours
between classes.

The various groups worked under different conditions
during the practice periods of 6, 12 and 3 minutes. Massed
practice (M) consisted of continued practice while dis-
tributed or spaced (D) practice consisted of alternating
intervals of 30 seconds of work and 30 seconds of rest.
Table 1 gives a concise picture of how the groups were

arranged.

11
TABLE 1
SCHEMA OF THE EXPERIMENTAL DESIGN

 

 

 

GI'OIIp 6 min. 5 mine 12 Min. 3 min. 3 min.
number practice rest practice rest practice

1 D - D - D*

a

2 D - D - M

3 D - M - D

H D - M - M

S M - M - M

6 M - M - D

7 M - D - M

8 M - D - D

 

*Hereafter D signifies distributed or spaced practice
and.M signifies massed or continuous practice.

It is readily observed that even though 8 groups were
used the conditions were not separated into 8 as such until
the final 3-minute period. During the first 6-minute
period we had only two conditions, D and M. With the
addition of the l2dminute period there were four conditions,
D-D”, DAM, M-M, and M-D. The final 3-minute period brings
the total of different conditions to eight as shown on the

chart.

Instructions to Subjects

Each subject was given the following written instructions:

*D-D means distributed-rest-distributed, etc.

 

 

l2

PURSUIT ROTOR EXPERIMENT
INSTRUCTIONS

The Object of this experiment is to see how
well you can follow a moving target with a hand
stylus. Stand in front of the pursuit rotor
with the stylus grasped firmly but in a relaxed
manner in your preferred hand. Keep the stylus
horizontal and move it around the turning disk
with lazy rotary movements. DO not begin until
I tell you to start, and step only when I say

St0pe

The experimenter then demonstrated the Operation as
the instructions were repeated. The subjects were told
not to begin until the experimenter said "start” and to
step when he said "step". The subjectsvvere given the
”ready” signal two seconds befOre the starting signal
and were permitted to pick up the stylus in preparation.

After the first work period in the distributed
practice groups, the subject was informed that he would
alternately work and rest. When a subject who had formerly
worked under distributed conditions was put on massed
practice he was told there would be a period of continuous
work.

The subject stood quietly in front of the rotor
during the rest interval in the case of distributed groups.
During the S and 3-minute rest periods, all subjects were

13
allowed to sit down and converse or read. If the subject

violated any of the instructions during the practice

period, he was corrected at once without interruption Of

activity.

RESULTS AND DISCUSSION

The subject's time on target was recorded to the
nearest 0.01 second. The performance curves in Figs. I
and II were Obtained by converting the time on target
for the groups into percent of time on target as has been
the usual procedure of other investigators. Analysis of
variance was used as the statistical technique for com-
parisons between groups wherever apprOpriate. A funda-
mental condition for validity of the F-test is that the
two mean squares be independent. This condition was not
met in the case Of percentages. Therefore, the percentages
were transformed into angles with mean and variance inde-
pendent as recommended by Snedecor (10). Reminiscence,
or gain over rest, was measured as the difference between
the last pro-rest trial and the first post-rest trial.
Here the gain was measured as the difference in time On
target for each individual and was not converted into
percentages. The gain distributions appear to have inde-
pendent mean and variance; consequently, a transformation
was not deemed necessary and the analysis was performed on
the raw data. In each case where analysis Of variance was
used Bartlett's test of homogenity (10) was first applied
to assure that the data were sufficiently homogeneous to

make the F-test applicable.

15

mamfiap ocooomnom MO mason on» no comes mm cofimmom ooauoean
oaanu no Adda» umaah one Ahao>a oonmoa .mopanE NH one my macammom
ooapoman ocoooo one pagan on» 90% museum anew pom mo> no ooceFLOcaoa .H ease mam

mamas» odooom banana
b ma an o p n H
-L# 1hmz_fl 0H A an r p

n dosmzna
.q fiflde4fi~fiqHfi—_d .

.

 

 

ququ uo emtq queoaed

l6

_.
.....i‘lq...‘ . - ail

mead: ocoeemnoa no modes on» no women we scammom coauoead
Apmoanpmonv ocooom one no meadows 0 use defiance ooapooad
Apmoanoadv pagan on» son masoam adom pom me>aSO oocesaomaoa .HH easmfim

means odooon sea

fifignmmumdodnugben annuanuumaaodflnoabea
__ __ __L_.La_ne.dfiwn +e_eud_e~an_4_di_aa._an_q.fi_n_an—L.q

.q
—-4
—

‘)
—
.—J

3%

> \i/V \?,‘\ (b. . L
.) ~ s
) s x ‘P« .K .
.v r \ 1
.0. ~ o
. 3 a .L
. d5
a _
...... ... . 1
...T 4
e 3 a w _ .a <
.L \ ... .. A. ..1 . i
a .
‘ ow ..A.’ .21 N. I x“
7 g .u
. ~ . .
a ..a ..x .. AI‘. ..u l
as. ..R N,” a). . «Ix .a
.....e. . .§:.~\’ .< t m
x c r. A a . a a u‘. . L
... . ,ra .u r : . _)
r .. .... a v a.
. fl QlZY.-O
. . :‘Q’086
Q89 I I...

 

.9.

m
H

qefiasq uo emyq queeaea

O
N

m
N

3 3 52 R

17

Individuals were assigned to groups on the basis Of
their performance during the first 30-second work period
in such a manner as to match the groups. The graph in
Figure I and the statistical analysis in Table 2 give
evidence that the groups were well matched at the beginning

of practice.

TABLE 2

ANALYSIS OF VARIANCE OF THE FIRST 30-SECOND WORK
PERIOD FOR 8 GROUPS WITH DATA CONVERTED TO ANGLES

 

 

 

 

 

:___

 

 

Source of variation d.f. Sggaggs 323:30 F
Between groups 7 17.29 2.h71 0.106
Within groups (error) 56 1307.58 23.350

Total 63 132h.87

 

For d.f. 7 and S6, F.05 = 3.32, F.01 = 5.85

The superiority Of groups working under distributed
conditions over groups working under massed conditions
during the first 6-minute practice period is clearly
evident from Figure I and Table 3. Also, there is no
significant difference between the groups working under
like conditions, D-D and DAM, M-M and M-D. Therefore,
hypothesis 1 is confirmed. This phenomenon was one of the
first observed to be typical of motor learning situations
and the results here are in agreement with those of other

investigators.

18

TABLE 3

ANALYSIS or VARIANCE OF THE LAST 30-SECOND WORK
PERIOD OF 6-MINUTE PRACTICE FOR A GROUPS WITH DATA
CONVERTED T0 ANGLES

 

Sum Of Mean

Source of variation d.f. squares square

 

D groups vs. M groups 1 1922.05 1922.05 32.825**

 

 

 

Groups treated alike 2 7.87 3.9M 0.067
Total between groups 3 1929.92 6h3.31 10.986**
Error 60 . 3513.31 58.56 \

Total 63 Shu3.23

For d.f. 1 and 60, p.05 = n.00, F.01 = 7.08

For d.f. 2 and 60, F.05 = 3.15, F 01 = “.098

e e '
FOP d.f. 3 and O, F.05 = 2.7 , F.01 = “-013
**Significant beyond .01 point

Gain in performance level over the 5-minute and
3-minute rest periods is Obvious from Figure I and the
statistical evidence for reminiscence is given in Tables
h and 5. All four groups with both rest periods show a
significant amount of reminiscence. The gain in each case
is a comparison Of the last 30-second pro-rest trial with
the first 30-second post-rest trial.

Due to the dissipation of IR during interpolated
rest intervals under distributed conditions we would
expect the massed groups to show more reminiscence than

the distributed groups over the same rest period. These

nan-W

__

“mm

. 19
TABLE A

MEAN REMINISCENCE SCORES IN .01 SECOND OVER THE
5-MINUTE REST PERIOD AS BASED ON 30-SECOND TRIALS

 

 

. Groups
M-M M-D D-D CDAM
Standard error of
the mean 55.98 55.27 72.7h 69.h9
t 6.060 7.808 3.236 n.698
P 4 .01 4.01 < .01 4.01

TABLE 5

MEAN REMINISCENCE SCORES IN .01 SECOND OVER THE
3-MINUTE REST PERIOD AS BASED ON BO-SECOND TRIALS

W

 

 

Groups
M-M M-D D-D D-M
Mean 5h0.62 190.62 211.56 547.00
Standard error of
the mean 5h.83 h9.71 h5.57 81.51
t 9.860 3.834 h.6h3 6.711

p < .01 (.01 (.01 4.01

 

 

20
comparisons are presented in Tables 6 and 7. The difference
between groups practicing under like conditions is not
significant. However, hypothesis 2, that massed groups
will show greater reminiscence than distributed groups,
is confirmed. Weaver found the massed groups showed
greater reminiscence than the distributed groups only at
the 10 percent level of significance. This lower level
of significance may be due to the fact that counter-clock-
wise rotation increased the difficulty Of the task and

consequently set up more I in the distributed groups

R
within a 30-second work period.

The graph in Figure II which shows the 6-minute
practice period and 6 minutes of the 12-minute period
was Obtained by plotting performance level of the h groups
using lO-second units of measurement. A statistical
analysis Of reminiscence gain from the last pre-rest
lO-second unit to the first post-rest lO-second unit is
presented in Tables 8 thrOugh 11. Tables 8 and 9 indicate
that when only the right tail Of the t distribution is
considered, the hypothesis Of no gain can be rejected at
least at the 5 percent level of significance for all groups
except the M-D group over the 34minute rest period. The
M-D group is approaching the maximum performance level
for this task and consequently reminiscence over this
and any succeeding rests may not prove to be significant.

A comparison of reminiscence between the various groups

and conditions over the 5 and 3-minute rest periods is

go. .~ --.-:*.—r~t~—.-----.*

21
TABLE 6

ANALYSIS OF VARIANCE OF REMINISCENCE GAIN IN .01 SECOND
FOR A GROUPS OVER THE S-MINUTE REST PERIOD AS
BASED ON 30-SECOND TRIALS

W

Sum of Mean

Source of variation d.f. squares square F

 

D groups vs. M groups 1 6,5h8,518 6,5h8,518 100.57S**

 

 

Groups treated alike 2 138,386 69,193 1.063
Total between groups 3 6,686,90h 2,228,968 3h.233**
Error 60 3,906,680 65,111

Total 63 10.593.58h

 

See footnote Table 3 for .01 and .05 F values
**Significant beyond .01 point

TABLE 7

ANALYSIS OF VARIANCE OF REMINISCENCE GAIN IN .01 SECOND
FOR h GROUPS OVER THE 3-MINUTE REST PERIOD AS
BASED ON 30-SECOND TRIALS

"

 

Sum of Mean
Source of variation d.f. squares square F

D groups vs. M groups 1 9,869,6h7 9,869,6h7 l7h.086**

 

 

Groups treated alike 2 3,832 1,916 .03h
Total between groups 3 9,873,h79 3,291,159 58.OSI**
Error 60 3,A01,657 56,694

Total 63 13,275,136

 

See footnote Table 3 for .01 and .05 F values
**Significant beyond .01 point

TABLE 8
MEAN REMINISCENCE SCORES IN

22

.01 SECOND OVER THE

S-MINUTE REST PERIOD AS BASED ON lO-SECOND UNITS
W

 

 

 

 

 

 

 

 

Groups
M-M M-D D-D D-M
Mean 62.81 82.19 56.06 99.87
Standard error of 1
the mean 35.66 21.6h 30.78 30.50 g
1%
t 1.761 3.797 1.821 3.27M }
P 4 .05 4.01 4. .05 < .01
TABLE 9
MEAN REMINISCENCE SCORES IN .01 SECOND OVER THE
3~MINUTE REST PERIOD AS BASED ON 10-SECOND UNITS
Groups
M-M M-D D-D DéM
Standard error of
tha mean 20014.8 3606).... 38.22 29.65
t 5.998 1.100 2.182 h.018
P 4 .01 >.05 < .05 4.01

 

23
TABLE 10

ANALYSIS OF VARIANCE OF REMINISCENCE GAIN IN .01 SECOND
FOR A GROUPS OVER THE 5-MINUTE REST PERIOD AS
BASED ON lO-SECOND UNITS

   

 

 

 

Source of variation d.f. Sum of Mean F
squares square
D groups vs. M groups 1 A78 R78 0.330
Groups treated alike 2 17,359 8,680 0.599
Total between groups 3 17,837 5,9h6 0.th‘
Error 60 869,216 lh,h87
Total 63 887:053

 

See footnote Table 3 for .01 and .05 F values

TABLE 11

ANALYSIS OF VARIANCE OF REMINISCENCE GAIN IN .01 SECOND
FOR h GROUPS OVER THE 3AMINUTE REST PERIOD AS
BASED ON 10-SECOND UNITS

 

 

 

 

Source of variation d.f. Sum 0f Mean F
squares square
D groups vs. M groups 1 55,932 55,932 3.032
Groups treated alike 2 lh,9hh 7,h72 0.h05
Total between groups 3 70,876 23,625 1.281
Error 60 1,106,83h l8,hh7
Total 63 1.177.710

 

See footnote Table 3 for .01 and .05 F values

..m may... gar—‘—

211
presented in Tables 10 and 11, reapectively. The analysis
shows that no significant difference is found between
distributed and massed groups or between groups under like
conditions in either case. These data substantiate the
results obtained by Weaver when he calculated reminiscence
gain by means of lO-second units Of measurement.

Examination of Figure I shows that after approximately
6 minutes Of post-rest practice the performance of the D-M
group has fallen to that of the M-M group and the per-

formance Of the M-D group has risen to the level of the

 

D-D group and that groups practicing under like conditions
remain at the same levels of performance. Table 12 compares
the groups at the end of the practice period and indicates
that the D groups are superior to the M groups and that
groups under like conditions are not significantly different.
Therefore, hypothesis 3 is confirmed. The length Of the
second practice period used by Weaver was 8 minutes. At

the end Of the period his groups practicing under like
conditions had not converged although they were not statis-
tically different. An unpublished study conducted at
Michigan State College which also used clockwise rotation
found as in the present investigation that the like groups
converge in less than 8 minutes. The discrepancy between
these results and those found by Weaver may be due to the
fact that counter-clockwise rotation generates a higher
level of conditioned inhibition.

Figures III and IV are develOpmental curves of con-

25
TABLE 12
ANALYSIS OF VARIANCE OF THE LAST 30-SECOND WORK

PERIOD OF lZ-MINUTE PRACTICE FOR A GROUPS WITH DATA
CONVERTED TO ANGLES

 

 

 

Sum of Mean

Source of variation d.f. squares square F

 

D groups vs. M groups 1 3h65.2h 3h65.2k 76.546**

 

 

Groups treated alike 2 68.07 3h.03 0.752
Total between groups 3 3533.31 1177.77 26.016**
Error _ 60 2716.19 h5.27

Total 63 62E9.SO

 

See footnote Table 3 for .01 and .05 F values

**Significant beyond .01 point
ditioned inhibition. The curve in Figure III shows the
develOpment of SIR’ during the pro-rest 6-minute practice,
in the groups practicing under M conditions. It was con-
structed by using the data from Figure II. The successive
differences between mean performance level of the distributed
groups and the massed groups were taken starting with 10-
second unit #6 and using every third lO-second unit there-
after. These differences were plotted and a free-hand
curve was drawn. This method is based on the assumptions
that IR develOps to a maximal value in approximately 30
seconds of work in the distributed practice group and dis-
sipates almost completely in 30 seconds of rest.

During post-rest performance the DAM group should

26

kl’lll.:lr‘lfl . ...

mead: ocooomtoa do women we oOHpOeaa umoaloan

m m

EH H mo o>a30 Heudoedoao>eo Hmoapoaoone

bu

mead: ozooom no» unenloam

mu nn «H m o

n

.HHH enemas

 

—4 4...“ 4.110% 4

w

.jLfleqdn.a.1; S—I__ _J L— ea LATE

d

 

‘fl *0 F\ (N :4

r- so
quasq uo emIq queoaed

(O

543‘”

27

mamaap ocooonnom do

Conan mm compound peonuumoe EH mHm
ho ophfio HensoSdoao>oo Hooapoaoose .bH opsmwm

panda» pdooon madame pmoalpmoa

 

m . ... e A
_ _ _ _ _

 

a “l '1
qeflasq uo equ queoaed J0 [sooadtoeu

‘t

h
e

28
develOp SIR and therefore the reciprocal of the difference
between the DéM and the M-M groups should indicate the
develOpment of SIR‘ This was the method used to obtain
the curve in Figure IV.

Figure III indicates that the growth Of conditioned
inhibition may well be a negatively accelerated (habit)
function. Figure IV lacks sufficient points to be a good
indicator because the groups converged rapidly. However,

hypothesis h is adequately confirmed.

I“! ‘m‘h “HRH

Figure V is an extinction curve of SIR derived by

 

taking successive differences between 30esecond work periods
of groups M-D and D-D during the post-rest practice period.
The curve is a decay function similar to curves found for
the extinction Of other habits. This supports hypothesis

5. These curves are similar to those found by Weaver in
his study and the SIR curve resembles those found by

Kimble (h) using a considerably different method of der-l
ivation.

The use Of lO-second measuring units within the 30-
second work intervals permits further analysis of the per-
formance not otherwise available. From Figure II it can be
seen that the third lO-second unit of the first 30-second
post-rest trial is higher than the first lO-second unit.
Also the graph shows that in the case of the pro-rest
30~second trials the Opposite relationship is true. The
statistical analysis is given in.Table 13. This finding

confirms hypothesis 6 concerning the variable of warm-up.

29

 

mamas» ocooomaom so oommL mm
evapoead umoauumon ca mHm mo o>ndo cofiuocHDXo Heoapmaooze

mamas» odooom hpafinu pmoanumoa
ma NH. m3” Wm «H nH NH Hm on a w b e m w m

e> madmam

 

   
 

_ .LL~flNLL—L—d_

——1

  

 

 

 

as a: r- ‘o in -e ex ea
403193 uo emIq queoaea

Y

 

:11 S

30
TABLE 13
MEAN DIFFERENCE BETWEEN FIRST AND THIRD 10-SECOND

PERIODS FOR FIRST POST-REST 30-SECOND PERIOD FOR
ALL GROUPS IN .01 SECOND

 

 

All groups
Mean 57.95
Standard error Of the mean 16.18 p:
t . 3.583
P .4;.01

 

 

‘F‘. ;=‘— '

The results at this point are predicatable from the theoret-
ical analysis of Hull (3), Ammons (1), and Kimble (E, 5)

and are in basic agreement with the empirical findings Of
Ammons (2), Kimble (h, 5, 6), Kimble and Horenstein (8),

and Weaver (11).

Still other observations are made possible using the
lO-second unit Of measurement as a basis for analysis.
Figure II shows that reminiscence is present after the first
30 seconds of work in the distributed practice groups as
indicated by the gain in performance level between 10-
second units #3 and h. The distributed groups continue
to show considerable reminiscence over each succeeding
30-second rest while no such increases are evident in the
massed practice groups. These results are in agreement
with the findings of Weaver (11) upon which hypothesis 7

was based.

31
A comparison of the amount of reminiscence in the
distributed groups between 10-second units #33 and 3A and
between lO-second unit #36 and the first lO-second unit Of
the 12-minute post-rest practice period is made in Table 1h.
The last 30-second rest period was chosen for this compar-
ison because the performance level at this point closely

approximates that of the 5-minute rest interval.

TABLE 1k

ANALYSIS OF VARIANCE OF REMINISCENCE SCORES OVER THE
30-SECOND AND 5-MINUTE REST PERIODS FOR THE
DISTRIBUTED GROUPS

 

 

 

 

 

 

Source Of variation d.f. 333.35: 823239 F
30-Second vs. 5-minute

rest 1 13,806 13,806 0.892
Groups with same

interval 2 23,u90 11.7u5 0.759
Total between groups 3 37,296 12,h32 0.803
Error 60 928,707 15,h78

Total 63 966,003

 

See footnote Table 3 for .01 and .05 F values

Hypothesis 7 is confirmed as Table 1h indicates no
significantly different amount of reminiscence over the
two rest periods.

The breakdown of the four groups into eight groups
for the final practice period of 3 minutes is presented in

VT

32
Figure VI. In the M-DéM group conditioned inhibition was
built up in the first practice session and extinguished
in the following distributed session. Because recondition-
ing is a rapid process, SIR should build up quickly and
consequently this group should show the greatest decline
in performance of any group under massed practice in the
third work period. This is evident from Figure VI and
thus hypothesis 8 is substantiated. The characteristic
post-rest hump phenomenon predicted and found by Ammons
(l, 2) is also definitely observed in the DHM-D, M-D-D,
M-M-M, M-M-D, and DAM-M groups.

In addition to supporting Ammons' notion of warm-up
decrement (1), use of the lO-second unit has made possible
three important findings. First, reminiscence in the dis-
tributed groups is not significantly less than that found
in the massed groups over rest periods Of the same length.
This implies, among other things, that when measurement is
made at the last lO-second unit before rest the difference
in performance level between the distributed and.massed
groups is due entirely or almost entirely to the presence
of SIR‘ Second, the distributed groups show reminiscence
after the first 30 seconds of work in the pro-rest period.
Third, in the distributed groups reminiscence over the
5~minute rest does not differ significantly from the
reminiscence over the last 30-second rest period in the
pro-rest practice session. Consideration Of these points
lead us to conclude in accord with Kimble (A) that there is
a critical level to which IR can rise. Also, the maximal

33

manage Uncoomuom LO macra on» C0 Somme mm nowmmom oofipomcn

AopchE mv pads» on» peg museum unmam no“ mo>a50 oocmchomaoe

 

 

 

EIEIQ
m M” A
.1 1 L .1 1 mm
13
13.
QIQIZ
m n H
fiafiBfiS “Q
Lem
.1 mm

 

madame Uncoom hepana

 

 

 

 

 

Slat: QIEIE
.o¢ 1.m 7. m R M7 on
ulni Ilnn
Q8219 SIQIQ
n n H m n H
1|o7414 fiflm mm _ .737 1 _ _ on
J3 I an

 

 

 

 

 

El 2.. E
m m H
. 7. a . _ 4
QDQIQ
n n H
1 own To _ _<

 

.H> enemaa

ow

mw

O
in

n
Q

ququ uo emIq queoaed

3

mm

3h
value of IR possible under well distributed conditions is

approximately equal to the maximal value possible under

massed conditions. In other words, I builds up rapidly

R

to this maximal level - in approximately 30 seconds; and

with 30 seconds of rest shows almost complete dissipation.
Conditioned inhibition has been postulated to be a

habit as defined by Hull (3). Consequently, curves Of

develOpment and extinction of SIR should resemble those

of other habits. We have presented two curves showing

the develOpment of SIR derived by different methods. The

derived curve Of extinction resembles that of other extinction

curves and the expectation of rapid reconditioning is

fulfilled. These various lines Of independent evidence

clearly substantiate the concept Of conditioned or learned

inhibition as postulated in the recent theoretical treat-

ments of motor skills phenomena.

SUMMARY

Eight groups of 8 subjects each worked on the pursuit
rotor task under alternate conditions of distributed and
massed practice. The groups were: (1) massed-massed-
massed, (2) massedpmassod-distributed, (3) massed-distributed-
massed, (h) distributed-massedamassed, (5) distributed-
distributeddmassed, (6) distributed-massed-distributed,

(7) massed-distributed-distributed, and (8) distributed-
distributed-distributed.

There were three practice sessions of 6 minutes, 12
minutes, and 3 minutes separated by two rest periods of
5 and 3 minutes, respectively. Massed conditions were
continuous practice, while distributed conditions were
alternately 30 seconds of work and 30 seconds of rest.

The pursuit rotor disk rotated in a clockwise direction
at a Speed of 60 revolutions per minute.

The most important experimental technique involved
measurement of performance in lO-second units of work
within a 30-second work trial for the distributed practice
groups.

Eight statements relating to the constructs of reactive
inhibition, conditioned inhibition, and set were tested and

confirmed. They are:

36

1. Pre-rest performance of groups under distributed
conditions will be superior to that of groups under massed
conditions.

2. More reminiscence will be present in the massed
groups than in the distributed groups after an equal
amount of rest.

3. In post-rest practice groups practicing under
like conditions will converge and the distributed group's
performance will be superior to that of the massed at the
end Of the period.

A. The develOpmental curve Of conditioned inhibition
will increase as a negatively accelerated (habit) function
of the number of trials.

5. The extinction curve of conditioned inhibition
will be a decay function similar to extinction curves of
other habits.

6. 0n the first post-rest 30-second trial the per-
formance on the third lO-second unit will be higher than
on the first lO-second unit. The Opposite relationship
will be true in the pro-rest 30-second trials.

7. The amount of reminiscence diSplayed in the dis-
tributed groups over the last 30-second pro-rest interval
will not differ significantly from that found over the
5-minute rest period for the same group.

8. The massed-d1stributed-massed group will show the
most immediate and greatest decline in performance Of any

group working under massed practice in the third work period.

37

The conclusions regarding reactive inhibition were:

1. There is a maximal level to which reactive inhi-.-
bition can rise which is approximately the same under
distributed or massed conditions.

2. Reactive inhibition builds up to a maximal value
in approximately 30 seconds.

3. Reactive inhibition dissipates almost completely
in 30 seconds.

With regard to conditioned inhibition it seems that
several independent lines Of evidence substantiate the

notion of a learned type of inhibition in motor learning.

l.

2.

3.

7.

8.

9.

10.

11.

BIBLIOGRAPHY

Ammons, R. B. Acquisition Of motor skill: I. Quan-
titative analysis and theoretical formulation.
Psychol. Rev., 54: 263-281. 1947.

Ammons, R. B. Acquisition Of motor skill: II. Rotary
pursuit performance with continuous practice before
and after a single rest. J. EXp. Psychol., 37:

393'u11e 19u70

Hull, C. L. Principles of behavior. D. Appleton-
Century, New'York. 1943.

Kimble, G. A. An experimental test of a two factor
theory Of inhibition. J. Exp. Psychol., 39: 15-23.
1949.

Kimble, G. A. A further analysis of the variables
in cyclical motor learning. J. Exp. Psychol., 39:

332-337. 19h9.

Kimble, G. A. Performance and reminiscence in motor
learning as a function of the degree Of distribution
of practice. J. Exp. Psychol., 39: 500-510. l9h9.

Kimble, G. A. and E. A. Bilodeau. Work and rest as
variables in cyclical motor learning. J. Exp.“
Psychol., 39:150-157. 1949.

Kimble, G. A. and B. R. Horenstein. ReminiscencO in
motor learning as a function of length of interpolated
rest. J. Exp. Psychol., 38: 239-244. 1948.

L__..

\

"-L

McGeoch, J. A. and A. L. Irion. The psychology of
hugan learning. Longmans, Green & Company, New'YOrk.
l9 2. '

Snedecor, G. W. Statistical methods. Iowa State
College Press, Ames, Iowa. 1946.

Weaver, J. An experimental investigation of the
comparative effect Of massed and spaced pro-rest
practice upon both massed and Spaced post-rest
performance on the pursuit rotor task. Unpublished
Masters Thesis, Michigan State College. 1950.

 

 

 

”Tfll'filfjfllflljijjTH flifljflijflifijflifljfififl“