i .r. it.

m.“ nu a? x...“

2»; ”U 3 ”A...” “m...

a! . :5 II

MM“. AKIN am... I u «Vb
I338. u-.. M015 :5!

.' . HF... .
an... 2 :5 «0.0“ «Me. U T m Q I M

I

q a .2. 33:... {Q ..
”WM... pt I)...“ “My”
CL an. n .. Mr.” «If
Hm” EL. RV an“... am.
9. U? c... Q Ma.

H. igu‘u 09%.”! J‘nd
«Mu ”nu 3 .a... Pi

 

u, alr- .«_ ‘..- am- ’9‘”wa
3-“ “
. IH'

l
V...” 3 , a, w. _.
E r ~_.- ‘
‘fi.’ 1“} A" n. k -i
I: 2 . n l ‘ -'
Pmc £213.21 $€ "

U
; ~ L 5. a» _'
-»;;,_:3, L314 fins: 3:

1‘

  

we

”iv-erase”; .» ‘~ r7

 
  
    
      

  

‘ animus av ‘
HOME 8: SDNS’
300K BINDERY iNC.

LIBRARY BINDERS
9mm: “mum-u

 

 

ABSTRACT
DEVELOPMENTAL INFLUENCES ON DIVERGENT THOUGHT
BY

Leonard Paul Sawisch

Results of previous studies manipulating stimulus
presentation suggested that on a divergent thinking task'
(Unusual Uses for a Penny) subjects at Piaget's concrete
operational level of cognitive development (5th grade)
would show facilitated performance when stimulus presen-
tation was concrete, while subjects at Piaget's formal
operational level of cognitive development (college) would
show facilitated performance when stimulus presentation
was abstract.

Measures of flexibility and originality did not
support the hypothesis, while the measure of subject
fluency did, especially at the 5th grade level. Some
procedural problems with originality indices are discussed,
and a measure of specific response types is suggested as
a more appropriate measure of subject performance for
future studies in this area. In addition, a presentation
gender interaction was found with a percentage of common
responses index, suggesting a gender effect not discussed

before, possibly due to task designs in previous studies.

 

DEVELOPMENTAL INFLUENCES ON DIVERGENT THOUGHT

BY

Leonard Paul Sawisch

A THESIS

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

MASTER OF ARTS

Department of Psychology

1974

 

ACKNOWLEDGMENTS

I am pleased to express my appreciation to
Dr. Donald M. Johnson, chairman of my committee, for
his guidance and assistance in the preparation of this

thesis.

ii

TABLE OF CONTENTS

Page
INTRODUCTION . . . . . . . . . . . . . . . . . . . 1
METHOD 0 O 0 O O O O O O O O O O 0 O O O O O O O O 5

Subjects . . . . . . . . . . . . . . . . . . 5
Procedure . . . . . . . . . . . . . . . . . . 5

. RESULTS 0 O O O O O O O O O O O O O O O O O O O O 10
Fluency . . . . . . . . . . . . . . . . . . . 10
Flexibility . . . . . . . . . . . . . . . . . 12
Originality . . . . . . . . . . . . . . . . . 12

DISCUSSION . . . . . . . . . . . . . . . . . . . . 18

SUMMARY . . . . . . . . . . . . . . . . . . . . . 24

REFERENCES . . . . . . . . . . . . . . . . . . . . 27

APPENDICES . . . . . . . . . . . . . . . . . . . . 28

Appendix I - Flexibility Categories . . . . . 28
Appendix II- Rating Scale for Originality . . 30

iii

LIST OF TABLES

Table Page
1 Correlations and corresponding F values
for the fluency index . . . . . . . . . . . ll
2 Correlations and corresponding F values

for the flexibility index . . . . . . . . . l4

 

3 Correlations and corresponding F values
for the mean originality index, the
percentage of unusual responses index,
and the percentage of common responses
index . . . . . . . . . . . . . . . . . . . 15

4 Examples of unusual responses given by
5th grade and college subjects . . . . . . 21

iv

Figure

LIST OF FIGURES

Page
Presentation-age interaction for
fluency O O O O O O O O O O I O O O O O O O 13
Presentation-gender interaction for j
percentage of common responses . . . . . . l7

INTRODUCTION

Underlying Piaget's description of cognitive develop-
ment is the increase in ability to think logically and
eventually the increase in ability to think logically
about abstract concepts. For example, while subjects at
the concrete operational level of cognitive development
have the ability to reverse operations, a necessary step
in logical thought that promotes solutions to simple
conservation tasks, this ability is somewhat limited to
concrete applications. It is not until subjects reach
the formal operational level of cognitive development that
the immediate situation can be transcended and thought can
be directed to more abstract applications (Piaget and
Inhelder, 1969).

The value of any theory, of course, lies in its
predictive ability. While knowing a subject's level of
cognitive development provides a degree of predictability
of subject performance on typical Piagetian tasks that
involve a specific solution, or limited set of solutions
(e.g., conservation and classification tasks), to what
degree does this same information facilitate prediction
of a subject's performance on tasks that do not have a

specific solution, or set of solutions?

 

Though this question has not been directly investi-
gated in the past, there are a few studies that in
conjunction suggest an interesting relationship between
cognitive level and ability to deal with abstractions.
Goodnow (1969), using Unusual Uses tasks with kindergarten
subjects, found that they made more original (non-
standard) responses when allowed to manipulate the
stimulus object, than when simply allowed to View it.

Otte (1962), in studying the effects of verbal (word) and
pictorial presentation of stimuli on concept attainment,
asked subjects to explain how groups of three stimuli

were alike. In a pilot study with 20 undergraduates, Otte
found verbal presentation evoked significantly more
responses (.01 level) than pictorial presentation, with
the nature of the response sets differing significantly
(.05 level) by mode of presentation. When the study was
replicated with 80 4th grade subjects, however, significant
differences in the opposite direction were found! Davidon
(1952), in a similar study with college students, used
four modes of stimulus presentation (short names, long
names, line drawings, and photos), and found that both
short names and long names elicited significantly more
unique responses than did line drawings or photos (the
influence on number of responses was unclear).

The relationship suggested is an interaction between

level of cognitive development and level of abstractness

of stimulus presentation. A plausible Piagetian
interpretation would be that subjects at the formal
operational level of cognitive development perform best
when given a more abstract stimulus presentation, while
subjects at the concrete operational, or lower level of
cognitive development perform best when given a more
concrete stimulus presentation.

While both the Otte and Davidon studies used concept
attainment tasks which do have a finite, though large,
number of solutions, a more appropriate approach would
be to use a clearly divergent task with a very concrete
stimulus presentation (i.e., an actual, manipulable
stimulus object) and an abstract stimulus presentation
(i.e., a symbolic representation of the actual stimulus,
or a verbal label for the stimulus). With subjects
assumed to be at the formal operational level of cognitive
develOpment (e.g., college students), one would assume
that performance would be most facilitated with an
abstract presentation while with subjects assumed to be
at the concrete operational level of cognitive develop-
ment (e.g., 5th grade subjects) one would assume that
performance would be most facilitated with a concrete
presentation. Note that while a cognitive level by
stimulus presentation interaction is predicted, direct
comparison of 5th grade and college performance is not as

important to the question as is comparison of performance

between presentations within each cognitive (or age)
level.

A review of studies in the area of divergent thought,
especially the works of Guilford and Torrance, suggest a
number of possible divergent thinking tasks that allow
for a great variety of solutions (Johnson, 1972). For
this study, however, the Unusual Uses task, as designed
by Guilford and modified by Torrance, has a number of
advantages. First, the task itself is relatively easy,
and so can be used with subjects of varying ages. The
task is also brief enough to maintain the interest of
younger subjects. More importantly, however, the task
provides easily rated measures of three performance
variables: fluency, flexibility, and originality
(Torrance, 1974). The stimulus object used is a penny.
Previously used by this author, this object elicits a
wide range of responses, and normative data is available

for both 5th grade and college subjects.

METHOD

Subjects
Thirty-six 5th grade subjects (18 males & 18 females)

were drawn from two 5th grade classes at a lower middle-
class grade school (Holt school system in Michigan), and
thirty-six college subjects (18 males & 18 females) were
volunteers from introductory psychology classes at

Michigan State University.

Procedure

 

College subjects were randomly assigned to one of the
two stimulus presentation conditions and then run in two
groups (by presentation). The 5th grade subjects were run
by classes, with each class assigned to one of the two
presentation conditions. This division was necessary due
to the difficulties in obtaining grade school subjects.

Each subject received an 8" by 11" paper with direc-
tions followed by twenty numbered spaces, and an envelope
containing the stimulus object. The directions (identical
for all subjects) were then read by the experimenter and

any questions were answered.

Directions

 

"This is an Unusual Uses task. You are to
list below as many unusual uses as you can think
of for the object provided. By unusual we mean
any use that is different from the common,
everyday, or intended use of the object. In
the envelope provided is the object you are to
work with. Open the envelope, take out the
object, and begin work. Remember, list as many
Unusual Uses as you can think of for the object
you have."

For the concrete presentation condition the envelope
contained an actual penny; while for the abstract presen-
tation the envelope contained a 3" by 5" card with the
words 'a penny' typed in the center. (Review of the
response sheets and verbal reports by subjects in the
latter group indicated that no subjects responded to the
actual index card rather than the words 'a penny'.) The
envelope procedure was used to allow for identical direc-
tions with both stimulus presentation conditions, and to
insure that, at least initially, subjects in the concrete
group actually handled the penny. All subjects were
allowed 10 minutes to complete the task.

Each subject's response sheet was rated for fluency,
flexibility, and originality by two independent judges.
Judges were trained on each scale using response sheets
from an earlier study until a high degree of concordance
between judges was obtained.

Fluency is the number of different responses produced.

This is defined rather leniently in that elaboration of an

idea or response may contain a number of uses. For

example, a subject responding with "A penny can be used
in games, like spinning contests or pitching pennies" was
rated as giving two responses: "spinning contest" and
"pitching pennies". Unelaborated, "A penny can be used
in games", the response was rated as a single response.
The fluency index was the sum of the ratings by both
judges.

Flexibility is the number of response set changes;
every time a response falls into a different category
from the response immediately before it, a point was given.
A guideline list of categories, obtained from previous
responses by 5th grade and college subjects to the penny
item, is located in the Appendix. The flexibility index
is the sum of the ratings by both judges, divided by the
fluency index.

Originality of each response was rated on a seven-
point scale, from "very common" (1), to "very unusual"
(7). This type of rating scale has been used extensively
in research at Michigan State University (Johnson and
Kidder, 1972). A sample of the rating guide used is
located in the Appendix. While it may be argued that a
separate originality scale should be used for each grade
level, it was not the intent of this study to make a

cross-age comparison, but to simply look at changes within

 

each grade level due to the experimental manipulation.
The scale described above has given a good distribution of

ratings for both 5th grade and college subjects in the

past, suggesting that it will be sensitive to changes
within each cognitive level. Interestingly, the results
of this study could influence the choice of separate
rating scales in the future, if the researcher's intent
is to make a cross-age comparison.

The originality index for each response is the sum
of the ratings by both judges. Given that subjects tend
to give a number of responses, there are many ways of
using the originality ratings to assign an originality
index for each subject. Traditionally, two approaches
have been taken. The first is a simple count of the
absolute number of original responses (i.e., responses
rated "5" or higher). While this is the index used in
the Goodnow (1969) and Davidon (1952) studies, it is a
somewhat crude measure in that it can be influenced by a
subject's fluency score (Johnson, 1972). The second
approach controls for fluency by dividing the total of
both judges' ratings by the fluency index, resulting in a
mean originality index. This is the index used by
Johnson and Kidder (1972), and the same index was computed
for subjects in this study.

Two additional indices were also computed. Because
the focus of this study was on performance per s9, and not
specifically on original output, for each subject the
percentage of responses rated "5" or above (uncommon
responses) was computed, as was the percentage of

responses rated "3" or below (common responses). To allow

a clear distinction between the two measures, responses
rated "4" were not considered (i.e., a "4" response may

be one 'a bit better than a "3"' or one 'almost a "5"').
For clarity, for each subject the number of "l", "2",

and "3" ratings were totaled, then divided by the total
number of ratings to give a percentage of common
responses. The same procedure was used to determine the
percentage of unusual responses. Note that this procedure
controls for effects due to subject fluency in the same

manner as the mean originality index.

7

RESULTS

A correlational analysis of data was used separately
for each of the five indices, with relation to cognitive
level (5th grade or college), stimulus presentation
(concrete or abstract), and gender, in view of the
popular concern over possible sex differences. This
procedure provides all the information provided by the
classical analysis of variance (F=(N-8) Iggy), and, in
addition, it indicates the direction of effects and allows
easy and direct comparisons of the magnitude of different
effects, both within this study and across studies, as
correlations are independent of sample size.

To facilitate understanding of the direction of
effects, the following values have been assigned to
maximize the number of positive correlations. For stimulus
presentation, concrete is positive (+) and abstract is
negative (-). For assumed cognitive level (age), 5th
grade is positive (+) and college is negative (-). For
gender, females are positive (+) and males are negative

(-).

Fluency

The interjudge reliability for fluency ratings was

.99. Table 1 contains the correlations and corresponding

10

11

TABLE 1

Correlations and corresponding F values
for the fluency index.

 

 

 

Correlation F value

Presentation (P) .26* 5.28*
Age (A) -.10 .76
Gender (G) -.01 .01
PA .28* 6.36*
PG .00 .00
AG -.08 .53
PAG .21 3.42

*P<.05

Concrete Presentation = +

Abstract Presentation

5th Grade = +
College = -
Females = +

Males = -

12

F values for the fluency index. Stimulus presentation

had a significant main effect on fluency (r = .26, P‘<.05),
with a mean fluency index of 30.42 for concrete presenta-
tion and 24.17 for abstract presentation. There was also
a significant presentation-age interaction (r = .28,

P < .05). Figure l clarifies this interaction, indicating
that the major effect was at the 5th grade level, with a
mean fluency index of 32.67 for concrete presentation and
only 19.56 for abstract presentation. At the college
level, the abstract presentation was not significantly
greater than concrete presentation (mean fluency indices

of 28.78 and 28.17, respectively).

Flexibility

 

The interjudge reliability for flexibility ratings
was .99. Table 2 contains the correlations and corre-
sponding F values for the flexibility index. Age had a
significant main effect on flexibility (r = .35, P‘<.01),
with a mean flexibility index of .59 at the 5th grade

level and .74 at the college level.

Originality

 

The interjudge reliability for originality ratings
was .97. Table 3 contains the correlations and corre-
sponding F values for the three originality indices. For
mean originality, gender had a significant main effect
(r = .38, P‘<.01), with a mean originality index of 3.71

for females and 3.13 for males.

 

34
32
30
28

Mean

Number 26

of

Responses 24
22
20
18

16

13

32.67

   

r28.17 ------- * 98-78

 

19dfi5
5th grade ---
College - -
Concrete Abstract

Mode of Presentation

FIGURE 1. Presentation-age interaction for fluency.

14

TABLE 2

Correlations and corresponding F values
for the flexibility index.

 

 

Correlation F value
P .01 .01
A -.35** 9.60**
G -.09 .68
PA .04 .11
PG .09 .65
AG -.06 .31
PAG -.14 1.59

 

**P < .01

15

TABLE 3

Correlations and corresponding F values for the mean
originality index, the percentage of unusual responses
index, and the percentage of common responses index.

 

 

 

 

Mean Original Percentage Unusual Percentage Common
r F value r P value r F value
P -.01 .01 -.Ol .02 .06 .33
A .10 .87 .40** 16.53** -.20 3.35
G .38** 12.21** .41** l7.ll** -.33** 9.51**
PA -.10 .88 -.12 1.52 .19 3.04
PG .18 2.72 .12 1.59 -.25* 5.28*
AG .15 1.84 .14 2.05 -.19 1.42
PAG -.03 .08 .05 .30 .05 .19
*P<.05

**P'<.Ol

16

For percentage of unusual responses, both age and
gender had significant main effects, (r = .40, P‘<.01; and
r = .41, P‘<.01, respectively). The 5th grade subjects
had a mean percentage of unusual responses of 36.08,
while college subjects had a mean percentage of unusual
responses of 20.22. Also, female subjects had a mean
percentage of unusual responses of 36.22, while male
subjects had a mean percentage of 20.08.

For percentage of common responses, however, there

was no significant main effect due to age, only a

 

significant main effect due to gender (r = -.33, P <.01),
with a mean percentage of common responses of 59.06 for
males and 42.95 for females. There was also a signifi-
cant presentation-gender interaction (r = -.25, P‘<.01).
Figure 2 clarifies this interaction. Females have a mean
percentage of common responses of 38.45 with concrete
presentation and 47.45 with abstract presentation, while
males have a mean percentage of 66.56 with concrete

presentation and 51.56 with abstract presentation.

 

Mean
Percentage
of

Common
Responses

FIGURE 2.

17

66 _. 66.56 (males)

62 -

58 -

54 -
52 " 51.56

50 -

48 ‘ L7.h§
46 -

44 -

4O -

38 " 38.h5 (females)
36 -

 

Concrete Abstract

Mode of Presentation

Presentation-gender interaction for percentage
of common responses.

 

DISCUSSION

With a Piagetian overview of results from three
published studies, it was predicted that subjects at the
concrete operational level of cognitive development would
show superior performance on a divergent thinking task
when the stimulus was presented in a concrete manner, as
opposed to an abstract manner. Similarly, it was assumed
that subjects at the formal operational level of cogni-
tive development would show superior performance when the
stimulus was presented in an abstract manner, as opposed
to a concrete manner. Performance was measured in three
domains; fluency, flexibility, and originality, with
three indices of originality being used.

The fluency data, to a degree, supports the hypothesis
in that 5th grade subjects were more fluent with a con-
crete presentation than with an abstract presentation.
College subjects, however, had nearly identical
performance under both conditions (Review Figure 1). It
is this lack of difference at the college level with the
large difference at the 5th grade level that explains the
appearance of a main effect by presentation for fluency.

The flexibility data does not support the original

hypothesis, as no differences between stimulus

18

 

19

presentations were found within either age group. There
was, however, a significant effect due to age, with
college subjects more flexible than 5th grade subjects.
This finding is consistent with the findings of other
divergent thinking tasks as reported by Torrance (1974),
where stimulus presentation is abstract for both age
levels.

The three originality indices also fail to support
the original hypothesis in that there were no differences
due to presentation within each age level. The use of
three indices of originality, however, does have some
advantages. While all three indices indicate a gender
effect (i.e., females have a higher mean originality, a
higher percentage of unusual responses, and a lower
percentage of common responses than males), consistent
with data reported by Goodnow (Otte and Davidon did not
consider gender), the percentage unusual and the
percentage common indices seem to provide more information
concerning performance per gg than does the mean
originality index.

Specifically, percentage of unusual responses has a
significant correlation with age, in that 5th grade
subjects appear to give a higher percentage of unusual
responses than do college subjects. The strength of this
correlation, however, may be a function of the originality
rating scale used. Because the focus of this study was

on differences due to stimulus presentation within groups,

 

 

20

the identical originality scale was used for all subjects.
Other researchers, for example Torrance, when making
direct cross-age comparisons, use a weighted score

(based on large normative samples) for each age group
investigated.

The percentage of common responses does not show a
significant correlation with age, which also tends to
support the interpretation that the age effect discussed
above may indeed be a function of the rating scale.
Considering what a common response is (refer to the
originality scale in the Appendix), it does not seem
likely that the kinds of common responses given would
be different for 5th grade and college subjects, while
the variance at the other end of the scale would make it
possible to favor one age group over the other. While
not enough information is available to look specifically
at differences in response types between age groups
(nor was it the intent of the study to do this), Table 4
contains a sample of the most original responses given
by 5th grade and college subjects, and suggest that such
differences may exist.

In addition, with the percentage of common responses
index, a significant presentation-gender interaction was
found. This interaction is of special concern because
no age effect or interaction was found, suggesting that
this effect was not a function of the scale, as it was

found at both the 5th grade and college level. Given the

21

TABLE 4

Examples of unusual responses given by 5th grade
and college subjects.

 

5th Grade Concrete

 

1)

2)
3)

4)

5)

Put holes in it and hook
it onto the faucet.

Talk to that man on it
Go inside the building

Put some together and use
them for a ruler.

Contacts

5th Grade Abstract

 

l)

2)
3)
4)

5)

Put hands on it and make
a clock.

Make a stethoscope
As a record (stereo)
As a movie screen

For a map

College Concrete

 

1)
2)
3)
4)

5)

Under a skipping windshield wiper
Flat marble

To stop on

Dildo

Talk to it

College Abstract

 

1)

2)

3)

4)

5)

Rattle around in man's pocket
(feels nice)

To rhyme with Lenny, Penny,
Benny, etc.

Make a film of it.
Eat for copper plated hemorrhoids.

Use columns on back to build own
home. (very sturdy)

22

direction of this interaction, it lends support to the
significance of the difference between the concrete
presentation and the abstract presentation, although
this difference did not have the effect on performance
originally predicted.

An explanation of this interaction is not an easy
task. Possibly females, when given a concrete stimulus
presentation, find it easier to cut through the response
hierarchy and thus give less common responses than when
stimulus presentation is abstract. The converse for
males would indicate it is easier to cut through the
response hierarchy when the stimulus presentation is
abstract, while a concrete stimulus presentation tends to
somewhat limit the male subjects' ability to cut through
that same response hierarchy.

Why this would be true is difficult to say. As
apparently no other study with divergent tasks has
manipulated the abstractness of stimulus presentation
and considered gender as a factor, the literature is of
no help. Additional research to substantiate this inter-
action and investigate possible causes is needed.*
Considering that most divergent tasks are presented
verbally (abstractly, as defined by this study),

especially for older subjects (Torrance, 1974), a review

 

*Such a study is presently under way at Michigan
State University by Sawisch and Kidder, but the data is
not available at this time.

 

23

of the mean percentage of common responses given with
abstract stimulus presentation (i.e., 51.56 for males and
47.45 for females, as shown in Figure 2) suggest why this
interaction has not been suggested before.

Hopefully this study serves to emphasize the
importance of choosing an originality index that provides
the specific information the experimenter is looking for.
Given this concern, it may well be that the original
hypothesis was based on a somewhat spurious index in that
both the Goodnow and the Davidon studies used absolute
number original as their originality index, which can be
affected by fluency scores. Note that the fluency scores
in this study did support the predicted presentation
effects, especially at the 5th grade level. If the
fluency scores had not been separated from the originality
scores, then, the originality scores might have appeared

to support the predicted presentation effects.

 

S UMMARY

The data in this study does not tend to support the
original hypothesis that concrete operational subjects'
performance would be facilitated more by concrete
stimulus presentation than by abstract stimulus presen-
tation; with the converse being true for formal

operational subjects. There is a possibility, however,

 

that other dimensions of performance not measured here
might be affected by differential stimulus presentation
as a function of cognitive development as described by
Piaget.

The performance variables measured in this study
were suggested by the studies done by Goodnow (1969),
Otte (1962), and Davidon (1952). Both the Otte and the
Davidon studies, however, used concept attainment tasks,
which have a finite solution set. In addition, both the
Goodnow and the Davidon studies used an index of
originality that did not control for subject fluency
(the Otte study did not use an originality index), while
the data in this study indicates subject fluency is
affected by stimulus presentation and level of cognitive

develOpment.

24

25

Perhaps a more apprOpriate approach would have been
to look for differences in specific response types.
Review of the flexibility index, located in the Appendix,
indicates that while it is sensitive to changes in
response sets, it is not refined enough to indicate
specific response types. (If such refinements were made,
it might then demand the use of different scales for each
age level.) Note that while the Otte study used a con-
cept attainment task, a significant age-presentation
interaction was reported for response types.

The major implication of this study is more procedural
than theoretical. Researchers in the future considering
the use of an originality index should be sensitive to
the possible influences due to subject fluency, and in
addition, should consider the use of other originality
indices in addition to mean originality if subject per-
formance 233 fig is the measure they are really concerned
with. It might also be noted that response type, though
difficult to quantify, might be an important addition to
future studies investigating performance on divergent
thought tasks, especially if the study takes a
developmental perspective.

This study should also serve to sensitize future
researchers to the possible influences of stimulus
presentation. This is especially true in view of the

stimulus presentation-subject gender interaction found

26

with the percentage of common responses index, a finding

that obviously demands further investigation.

 

 

 

REFERENCES

 

REFERENCES

Davidon, S. The Effects of Symbols, Shift, and
Manipulation Upon the Number of Concepts Attained,
Journal of Experimental Psychology, 1952, 44, 70-79.

 

Goodnow, J. J. Effects of Active Handling, Illustrated
by Uses for Objects, Child Development, 1969, 49,
201-212.

 

Johnson, D. M. Systematic Introduction to the Psychology
of Thinking, Harper & Row, Publishers, New York,
1972, Chapter 7.

Johnson, D. M. and Kidder, R. C. Productive Thinking in
the College Classroom, E. D. P. Project Report,
January, 1972.

Otte, W. The Differential Effects of\Mrbal and Pictorial
Representations of Stimuli Upon Responses Evoked,
Journal of Verbal Learning and Verbal Behavior,
1962, 1, 192-196.

 

Piaget, J. and Inhelder, B. The Psychology of the Child,
Basic Books, Inc., New York, 1969.

 

Torrance, P. E. Torrance Tests of Creative Thinking,
(Norms-Technical Manual), Ginn and Company,
Lexington, Massachusetts, 1974.

 

27

 

 

 

APPENDICES

APPENDIX I

FLEXIBILITY CATEGORIES

 

12.

13.

14.
15.

16.

APPENDIX I

Flexibility Categoriesl

BUY: buy marbles, buy presents, etc.

MONEY: monetary uses other than to buy; tip, tax,
loan it, etc. [

SAVE: put in bank, collect, trade, etc.

CHARITY or GIFT: give to charity, give away, etc.

 

GAMBLE: bet, etc.

GAMES: spin, pitch, do magic tricks, use as a playing '
piece (checker), etc.
TRACE: circle template, trace Abe, etc.

ART: make pictures, collages, etc.
DECORATION: penny loafers, boarders, etc.
JEWELRY: bracelet, ring, watch face, etc.

GOOD LUCK: penny loafers FOR good luck, wishing wells,
etc.

WEIGHT: sinker, door stop, record player arm weight,
etc.

DESTRUCTION: smash (railroad tracks), melt, threw
away, etc.

WEAPON: slingshot ammo, weapon, target, etc.
ELECTRICAL: fuse, conductor, etc.

TOOL: screwdriver, window prop, scraper, shim (jam
machines, doors), etc.

 

1Based on data from 5th grade and College samples,

and modelled after the Flexibility scale described by
Torrance (1974).

28

29

Appendix I (Cont'd)

l7. COVER or PLUG: eye cover, fill holes, ear plug, etc.

18. SYMBOLISM: insult, pacifier, or use penny as a pun
(play on words), etc.

19. OBJECT: used simply as an object of action; bury,
swallow, bite, throw, freeze, etc.

New categories can be created, but only if the response
cannot be placed in one of the above categories.

APPENDIX II

RATING SCALE FOR ORIGINALITY

7.

APPENDIX II

Rating Scale for Originality

Common, or intended use of the object; using a penny
as money.

buy something, even out change, loan it, etc.

a. Also a common use, though not the original inten-
tion of the object's production.
fuse, collect, flip, draw circles, etc.

b. This rating is also applied when almost any object
can be used in the same way.
bury it, kick it, burn it, freeze it, etc.

The object is often used, or substituted for, (also
more specific than 2b).
checker replacement, screwdriver, book mark,
games, etc.

Apparent use, though less common than a 3.
record player arm weight, art projects, jewelry,
decoration, etc.

a. Not a common use, perhaps demanding additions or
alterations.

fish bait, slingshot ammo, melt for hardware, etc.

b. May be an elaboration of a 4 response.
earrings, make a belt, mosaic, etc.

0. May be a slight play-on-words
pacifier, conversation piece, etc.

Uncommon use, perhaps demanding a remote substitution,
or a more obvious play-on-words.
dike hole plugger, pencil sharpener, flat marble,
peace offering, etc.

Very uncommon or unusual, (remote association).
Opaque monocle, phallic symbol, ant hole cover,
standard of weight, etc.

Elaboration of a response can increase its rating, e.g.,

while "jewelry" is rated as a 4,

This can be true of any rating level; while "ant hole
cover" is a play on words (man hole cover), it is more
unusual than an average 6 response.

30

"earring" is rated as a 5.

””11 ((131111.(fiflijffligxiifllfll (117“