RELATEONSHIP BETWEEN THE ORiENTlNG
REFLEX AND SEREAL LEARMNG

Thesis for the Degree of M, A.
MECHIGAN STAYS. UNWERSITY
GERRH E. DeYOUNG
1969

   
 

  

7.4 9:515

LIB R A R Y
Michigan 8 me
University

    

     
 
 

  

g" BINDING BY

HMS 8: SONY
BINDERY WC.

W3“!

  
  

ABSTRACT

RELATIONSHIP BETWEEN THE ORIENTING REFLEX
AND SERIAL LEARNING

BY

Gerrit E. DeYoung

Isolation of an item in a serial list has been
found to produce faster learning of at least the iso-
lated item, an effect which is called the von Restorff
phenomenon. The orienting reflex, having physiological
concommitants, is an objectively measureable response
which has been shown to be related to verbal learning.
Since the isolation of a stimulus is likely to result
in a large orienting reflex to that item, it is possible
that the von Restorff effect may be explainable in terms
of the effect of the orienting reflex. Since the orient-
ing reflex was not measured in the studies reporting a
von Restorff effect, no direct evidence concerning the
relationship existed.

Three experimental conditions were selected for
the current research, each condition being run in a warm

1

Gerrit E. DeYOung

room and receiving a serial learning task using nonsense
syllables. In the puff isolate condition. the seventh
item of the list was isolated by being paired with a cool
air puff in the warm room; in the puff control condition.
the air puff was presented prior to the first item of the
list. The control condition received no air puff and was
simply the typical serial learning paradigm. The GSR as
a measure of the orienting reflex was recorded in all
conditions.

The following predictions were advanced: 1) The
learning of the isolated item in the puff isolate condi-
tion would be enhanced. since the air puff should elicit
an OR. 2) The learning of the first item in the puff
control condition should not be enhanced. since the first
item of a serial list is already isolated by its primary
position and the position alone should yield a large OR;
this condition should produce results similar to condi-
tion C. 3) In all conditions. the GSR should be nega-
tively correlated with the number of trials to the last
error for each item; that is large ORs should be asso-

ciated with faster learning.

Gerrit E. DeYbung

The expected von Restorff effect was not observed.
It was concluded that the lack of von Restorff effect was
due (1) to the long interstimulus interval and (2) to the
possible inhibitory effect of the puff itself upon reSpond-
ing. A significant correlation was found between the GSR

to each item and the Speed of learning of the item.

Thesis Committee:

David C. Raskin, Chairman
James L. Phillips

Richard Hart

Approved:

 

Date:

 

RELATIONSHIP BETWEEN THE ORIENTING REFLEX

AND SERIAL LEARNING

BY

Gerrit E. DeYoung

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF ARTS

Department of Psychology

1969

ACKNOWLEDGMENTS

The author wishes to express his thanks to
Dr. David Raskin, Dr. James Phillips, and Dr. Richard
Hart for their continued assistance and advice. Their
suggestions improved the thesis in many ways.
Gratitude is expressed to Dr. Charles Wrigley
and the Computer Institute for Social Science Research
for the provision of computer time and keypunching

assistance for the analysis of the data.

ii

TABLE OF CONTENTS

Page

ACKNOWLEDGMENTS . . . . . . . . . . . . . . . . . . ii

LIST OF TABLES. . . . . . . . . . . . . . . . . . . iv

LIST OF FIGURES . . . . . . . . . . . . . . . . . . v
INTRODUCTION. . . . . . . . . . . . . . . . . . . . l
The Orienting Reflex. . . . . . . . . . . . . . 2
Relationship between the OR and Learning. . . . 3
The von Restorff Effect . . . . . . . . . . . . 8
Statement of the Problem. . . . . . . . . . . . 9

METHOD. . . . . . . . . . . . . . . . . . . . . . . 12

RESULTS . . . . . . . . . . . . . . . . . . . . . . 16

DISCUSSION. . . . . . . . . . . . . . . . . . . . . 23

REFERENCES. . . . . . . . . . . . . . . . . . . . . 26

APPENDIX A: CONSTRUCTED LISTS. . . . . . . . . . . 28

APPENDIX B: INSTRUCTIONS . . . . . . . . . . . . . 29

iii

LIST OF TABLES

Table Page
1. Analysis of Variance of Mean Trial to Last
Error for All Serial Positions. . . . . . . 16

2. Analysis of Variance of Mean Trials to Last
Error, Serial Position 1. . . . . . . . . . 18

3. Analysis of Variance of Mean Trials to Last
Error, Serial Position 7. . . . . . . . . . 19

4. Analysis of Variance for Mean Trial to Last
Error, Serial Position 8. . . . . . . . . . 19

iv

L IST OF FIGURES

Figure Page

1. Mean trials to last error as a function of
serial position. . . . . . . . . . . . . . 17

2. Mean percent change in skin resistance as a
function of serial position. . . . . . . . 20

RELATIONSHIP BETWEEN THE ORIENTING REFLEX

AND SERIAL LEARNING
INTRODUCTION

Isolation of an item in a serial list by some
method such as changing the color of the item (Jensen,
1962) or applying shock in temporal contiguity with the
item (Raskin, Hattle, & Ruble, 1967) has produced faster
learning of the isolated item. This effect is called
the von Restorff phenomenon. Several explanations have
been advanced for this phenomenon, including one derived
from the principles of gestalt psychology (Koffka, 1935),
and one derived from interference theory (Gibson, 1940,
1942).

Recent experimental research of Sokolov (1963)
and Maltzman and Raskin (1965) and others has involved
the concept of the orienting reflex (OR). This reflex,
having physiological consommitants, can be measured dur-
ing ongoing activity of an organism, allowing inferences
to be made concerning the relationships between the OR

1

and the ongoing activity. Several lines of evidence sug-
gest that the OR may be strongly related to learning of
several types. Brown (1937) found high correlations be—
tween the order of learning of items in a serial list
and the GSR to the items. Belloni (1964) found high OR
gs superior to low OR gs on a paired—associate learning
task. It is the purpose of this study to investigate the
relationship between the OR and serial learning, includ-

ing the von Restorff phenomenon.

The Orienting Reflex

The orienting reflex is usually described as an
organism's first reSponse to any normal change in stim-
ulation. Its apparent function is to bring the organism.
into closer contact with its environment following the
occurrence of a novel stimulus. Although a broad range
of stimuli can elicit the reflex, the same components are
usually observed, and for this reason the OR is considered
to be nonspecific.

Among the several components of the orienting re—

flex, a simultaneous dilation of the blood vessels of the

head and constriction of the blood vessels of the periph-
eral areas is asserted by Sokolov (1963) to be the most
reliable and distinctive component. Secondly. alpha-
blocking occurs. that is the slow waves of high amplitude
in the EEG characteristic of the relaxed state are re—
placed by fast waves of lower amplitude (beta waves)
characteristid of an excited state. Other components

of the OR include the occurrence of the galvanic skin
response (GSR). dilation of the pupil of the eye. and

an increase in the sensitivity of the retina. Certain
somatic components are also observed. including cessation
of general activity. change in respiration. an increase
in muscular tonus. and turning of the receptors toward

the source of stimulation.

Relationships between the Orienting
Reflex and Learning

An early study by Brown (1937) related the GSR
to performance in a serial learning task. Stimuli were
19 nonsense syllables of 0%.associative value. and an 8

second ISI was used. Brown rank-ordered the number of

correct anticipations to each syllable and the GSR to each
syllable (summed over all trials) for each subject. For
individual subjects. correlation between the order of
learning and the rank order of the GSR ranged from +.14

to +.67. The correlation for the measures summed over
subjects was +.80. Brown also averaged the magnitude of
GSR for each trial up to and including the first correct
anticipation and found a correlation with order of learn-
ing of +.8l. A lower correlation (+.58) was found between
the average GSR on correct trials and the order of learn-
ing of each syllable.‘ Brown concluded that the typical
serial learning curve was observed because of "greater
attention. tension. awareness. or apprehension on the part
of the individual to those items occupying a primary and a
final position."

Thompson and Obrist (1964) observed EEG activity
during two control periods. a serial learning task. and
overlearning. Measurements of wave activity were made in
the 4 seconds immediately preceding the presentation of
each syllable. In general terms. they found that alpha
activity decreased and beta activity increased during

learning. During overlearning. both measures recovered

somewhat toward control levels. In order to examine the
hypothesis that the maximal EEG changes took place near
the point of learning of each syllable. the syllables
were divided into 4 groups according to the quarter of
learning within which the last error occurred. The mean
number of alpha and beta waves for each type of syllable
in each quarter was tabulated; the least number of alpha
waves and the greatest number of beta waves were found
during the quarter in which learning occurred for all of
the four syllable groups.

Using the GSR. Kintsch (1964) has obtained data
similar to the Thompson and Obrist (1964) data with the
EEG. In the first experiment. the numbers 1 and 2_were
associated with 12 nonsense syllables. a type of learning
which has been previously shown to be all—or-none in char-
acter. When GSR data were plotted forward and backward
from the last error. it was shown that the GSR increased
to the point of the last error. then decreased. In Exper-
iment II. gs learned a 10-item list of different numbers
paired with nonsense syllables. This type of learning
has been described by a two-stage model in which the first

stage is an initial string of errors. and the second is

an intermediate stage which follows the first correct antic-
ipation and in which response probability is expected to be
approximately constant. Following the intermediate stage.
there is a transition to the learned state. When the GSR
was plotted for the items in each of these states. a non-
significant increase was observed in the first stage. then
a steady state in the intermediate stage. and finally a de-
crease in the GSR as in Experiment I. In both experiments
there were no differences in GSR when correct trials were
compared with incorrect trials prior to the completion of
learning.

Maltzman and his associates have worked with the
hypothesis that the OR is related to discriminative ability
rather than to such gross excitatory mechanisms as drive.
Belloni (1964) based a paired-associate experiment upon an
experiment by Standish and Champion (1960). which supported
the drive position by contrasting the S's rate of learning
of easy and difficult lists. Belloni found no correlation
between anxiety and the OR (as measured by the response to
a word in a pretest). and Standish and Champion's results
with respect to anxiety were not replicated. With respect

to the OR. High OR males were superior to the Low OR males

in both the easy and the difficult lists. supporting the
position that the OR is related to discriminative ability.

Another more direct relationship between the OR
and learning is that the OR itself can be conditioned.
Furedy (1967) obtained a conditioned GSR using a tone as
the conditioned stimulus and a cool air puff in a warm
room as the unconditioned stimulus. Furedy employed a
control group which received the airpuff approximately
10 seconds before the tone. and found a significantly
greater amount of GSR to the tone in the forward condi-
tioning group.

Raskin (1969) reported a semantic conditioning ex-
periment in which the CS word "plant" was followed by a
110 db white noise. with filler words between the CS and
UCS. He reported that evidence of conditioning was indi-
cated by a reliably greater GSR to the CS word than to
the preceding control words. The difference in OR level
was highly significant. as was the evidence for semantic
conditioning. Some gs were informed that the white noise
might follow some specific word. The High and Low OR

gs who were informed conditioned faster than their

uninformed counterparts. Informed High Anxiety SS. on

the other hand. conditioned much faster than uninformed
high anxiety gs. while low anxiety gs conditioned at about
the same rate regardless of their information about the
experimental contingencies. Thus. the independence of

the OR from anxiety was supported in this context also.

The von Restorff Effect

Studies such as Kimble and Dufort (1955). Jensen
(1962). and Newman and Saltz (1958) have isolated an item
in a serial list. e.g. by changing its color or meaning-
fulness level. Typically it has been found that the
learning of the isolated item as a response is enhanced.
but that learning of the item immediately following the
isolated item is not improved. I

Raskin et a1. (1967) noted that the above studies
used relatively innocuous methods of item isolation. Using
mild shock to isolate the seventh item of a 12-item serial
list. Raskin et al. found significantly faster learning of
the 6th. 7th. and 8th items. as well as significantly

faster learning of the entire list in the shock-isolated

condition. It was reasoned that the facilitation of

learning was related to the elicitation of large ORs by
the shock isolated item. They suggested that other in-
vestigations have failed to find overall facilitation
because other methods of isolation did not produce strong.
sustained ORs to the items in the middle of the list.

The use of electric shock was said to have overcome this

difficulty.

Statement of the Problem

Since experimental research has established a re-
lationship between components of the OR and verbal learn-
ing. and since the isolation of a stimulus to obtain the
von Restorff effect is likely to produce a large OR to
that item. it is possible that the Von Restorff effect may
be explainable in terms of the effects of the OR. as sug-
gested by Raskin et a1. (1967). In the present experiment.
cool air was used in an attempt to isolate items. since it
was an appetitive stimulus instead of aversive. and had
been shown by Furedy (1967) to be capable of eliciting a

conditioned GSR.

10

Three experimental conditions were established for
the current research. each condition being run in a warm
room with a serial-learning task using nonsense syllables.
In the puff-isolate condition (PI). the seventh item of
the list was isolated by being paired with a cool air
puff in the warm room; in the puff control condition (PC)
the air puff was presented prior to the first item of the
list. Condition C (control) received no air puff and was
simply the typical serial learning paradigm. The GSR as
a measure of the OR was recorded in all conditions.

The following predictions were advanced: l) The
learning of the isolated item in Condition PI should be
enhanced. since the air puff should elicit an OR. The
critical items with respect to this prediction were items
7 and 8. Isolation was expected to affect item 7 in its
function as a response; and was expected to affect item 8
insofar as it changed the stimulus pr0perties of item 7.
2) The learning of the first item in Condition PC should
not be enhanced. since the first item of a serial list is
already isolated by its primary position. and the position
alone should yield a large OR. This condition was expected

to produce results similar to Condition C. 3) In all

11

conditions. the GSR should be negatively correlated with
the number of trials to the last error for each item;
that is large ORs should be associated with faster

learning.

METHOD

Subjects.--Forty-seven male §s from the introduc—

 

tory psychology course at Michigan State University volun-
teered for the experiment. Each §_received extra course
credit for his participation. Two gs were rejected from
the subject pool. One had previously participated in a
similar experiment. the second failed to meet a criterion
of six correct responses by the twenty-fifth trial. Fif-
teen of the gs were assigned to each of the three experi-
mental conditions. Assignment of §s to the experimental
conditions was made by predetermined random order. based
upon the order of the Ss' appearance at the laboratory.

Apparatus.--The learning material was presented

 

by an MTA Scholar 100. The GSR was recorded with a Mari-
etta Apparatus 12-l3R recorder. The first and third fin—
gers of the §fs left hand were moistened with electrode
paste and an electrode was placed on each of the two fin-
gers. The teaching machine controlled an event marker

which recorded the time of presentation of each stimulus

12

13

on the recording of the GSR. The teaching machine also
signalled a Hunter timer at the proper time for presenta-
tion of the air puff.

The air puff was administered to the S by a sys-
tem involving compressed air. the timer. and a solenoid
switch valve. The timer completed an electrical circuit.
opening the solenoid switch valve. Air was then released
from the compressed air cylinder for 1 second at a regu-
lated pressure of 25 psi. and was cooled by being passed
through 25 feet of 1/4 inch copper tubing immersed in ice
water. The final section of tubing was plastic of 3/8
inch diameter. attached to the teaching machine such that
it would emerge on the forehead of the S, from a distance
of 2 feet from the g.

Learning Materials.——Twelve CVC nonsense syllables
of 42% associative value according to the Krueger (1934)
norms were selected for this experiment. In order to con-
trol for any specific effects of any stimulus or of stim-
ulus order. three different randomized orders of the list
were constructed; 5 S5 of the 15 in each group received
one of the three different lists. Each of the lists be-

gan with four asterisks as a signal to anticipate the

14

first item. The three versions of the list are shown in
Appendix A. The lists were constructed such that succes-
sive nonsense syllables had no letters in common. No syl-
lable was used in the same position in more than one list.
and each syllable was used in both the first half of one
list and the second half of another. Also no identical
strings of successive syllables longer than two syllables
were repeated between lists.

Procedure.--The gs were seated in front of the

 

teaching machine in a room maintained at 95°F. The elec-
trodes for measurement of the GSR were attached. and gs
were read Part I of the instructions in Appendix B. gs
learned a pretest serial list consisting of four common
words in order to familiarize them with the serial learn—
ing task and to allow the GSR to become habituated to the
general experimental situation. Part II of the instruc-
tions in Appendix B were then read to the S. the GSR re-
corder was adjusted. and the serial list was presented to
the S.

The syllables were presented at a 6—second rate.
a time judged to be long enough for the occurrence and

recovery of the GSR to each stimulus. The inter-trial

l5

interval was 18 seconds. consisting of two 6-second periods
of nothing in the window followed by 6 seconds of a row of
4 asterisks.

In conditions PI and PC. an air puff was admini-
strated within the list. The gs in Condition PI were ad-
ministered an air puff at the onset of the seventh item
of the serial list. i.e. during anticipation of the eighth
item. The gs in Condition PC received the puff during
presentation of the asterisks. i.e. during anticipation
of the first item. Condition C involved no air puff and

was simply the standard serial-learning paradigm.

RESULTS

In order to assess the speed of learning. the
mean trial to last error was calculated for each syllable.
Figure 1 presents these results for the three conditions.
In general. Condition PI produced faster learning with re—
spect to trials to the last error. According to the analy-
sis of variance in Table 1. only the effect of serial po-

sition was significant F(ll.462) = 10.86, p < .001.

TABLE 1

Analysis of Variance of Mean Trial to Last Error for
All Serial Positions

 

 

 

m n a
Source of Variation SS DF MS F P
Between Subjects 8415.90 44
Groups 393.41 2 196.71 1.03
gs within groups 8022.49 42 191.01
‘Within Subjects 7881.83 495
Serial position 1546.89 11 140.63 10.86 .001

Groups X serial
position 351.03 22 15.96 1.23 .2

Serial position X
position within
groups 5983.91 462 12.95

16

l7

 

 

12 a)
s 0.
_- /\\
. // . \
. A . \
10 -- \ /__ —/
/ \ \
/ / O O \
4- / '
O
’—
B—r / o
/
u . I.
o
H I
H
0
U 6“ I o
O
,9. I
"a I
.-o "“ I — Group PI
3.
.‘3 I 0—‘0 Group PC
a a I
a . Group C
, I
3 -.—
l i, 1 Ir 1 ‘1: § % fil 1r f
113 £5167 89 10 1112
Serial Position
FIGURE 1

loan trials to last error as a function of serial position

18

A finer examination of the learning data in terms
of trials to last error was made on individual serial po-
sitions. Table 2 shows that learning of the first item
in Condition PC was slower than the learning of that item
in the other (PI and C) conditions but the differences
were not significant. F(2.42) = 2.42, p < .10. Tables 3
and 4 show that. with respect to the number of trials to
the last error. there were no significant differences be—
tween the PI. PC. and C conditions at either item 7 or

item 8.

TABLE 2

Analysis of Variance of Mean Trials to Last Error.
Serial Position 1

 

 

Source of Variation SS DF MS F P
Total 989.11 44
Groups 101.91 2 50.96 2.41 .10

Error 887.20 42 21.12

 

19

TABLE 3

Analysis of Variance of Mean Trials to Last Error.
Serial Position 7

 

 

Source of Variation SS DF MS F P
Total 1231.24 44
Groups 16.84 2 8.42 .29
Error 1214.40 42 28.91
TABLE 4

Analysis of Variance for Mean Trial to Last Error.
Serial Position 8

 

 

 

Source of Variation SS DF MS F P
Total 1040.80 44
Groups 18.53 2 9.27 .38
Error 1022.27 42 24.34

 

It was predicted that the GSR within each of the
three experimental conditions would be inversely related
to trials to last error within each of the three condi-

tions. Figure 2 shows the mean GSR. averaged over all

20

8 “i“

7—(r— "'—
o-——--O

5 __

 

Mean per cent change in skin resistance

 

l L 1 J 4-

] 1 I l 1 j
0 1 2 3 4 5 6 7 8 9 10 1 l 12
Serial Position

FIGURE 2
Mean percent change in skin resistance as a function of serial position

21

trials. for the anticipation interval preceding each item
in each condition. GSR was measured in terms of per cent
change of conductance during the anticipation of each
item. In Condition C. the GSR was inversely related to
trials to last error. the GSR being highest at the primary
items of the list. dropping in the middle and rising
slightly at the final position. The reliability of this
relationship was supported by a significant correlation.
r(ll) = -.73. p < .01. between GSR and trials to last
error for Condition C. Figure 2 also shows that the GSR
data for group PC were similar to the GSR data for Condi-
tion C. with the exception of item 1. Item 1 in the PC
group received the air puff during the anticipation period.
and correspondingly had an elevated GSR. The learning data.
as previously mentioned. showed a nonsignificant increase
in number of trials to last error for this group with re-
spect to item 1. Correspondingly the correlation between
GSR and trials to last error for the PC group was not sig-
nificant. r(ll) = .38. p = .22. The results for GSR in
Condition PI evidenced a basically inverse serial learning
curve with the exception of item 8. the item during the

anticipation of which the air puff was administered. The

'22

learning data at this point did not show a decrease in
trials to last error. and the correlation between trials
to last error and GSR within this condition was not sig-
nificant. r(ll) = -.43. p = .16. However. the correla-
tion between the same two variables for all items exclud-
ing the eighth item was -.72 df - 10 (p - .01).

An analysis of variance of the per cent change
of skin resistance showed that there was not a statistic-
ally significant difference between the three groups on
this measure. F(42.2) = .12.

The base level of skin resistance was also com-
puted. However. plots of those data showed that the base~
level remained nearly constant over the various serial
positions. yielding no reliable relationship between base

level and either the GSR or serial-learning data.

D ISC USS ION

It was expected on the basis of previous research
that three main predictions would be upheld by this exper-
iment. The prediction that the learning of the isolated
item in Condition PI should be enhanced was not supported
by the data. Item 7. where the isolated item functioned
as a response. and Item 8. where the isolation of the 7th
item affected its stimulus properties were the critical
items in this prediction. One reason for the failure of
this prediction with respect to item 8 may be related to
the next prediction. The prediction that the learning of
the isolated item in Condition PC should not be enhanced
was supported by the finding that the learning of item 1
in the PC group was inferior to the learning of the item
in the other two groups. although the difference was not
significant. The occurrence of the air puff. therefore.
may have interfered with the emission of the response be-
ing anticipated. Interestingly. this result was found
even though all gs in the PI and PC groups reported during

23

24

post-experimental questioning that the airpuff did not
interfere with their response to those items.

The final prediction that the GSR should be nega-
tively correlated with the measure of trials to last
error was partially supported by the data. Group C. the
group most similar to that used by Brown (1937). produced
the correlation of -.73. Brown's correlation of -.80 was
obtained by correlating rank order of amount of GSR with
the rank order of learning. Thus. this replication has
been obtained using measures with stronger quantitative
assumptions. The correlation of —.72 for the eleven non-
isolated items in group PI supported this prediction also.

A possible factor in the failure to obtain a VODI
Restorff effect or a spread of the isolation effect may
have been the long presentation rate of six seconds. Al-
though similar items were used in the experiment by Ras-
kin et a1. (1967). the mean trial to last error ranged
between about 3 and 12 in this experiment as compared
with a range of about 10 to 30 in the Raskin,et al. ex—
periment. This difference in speed of learning may have
been a result of increased rehearsal time allowed by the

6-second presentation rate in contrast with the Raskin et

25

al. rate of 2 seconds. Thus. a limitation may have been
placed upon the amount of improvement in learning which
could have been produced by the air puff. This factor
is especially relevant in view of the Jenkins and Postman
(1948) hypothesis that isolation attracts attention and
attention leads to increased rehearsal time.

Another factor in the failure of the prediction
of a von Restorff effect appears to be that the air puff
interfered with emission of correct responses. The exper-
iment. if run to adequately test the prediction. should
use a puff strong enough to elicit an OR. but not strong
enough to interfere with responding. as was apparently the
case in this experiment. With respect to the spread of
isolation effect. an additional possibility is that the
air puff. in contrast to electric shock. may not have
been salient enough to produce heightened ORs to other

items in the list.

REFERENCES

Belloni. Marigold L. The relationship of the orienting
reaction and manifest anxiety to paired associate
learning. Unpublished doctoral dissertation:
University of California. Los Angeles. 1964.

Brown. C. H. The relation of magnitude of galvanic skin
responses and resistance levels to the rate of
learning. J. exp. Psych.. 1937. 39, 262-278.

Furedy. J. J. Classical appetitive conditioning of the
GSR with cool air as UCS. J. exp, Psychg. 1967.
15..73-80.

Gibson. Eleanor J. A systematic application of the con-
cepts of generalization and differentiation to
verbal learning. Psyphol. Review.. 1940. 41,
196-229.

Gibson. Eleanor J. Intra-list generalization as a factor
in verbal learning. J. exp. P§ych.. 1942. 39,
185-200.

Jenkins. W. O. and Postman. L. Isolation and spread of
effect in serial learning. Amer. J. Psych..
1948. pg, 214-221.

Kimble. G. A. and Dufort. R. H. Meaningfulness and iso-
lation as factors in verbal learning. J. ex .
PS!Ch.I 1955: 29: 361-368.

Kintsch. W. Habituation of the GSR component of the ori-
enting reflex during paired-associate learning
before and after learning has taken place. g,
Math. Psychol.. 1965. 2, 330-341.

Koffka. K. Principles of Gestalt Psychology. New York:
Harcourt. Brace. 1955.

26

27

Maltzman. I. and Raskin. D. C. Effects of individual dif-
ferences in the orienting reflex on conditioning

and complex processes. J. exp. res. Personality.
1965. l, l-l6.

 

Newmann. S. E. and Saltz. E. Isolation effects: stimulus
and response generalization as explanatory con-
cepts. J. exp. Psych.. 1958. 55, 467-472.

 

Raskin. D. C. Semantic conditioning and generalization of
automatic responses. J. exp. Psych.. 1969. 12.
69-76.

 

Raskin. D. C.. Hattle. M. S.. and Ruble. E. W. The effects
of electric shock isolation in serial learning.
Psych. on Sci.. 1967. 8, 413-414.

 

Saltz. E. and Newman. S. E. The von Restorff isolation
effect: test of the intralist association assump-
tion. J. exp. Psych.. 1959. 58, 445-451.

 

Sokolov. E. N. Perception and the conditioned reflex.
Oxford. Pergamon. 1963.

Standish. R. R. and Champion. R. A. Task difficulty and
drive in verbal learning. J. exp. Psych.. 1960.
52. 361-365.

Thompson. L. W. and Christ. W. D. EEG correlates of verbal
learning and overlearning. Electroenceph. Clin.
Neurophysiol.. 1964. lg. 332-334.

APPENDIX A

CONSTRUCTED LISTS

List 1 List 2 List 3
XEC VUB WUC
KYV YAD PAF
ZOB TEV BYP
YAD BYP XEC
VUB GOX ZOB
PAF QIF JUF
QIY JUF TEV
TEV ZOB GOX
JUF XEC KYV
GOX IYV YAD
WUC PAF VUB
BYP WUC QIY

28

APPENDIX B

INSTRUCTIONS

PART I

This is an experiment in verbal learning. The
electrodes which I have just attached will measure elec-
trical responses in the skin. Please rest your hand
quietly on the table in front of you; any movement of
the fingers will be picked up by the recorder and will
interfere with the measurement of the electrical responses.

The first list that you are to learn is a practice
list of four items. The items will appear in the window
of the machine in front of you. The first time through
the list. simply watch to see what the items are and their
order. Starting with the second trial. when the asterisks
appear try to anticipate the first word. when the first
word appears try to say the second. and so on. In general.
then. your task is to try to anticipate the next item on

the list each time an item appears.

29

30

Appendix B (cont.)

PART I I

This second list is a list of twelve nonsense syl-
lables which is to be learned similarly to the first list.
Again. when each preceding item appears you are to try to
anticipate the next items; in this case. however. please
spell out the syllables. Occasionally. there may be puffs
of air from the tube in front of you.

The experiment will end when you correctly antici-
pate all items of the list on two consecutive trials. Are

there any questions?

1042

I
l
l
l
I
l
i
l
l
l
I.
I
II
I
l
l.
l
l
Ill-l
l
l
I
l
II
II
l
I
[III
'-
nlll'
Ill
1"-
l
'-
I
I
III
I
II
l
I
III
III
I.
I
III

3 1293 03