Download as pdf or txt
Download as pdf or txt
You are on page 1of 14

JOURNAL OF RESEARCH IN SCIENCE TEACHING VOL. 12, NO. 1, PP.

49-62 (1975)

THE EFFECT OF EXPERIENTIAL SCIENCE ON


DEVELOPMENT OF LOGICAL THINKING
IN CHILDREN

MARCIA C. LINN AND HERBERT D. THIER


University of Gdifornia,Berkeley, Gdifornia 94 720

Introduction
The development of logical thinking has been studied extensively by Inhelder and
Piaget’ ’ *. The Science Curriculum Improvement Study (SCIS) has developed an experiential
physical and life science curriculum based, in part, on Piaget’s work. One of the goals of the
curriculum is to foster logical thinking in children. The upper-level units of SCIS emphasize
aspects of logical thought concerning isolation and control of variables.
The present study was undertaken to gather evidence of the effect SCIS has on the
development of logical thinking in children. In order to investigate this effect, comparisons
were made between fifth graders who had studied at least the Energy Sources unit from SCIS
and fifth and eighth graders who had not studied SCIS. During 1971-72 the fifth-level physical
science unit Energy Sources3 was available in final edition for the first time. This provided an
excellent opportunity to select classes for the study, since it was known which districts had
received this unit and since all teachers using the unit would necessarily be using it for the first
time.
An investigation of logical thinking was undertaken because this is one of the general
objectives of SCIS. Increasing logical thinking is a goal of many programs and a capability
considered necessary for adults. This goal is frequently put forth by school districts as a general
objective of their program. Therefore, this study is an attempt to pinpoint the effect of a
particular program in achieving a well-accepted educational goal.

Logical Thought
Inhelder and Piaget’ describe the growth of logical thought from age eight to age fifteen by
examining how subjects conceive, interpret, and execute experiments. During this period,
subjects move from what Piaget calls “concrete” reasoning to what he calls “formal” reasoning.
Concrete thought is characterized by limited extensions of empirical reality. Thus the concrete
thinker can organize observed data using classification, seriation, correspondence, and other
operations which maintain the information in the same form as it was presented to the child.
The concrete thinker begins to recognize physical properties of objects as variables, although
certain properties are easier to isolate from an experiment than others. During the concrete
stage, the number of properties that the subject recognizes to be variables increases, but his
ability to apply the operations of the concrete stage is limited by his ability to recognize a given
property as a variable.
49

01975 by the National Association for Research in Science Teaching


Published by John Wiley & Sons, Inc.
50 LINN AND THIER

Formal thought, according to Piaget, is characterized by a reversal of the importance of


“reality” and “possibility” in the subject’s approach to a problem. Thus the concrete thinker is
concerned with the “reality” of a situation (with the actual experiments that have been
performed), while the formal thinker is concerned with the “possibilities” inherent in an
experiment. The formal thinker looks at a given experiment as but one of all those possible to
verify a set of hypotheses. He is capable of generating many possible experiments to determine
the relationships among variables.
The phenomenon of compensating variables has been examined by Piaget frequently. To
explain why a heavy ball released from a low position travels as far as a light ball released from
a high position involves understanding compensation of variables. Inhelder and Piaget‘ have
shown that subjects at the beginning of the concrete stage cannot explain compensations. Even
if these subjects are shown how each variable (height and weight) affects the outcome of an
experiment, they cannot combine these in a compensation. Instead, they tend to pick one
variable or the other to explain the result. Toward the end of the concrete stage, the subject can
explain compensations when the relevant variables are isolated for him. At the end of the
concrete stage, the subject understands compensations involving two known variables but
cannot generalize this concept of compensation to the mutual compensation of three or more
factors. At the beginning of the formal stage, he can isolate the variables in a compensation as
well as explain results of experiments by using a compensation.
Thus, a compensating variable task, when direct experience with the relevant variables is
provided, is an appropriate measure of logical development to use with fifth grade subjects who
are ten or eleven years old. According to Piaget,’ subjects at the beginning of the concrete stage
(about age eight) cannot perform the task adequately while those at the end of the concrete
stage or the beginning of the formal stage (about age thirteen) can.

The SCIS Bogram


In the SCIS program, students are introduced to scientific content through their
experiences with physical and biological phenomena. In the course of their investigations, they
engage in observation, measurement, interpretation, prediction, and other scientific processes.
Central to the conceptual framework of SCIS is the view that objects interact (“do something
to one another”) in reproducible ways under similar conditions. The four major scientific
concepts used to elaborate interaction are matter, energy, organism, and ecosystem.
The Energy Sources unit focuses on matter and energy as vehicles for illustrating
interaction. The unit begins with a review of the concepts of property, interaction, system,
subsystem, and variable-all from previous units. During the remainder of the unit, the concepts
of energy source, energy receiver, energy transfer, and energy chain are introduced. Students
apply these concepts to situations in which either temperature change or motion provides
evidence of energy transfer. In each situation students, with the help of their teachers, choose
the major variables and set up experiments to investigate the effect of these variables on the
amount of energy transferred. The concept of compensation is not taught in the Energy
Sources unit.
The role of the teacher in the SCIS program is different from the teacher’s role in the
traditional book-oriented science program. According to one SCIS staff member, “The teacher
is not in the classroom to tell children what science is; she is there to help them find out what
science may be. She has to create an environment that invites and supports curiosity,
investigation and i n q ~ i r y . ” Teachers
~ of SCIS, according the SCIS staff, focus on the child’s
cognitive growth, take a guiding rather than a controlling role in the classroom, and believe that
children must take responsibility for their own learning? Other investigator^^"^^ have shown
EXPERIENTIAL SCIENCE ON DEVELOPMENT OF LOGICAL THINKING 51

that SCIS teachers tend to ask divergent, or higher-order, questions rather than convergent, or
“fact,” questions.
In developing the SCIS program, Piaget’s research on the stages of children’s intellectual
development was taken into account. Karplus: the director of SCIS, documents three
influences that Piaget’s work had on the SCIS program: (1) activities for the early grades had a
perceptual emphasis, for the middle grades the emphasis is conceptual and some abstract
referents are introduced in the upper grades; (2) concepts in the teaching program were chosen
to help advance students from one developmental stage to another by calling attention to
generalizable aspects of the student’s observations; and (3) since each student is free to use the
materials in the program as he wishes, he can adapt them to his immediate needs and is likely to
find satisfaction rather than frustration with most activities.
The SCIS program encourages intellectual development and logical thinking through both
the recommended teaching style and the design of the student activities. In order t o assess the
effect of Energy Sources on logical thinking it is appropriate to choose a task such as
compensating variables which is based on ideas taught in the unit but is not explicitly covered
by the unit and which is relevant to the general developmental level of students of Energy
Sources but not easily explained by them. Thus, the present study was undertaken to determine
the effect of SCIS on logical thinking.

Method
Evaluation Instrument
The Cart Experiment, used to evaluate ability to explain compensating variables, was a
group test presented on silent, 16mm film. The filmed demonstration of the experiment had
the advantage, of course, of being identical each time. Also, copies of the film could be used
across the country, thus insuring that the tests were all administered in a relatively short time
span. This proved to be an inexpensive, well-received, and very clear method of presenting an
experiment to large groups of children.
The apparatus used in the Cart Experiment film is shown in Figure 1. The cart had a
wooden frame and three roller skate wheels. The bumper at the front consisted of a plastic strip
attached to the wood and held at the appropriate angle by a string between the plastic strip and
a stick attached to the cart. The cart went down the ramp-from either the high or low
position, on either a blue or brown surface-hit the wall, and bounced back until it came to
rest. The blue surface was wood; the brown surface was soft packaging material. A calibrated
runway indicated how far the cart went on each trial, as shown in Figure 1. The cart always
came to rest between the number 15 and the number 75.
The experimenter administering the Cart Experiment had a single sheet containing
instructions for himself and the class. First the experimenter said: “Today you are going to see
a film of some experiments with a cart. The cart rolls down a ramp, hits a wall, and bounces
back. After each two trials, you will be asked to explain what you observed. Watch the film
carefully. We are interested in your explanation of what happens in the film. Explain as best
you can what you observe.” He then passed out a response sheet for Parts A and B of the
experiment. The response sheets are shown in Figure 2. Note the drawings of each trial on the
response sheets.
Part A. The experimenter then said: “Please place your name on the response sheet for
Parts A and B. We will now read Part A and then watch the first part of the film.”
He read Part A and started the film. The film showed the ramp and the cart and then
showed the cart being released from the high position, going down the ramp, hitting the wall,
and coming to rest at 75. Then the cart was released from the low position, hit the wd, and
wall

brawn s1irfat-p.

I I I

starting point 2 'starting point 1

Fig. 1. Apparatus used in the Cart Experiment f h .


c
A. The s a m e r a m p Is used in all trials shown In the lllm, s o m e t i m e s a
brown s u r f a c e IS placed on It. me s a m e c a r t wUl be used tn a l l trhls shown The s a m e cart and r a m p a r e used In both trlals. In t h e flrat t r h l t h e r a m p
has the brown moterlpl on the surface and the c u t Btarls f r o m lhe h t h
in the film.
position. In the secmd trlal the r a m p does n d have the brown surface m It
First the cart wIU be r e l e a s e d from two duferent pmltlons on lhe r a m p wlthoul and the cart slarta f r o m the low pmttlon. You wlll not see the end d elther
the brown aurface. trlal.

When dld the c a r t go the s h o r t e r dlstance7 (Check c


~ l l below) 1. Suppwe the e u t went j u t v far In # 1 u It dld In #2. How would you
Hkh LOW explaln uul 7
Rxplatn the d:wt praitlm h
a on how far the cart goes.

8. In the next two trwS the earl I s r e k w e d f r o m ths low p Q t l b n on the ramp.
The brown s u r f a c e la placed 011 Ule r a m p lor the l l n t trhl.

Wheo did the eul go the s h o r t e r dlstance? (Check one belor)


With brmm Burlace Wlthout b r o m s u r f a c e z
Exphln (hc effect t h e brown Burface h a s on how far the cart goen. x
z
n

Fig. 2. Response sheets for the Cart Experiment.


(n
W
54 LINN AND THIER

came to rest at 35. After the second trial, the film went black and was stopped by the
experimenter. The students then answered the questions in Part A.
Part B. When everyone had finished Part A, the experimenter read Part B and started the
film. The fdm showed the brown surface being placed on the ramp and then showed the cart
being released from the low position with the brown surface, hitting the w d , and coming to
rest at 15. Then the cart was released from the low position without the brown surface, hit the
wall, and came to rest at 35. The film went black and was stopped. The students answered the
questions in Part B.
Part C. The experimenter collected the papers from Parts A and B and distributed papers
for Part C (Figure 2). The experimenter said: “Please put your name on Part C. I will now read
Part C to you. Get ready to view the last part of the film.” He then read Part C and started the
film. The film showed the cart being released from the high position with the brown surface.
When the cart hit the wall, the film of the trial was interrupted. In the second trial the cart was
released from the low position without the brown surface and the film of the trial was
interrupted when the cart hit the wall. In fact, the carts were going at the same speed and both
stopped at 35. The film was interrupted because we found in pilot testing that this was the best
way to get the subjects to explain why the carts went the same distance rather than to describe
extraneous factors that they might have noticed or imagined when the trials were completed.
When the film was over, the students answered Part C. The experimenter then collected the
papers.

Summary Form
The experimenter was asked to fill out a form when he administered the Cart Experiment.
He noted the class size, grade, school, extent of background in SCIS, when and/or if Energy
Sources had been taught, and when he administered the experiment. Indications of the
environment of the area (urban fringe, suburban, or rural) were noted. Any questions that the
students asked and the teacher’s responses were also recorded on the form. The students’
papers, the summary form, the instructions for the experimenter, and the film were then placed
in the envelope provided and mailed to the investigators. The entire classroom procedure took
less than 40 minutes.

Scoring the Cart Experiment


Parts A and B of the Cart Experiment were designed to show the students the effect of two
variables: height and surface texture. If the student marked the effect of a variable incorrectly,
his written response was examined. Less than 5 percent of the subjects marked the effect
incorrectly and explained the effect incorrectly. Thus, Parts A and B insured that the subjects
had common knowledge about the two variables under investigation.
Responses to Part C for about 20 percent of the classes were examined by both the
investigators. Standards for unacceptable and acceptable responses for explaining a compensa-
tion were determined from this examination. They were as follows:
(a) Unacceptable-wrong: Subject gives no answer or says “I don’t know.” Subject reports that one
Cart goes faster or slower or further or shorter than the other, or describes only the condition of one
Cart (e.g., “High has brown surface”).
(b) Unacceptable-one variable: Subject explains the effect of one variable but not the other (e.g.,
‘‘Because brown slows it down”). Subject mentions second variable only as description of cart (e.g.,
EXPERIENTIAL SCIENCE ON DEVELOPMENT OF LOGICAL THINKING 55

“Brown is higher, other lower”). Subject does not explain second variable (e.g., “Both have
advantage and disadvantage because brown slows it down”).
(c) Acceptable-two variubks: Both variables are explained (e.g., “One cart has a higher start but
brown slows it down. Other starts low but has smoother surface”). Both variables are implied, even
if language is not precise (e.g., “One starts at shorter position but has smoother, other is high and
rough”). The variables are described and result is stated (e.g., “One is high on brown, other is low
on smooth so they come out even”). Only one trial is described clearly but compensation is implied
by use of “but” or similar connective which suggests opposition or compensation (eg., “High goes
faster but brown slows it down so come out the same”).

As a reliability check on the scoring, six classes (three experimental and three control) of
the ninety-two classes tested were selected at random. A member of the staff was given the
above standards and asked to score Part C of each paper. His ratings agreed with those of the
investigators in 90 percent of the cases. Where there were discrepancies, they did not follow any
pattern; some responses were given a higher rating by the staff member than by the
investigators and some a lower rating.

Selection Crireria
All the districts where SCIS had been used extensively for several years and which had
ordered the Energy Sources unit were considered for participation in this study. A list of public
school districts that had ordered over five Energy Sources kits was formed. Selection was made
in order to get a reasonable cross section of types of schools, types of students, and approaches
to teaching. Schools in rural areas, suburban areas, and urban fringe areas were chosen. Schools
where science was taught by a specialist and where science was taught by the classroom teacher
were included. Also schools in SCIS Trial Center areas and in areas far from Trial Centers were
involved.
When a district was selected, the science coordinator was contacted. All science
coordinators who were contacted wanted to participate but a few were unable to either because
by May school was nearly over or because the Engergy Sources kits had not been used for one
reason or another. The science coordinator was asked to describe the classes in his district
where Energy Sources was being taught. Other nearby classes which had not had Energy
Sources but were otherwise similar to these classes were then sought. These were usually in
nearby districts which had not used SCIS. Only intact classes of fifth graders were used. Classes
were matched on socioeconomic status of the school area and wealth of the district. Also, if
possible, one or two eighth grade classes similar to the fifth grade classes were selected and
tested. Since Energy Sources was available in final edition for the first time in 1972, no eighth
grader could have studied the unit.
In general, every class using Energy Sources or every fifth grade in a particular school was
used. All students present on the trial day were tested. The type of science taught in the
non-SCIS classes varied considerably, ranging from incidental science or the planned use of a
textbook to other laboratorycentered programs. Thus, a cross section of SCIS and non-SCIS
teachers was used.
Almost aU the SCIS teachers included in this study had some introduction to XIS before
teaching the program. This ranged all the way from contact with a trained science coordinator
to participation in a formal summer course as part of an SCIS Trial Center or a federally
subsidized training program. In addition, the teacher’s guide for each SCIS unit gives specific
instructions on how to teach the activities in the unit. Science in the elementary school is
generally taught by the classroom teacher. A few communities use science specialists to teach
science; one district using this approach was included in the study.
56 LINN AND THIER

Teacher's Experience
Since this was the first year that Energy Sources was available, this was the first time that
any teacher could possibly have been teaching the unit. Many of the SCIS teachers had used
other SCIS units in the past, while some non-SCIS teachers had taught SCIS lower-grade units
in previous years. Districts using SCIS involved in this study had all been using the program for
several years. Districts that adopted SCIS before the upper-grade units were available generally
used the lower grade units in the upper grades. Thus, this was nor the first time that most of the
SCIS teachers had used a materials-centered program. It is, nevertheless, possible that teachers
using Energy Sources for the first time were enthusiastic about a new unit and taught it
especially well. On the other hand their lack of familiarity with the materials may have made
them less effective in teaching the unit for the first time than they would be in teaching it the
second and subsequent times.

Subjects
Following the above selection procedure, science consultants from four urban fringe areas,
four suburban areas, and one rural area were contacted and agreed to cooperate. These areas
were in seven states: California, Colorado, Illinois, Maryland, Michigan, New Jersey, and
Pennsylvania. A total of seventeen school districts, ninety-two classes, and 2290 subjects were
involved.
The number of classes from each area in each group (experimental, control, eighth grade) is
given in Table I. As can be seen, equal numbers of classes in the experimental and control
categories were not always supplied. A greater effort was made to insure that experimental and
control classes were parallel (as far as socioeconomic status of the area and wealth of the district
were concerned), than to insure that equal numbers of experimental and control classes were
secured.

TABLE I
Means and Standard Deviations by Area of Percentage Scores
for Classes on Compensating Variables

Area Control Experimental Eighth Grade


State
Type' N2 M S.D. N* M S.D. N2 M S.D.
1. California S 4 21 8 2 58 6 3 58 4
2. Illinois S 5 28 13 5 54 14
3. Maryland S 11 31 10 8 55 20
4. New Jersey S 2 50 3 10 45 12 1 64
5. California UF 1 35 2 44 13
6. Colorado UF 2 42 2 4 49 13 2 54 24
I. Michigan UF 4 33 20 5 46 15 2 60 1
8. Pennsylvania UF 2 48 18 3 54 21 1 80
All Non-Rural I 31 33 13 I 39 50 15 9 60 12
9. Michigan R 5 21 11 8 43 6
All Classes 36 32 13 41 49 14 9 60 12

S = Suburban, UF = Urban Fringe, R = Rural


* Number o f classes
EXPERIENTIAL SCIENCE ON DEVELOPMENT OF LOGICAL THINKING 57

Results and Discussion


Since we were interested in the effect of the Energy Sources unit as presented by the
teacher, the class was chosen as the variable of interest. For each of the ninety-two classes in
the study, the percentage of subjects who gave a two-variable response to Part C of the Cart
Experiment was computed. Results for Part C of the Cart Experiment are given in Table I.
Looking at the means for the controls, it is clear that rural subjects had the lowest scores
and that subjects in New Jersey and Pennsylvania had the highat scores. It should be noted
that New Jersey and Pennsylvania controls were the only controls who were studying other
SCIS units (not Energy Sources) when they were tested. Looking at the experimentals, again,
the rural subjects have the lowest mean. For both experimentals and controls there are no
consistent differences between urban fringe and suburban districts. It appears that such a gross
variable is not important for success in compensating variables. Results from urban fringe and
suburban subjects are therefore analyzed together. The rural subjects, on the other hand, did
score below urban fringe and suburban subjects. These findings parallel regional results for
seventeen-year-olds reported by the National Assessment of Educational Progress." No
consistent pattern of differences emerged for urban fringe and suburban populations in the
national assessment results. On the other hand, where there were differences between rural
subjects and other subjects, the rural subjects always had lower scores than the other subjects.
In reading the responses from this study, the authors concluded that rural subjects in this
sample had poorer writing skills than nonrural subjects. It is possible that scores for these
subjects would increase somewhat more than would scores for nonrural subjects if the test
were given in an interview.

Urban Fringe and Suburban Populations


Since the number of classes tested in each urban fringe and suburban area was different,
the mean and standard deviation of class scores for the area were computed. These areas
represent a cross section of all urban fringe and suburban areas using SCIS. Thus, to investigate
the effect SCIS has on the understanding of compensating variables, it is appropriate to
compare experimentals in each area with controls in the same area. The mean and standard
deviation of the scores for classes in each area were computed. For the eight areas, the control
mean was 0.37, standard deviation 0.09. The experimental mean was 0.51, standard deviation
0.05. To test the significance of the difference between the means, a t ratio was computed. This
analysis yielded a t of 3.8 p < .01.* Thus, we can conclude that classes studying SCIS in urban
fringe and suburban areas understand compensating variables better than classes not studying
SCIS in these areas. The difference is illustrated in the histograms in Table 11.

Rural Population
The rural area included in this study was in Michigan. Comparing scores for experimental
and control classes, those for experimentals were significantly higher than those for controls
(t = 4.1, p < .01). This was the only rural area investigated, since it was the only one that
ordered five Energy Sources kits in 1971-1972. It comprises four separate districts (two
experimental and two control). Thus, students of SCIS in a representative rural area do better
on compensating variables than do students in this rural area not using SCIS.

*The Satterthwaite" correction was used to control for unequal standard deviations. This reduced the
degrees of freedom to 1 1.
58 LINN AND THIER

TABLE I1
Scores on the Compensating Variables Question

Control Classes
7
3
3 9
3 8 6
3 9 5 6
2 7 3 3 3
9 9 1 7 3 1 3 3 4 8
2 9 1 2 3 1 2 2 4 7
5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80

I Percentage of students correct in the class' 1


Experimental Classes
9 9
9 6 7
8 9 6 7
I 9 4 5 4
5 9 4 4 4
7 4 7 3 9 3 3
4 4 6 2 9 3 3 6 3 8
3 4 2 4 2 8 1 1 2 2 3
5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80

Percentage of students correct in the class'


1
Eighth Grade Classes
7
7
1
6 1 1 4 6 8

5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80

Percentage of students correct in the class'

' Numbersrefer to states and area types in Table I . Classes with the same number are from the same area.

Comparison of Fifth and Eighth Graders


There were only nine eighth grade classes in the study and they came from five of the eight
urban fringe and suburban areasinvolved. Unllke the fifth graders, all the eighth grade classes in a
given school were nor tested. Thus, although coordinators were asked to select representative
classes, the eighth grades do not necessarily comprise a random sample of eighth graders.
Nevertheless, as shown in the histograms in Table 11, the mean for experimental classes is closer
to the mean for eighth graders than it is to the mean for controls. Thus, experimentals
performed significantly better than controls and approached the ability level of eighth graders.
Comparison of fifth and eighth grade controls confirms research by Inhelder and Piaget3
showing that older children are more likely to explain compensations adequately than younger
children. Clearly since only 60% of the eighth graders could answer acceptably many could
EXPERIENTIAL SCIENCE ON DEVELOPMENT OF LOGICAL THINKING 59

profit from additional experience with variables. Many eighth grade pupils do not demonstrate
ability to perform at the level of formal operations.

Sex Differences
The percentage of boys and of girls giving a two-variable response to the compensating
variables question was computed. For all classes 41 percent of the boys and 47 percent of the
girls answered correctly (Xz= 6.4 p < 0.02).Looking at girls vs. boys in experimental, control,
and eighth-grade groups, girls scored slightly better than boys but none of the differences were
significant. Thus the small but significant overall difference between boys and girls can
probably best be attributed to the well-documented superiority of girls in language skill.’*
TeacherEffectiveness
Since experimental classes scored significantly better than control classes on compensating
variables, it follows that experimental teachers were generally effective in teaching Energy
Sources. The standard deviation for all experimental classes is very close to the standard
deviation for all control classes, and the histograms (Table 11) for the two groups have similar
shapes. Thus, the class scores for experimental teachers, all using the same materials, were as
vaned as the class scores for control teachers using different materials.
One might have expected a larger standard deviation for experimental classes than for
control classes, since some teachers might have succeeded with SCIS while others might have
failed. This was not the case. In fact, an inspection of the histograms of scores for experimental
and control classes (Table 11) reveals that only five experimental classes (9 percent) fall at or
below the control mean and eighteen (38 percent) exceed all but two control classes. As can be
seen in Table 11, the highest scoring experimental classes were from the same area as the lowest
scoring experimental classes. The one lowscoring experimental class was a special group. We
controlled for special grouping, which was infrequent in the schools studied, by always testing
every class in a given school.
Except for the classes in Colorado, each class in this study was taught by a different
teacher. In Colorado, specialist science teachers were used. Each of the two teachers taught two
experimental classes and one control class. These teachers were in different schools in the same
district. The differences between scores for experimental and control classes for these teachers
were not significant (t = 1.2). For such a small sample (N = 6) no definite conclusion can be
drawn. It appears that the teacher was more important than the curriculum in this case, but it
may be that the teacher carried ideas from one science program over t o another. Also, the
comparability of the classes in this case is unknown since they were all in the same school and
the children may not have had random access to classes.
Analysis of Student Responses to the Cart Experiment
A random group of experimental and control responses to Part C of the Cart Experiment
was analyzed in order to determine which aspects of the Energy Sources unit might have been
most effective in influencing performance on this compensating variables question. As was
expected, subjects who answered the question at all demonstrated understanding of the height
variable in one way or another. Height is understood empirically at this age and was the starting
point for most responses. Once the child looked at the question as a height question, he then
used the brown surface to explain the unexpected result of a low release point “going as far as”
a high release point.
As stated earlier, answers which were scored as “two variables” explained the role of height
explicitly and related height to surface by either opposition (using “but”) or convergence (using
“and”). Analysis of wrong responses from four classes revealed the following categories in
descending order of sophistication:
60 LINN AND THIER

1. Subject explains that “brown slows it down” so they are equal. Height is not
mentioned but is implicit in the answer. These subjects would probably give two-variable
answers in an interview.
2. Subject described the relevant variables but does not reason about them (eg., “They
had different heights and surfaces”).
3. Subject gives an incomplete reason using one variable (e.g., “High places goes further”).
4. Subject refers to some concept such as “energy” or “power” but does not explain how
it works (e.g., “They both had the same power”).
5. Subject does not answer question.

Both experimental and control subjects in the classes sampled made each of the types of errors
listed above, although experimentals made fewer errors altogether.
Students of Energy Sources have many experiences with explaining the roles of certain
variables in experiments although they never use cars on inclined planes or experiment with
compensation. Experimentals with more experience in explaining the roles of variables were
probably more likely than controls to refer to the height variable explicitly instead of
implicitly. Also, students of Energy Sources learn to isolate a variable and investigate its effect
while holding the others constant. Undoubtedly this would help them to reason about
individual variables in a complex situation instead of merely describing the variables. This
experience would also be likely to predispose the student of Energy Sources to describe the
variables in a situation instead of only referring to some concept such as “power” or “energy.”
Thus, analysis of the types of errors made on the Cart Experiment indicated that the
Energy Sources unit would be likely to help subjects on this activity by giving them experience
with explicitly naming and isolating variables. Although it is possible to hypothesize what the
experimental child learned so that he could explain compensating variables, it is not clear from
this study how he learned these things.
Part of the answer to how these things are learned can be determined by comparing the
kinds of experiences that occur in SCIS to those that occur in most other programs. The major
difference between SCIS and most other programs is that in SCIS children work with materials
and set up experiments in small groups. Children do their own experiments with materials
provided in the kit. They gather data, record observations, and discuss (and argue about) their
conclusions with their peers. Piaget rarely comments on how logical t h k i n g can be taught, but
he has stated that learning is the result of the constant interaction of the individual and his
environment and that children learn through arg~mentation.’~ Thus, learning about variables
by doing experiments with concrete materials and discussing them with peers would be
consistent with Piaget’s theory.
Linn and Pete~son,’~ in studying the effect of SCIS on the logical thinking of first graders,
found that experience in manipulating materials facilitated performance on explaining why
some objects float and others sink. These findings are in line with those of Almy” who found
that culturally disadvantaged children (who presumably have fewer opportunities to work with
intellectually stimulating materials) perform below middle-class chddren. Thus, experience with
materials as provided in the Energy Sources unit is probably important in fostering logical
development.
Summary and Conclusions
Evidence of logical thinking, as measured by ability to explain compensating variables,
shown by fifth graders who took the SCIS Energy Sources unit was compared with that of fifth
and eighth graders who did not take the unit. It was found that fifth graders who studied
Energy Sources appeared to be better logical thinkers than comparable fifth-grade students who
did not. Also, students of Energy Sources had scores that were close to those of eighth graders
in comparable areas. Gains on an important measure of logical thinking resulting from a
EXPERIENTIAL SCIENCE ON DEVELOPMENT OF LOGICAL THINKING 61

commercially available science program are very encouraging. These gains are even more
noteworthy since the teachers of the classes in this sample represent a cross section of all
teachers using the SCIS program. Although most teachers of SCIS have some contact with
science coordinators who have been introduced to the program by SCIS staff members, these
teachers do not ordinarily have extensive special training.
As a group, classes studying SCIS were distributed about the group mean in much the same
way as classes not studying SCIS were distributed about their group mean. Differences between
classes and teachers which presumably account for this dispersion were probably the same in
each group. This situation might change as teachers use the units for the second and subsequent
times.
Looking at which aspects of Energy Sources were most likely to have affected logical
thinking, it appears that students learn to explicitly name variables, and they use materials to
do experiments while interacting with their peers. These fudings indicate which aspects of
school experience are likely to be useful in promoting logical thinking. Additional research is
needed to determine if more experiences with SCIS-type materials, either at the fifth-grade level
or prior to fifth grade, would increase the number of subjects who could solve the
compensating variables problem. It would also be worthwhile to determine if other non-SCIS
experiences are necessary to teach compensating variables to those who do not succeed after
studying Energy Sources.
Looking at the eighth graders in the study, about 60 percent could solve the compensating
variables question correctly. Since skill in written expression as well as skill in logical thinking
increases between fifth and eighth grade, differences between the experimental fifth graders and
the eighth graders could be attributed to both improved language skill and improved logical
thinking. At any rate, fifh graders are performing close to eighth graders after studying SCIS.
This provides an opportunity to develop educational experiences for fifth graders whose logical
thinking ability approaches that of eighth graders after they study Energy Sources.
There were large individual differences in scores on the Cart Experiments for both fifth and
eighth grade classes in this study. Also some fifth grade classes that studied SCIS had as many
students near the formal level as many of the eighth grade classes. Thus educational programs
which encourage intellectual development need some mechanism for coping with wide ranges of
ability in the classroom. The XIS program accomodates these variations in ability by providing
materials for each child to use in an environment that encourages investigation without
demanding the same results from every student. Another method for coping with individual
differences in intellectual development is to provide opportunities for students to choose their
own materials and their own problems in an atmosphere that encourages experimentation.
Purely verbal responses, according to Piaget, are not sufficient to determine whether a
subject is a formal thinker. Piaget also states that ability to explain compensations when the
variables are demonstrated beforehand is indicative of the most advanced level of concrete
operations. The verbal nature of the Cart Experiment prohibits determining whether children
are at formal operations, but subjects who give “two variable” responses are at least at the
advanced concrete level. Any test which depended on manipulation of materials and individual
analysis of subjects would require considerably more time for administration. Since there is a
wide range of scores for classes on the Cart Experiment at this age level, it was suitable for the
investigation.
The districts and schools in this study were not selected randomly, but they do represent a
cross section of districts and schools using SCIS.The control schools are closely matched to the
experimental schools on the basis of socioeconomic status, wealth of district, and other
variables. Thus, the experimental and control classes in each area are from the same population.
The lower scores for the rural compared with the non-rural subjects, however, may be due to
62 LINN AND THIER

chance variation in selection of schools and districts rather than to real differences between
populations.

Availability o f the Cart Film


Copies of the Cart Film are available to interested researchers. Please contact the authors
t’or details.

*This research was supported in part by a grant from the National Science Foundation to AESOP
(Advancing Education through Science-Oriented Programs). We wish to express special thanks to Dr. Warren
Wollman for assistance in categorizing student responses. We wish also to thank all the leadership people
Lind teachers who assisted in administering the Cart Experiment.

References

1. Inhelder, B., and Piaget, J., The Growth of Logical Thinking from Childhood to
Adolescence, Basic Books, Inc., New York, 1958.
2. Inhelder, B., and Piaget, J., The Early Growth of Logic in the Child, Harper & Row,
New York, 1964.
3. Science Curriculum Improvement Study, Energy Sources, Rand McNally & Company,
Chicago, 1971.
4. Kageyama, C., “From Foreground to Background: The Changing Role of the
Teacher,” SCIS Newsletter, No. 9 , Winter 1967.
5. Haan, N., A n Exploratory Investigation of the Effect of an Initial Experience with
SCISS Material Objects Unit on First-Grade Children and Their Teachers. Far West Laboratory
for Educational Research and Development, Berkeley, 1968.
6. Moon, T. C., “A Study of Verbal Behavior Patterns in Primary Grade Classrooms
during Science Activities,” unpublished doctoral dissertation, Michigan State University, 1969.
7. Porterfield, D. R., “Influence of Preparation in Science Curriculum Improvement
Study on Questioning Behavior of Selected Second and Fourth-Grade Reading Teachers,”
unpublished doctoral dissertation, University of Oklahoma, 1969.
8. Wilson, J. H., “The ‘New’ Science Teachers are Asking More and Better Questions,”
Journal of Research in Science Teaching, 6:49-53, 1969.
9. Karplus, R., “Strategies in Curriculum Development: The SCIS Project,” to be
published by the Aries Corporation.
10. National Assessment of Educational Progress, Report 1, Science: National Results, U.S.
Government Printing Office, Washington, D.C., 1970.
11. Satterthwaite, F. E., “An Approximate Distribution of Estimates of Variance
Components,” Biometries Bulletin, 2 : 110-1 14, 1946.
12. Maccoby, E. E., “Sex Differences in Intellectual Functioning,” in The Development of
Scx Differences, edited by E. E. Maccoby, Stanford University Press, Stanford, 25-55, 1966.
13. Piaget, J., The Language and Thought of the Child, Routledge and Kegan Paul,
London, 1952.
14. Linn, M. C . , and Peterson, R. W., “The Effect of Direct Experience with Objects on
Middle Class, Culturally Diverse, and Visually Impaired Young Children,” Journal of Research
in Science Teaching, 10: 83-90, 1973.
15. Almy, M., Young Children’s Thinking: Some Aspects o f Piaget’s Theory, Teachers
College, Columbia University, New York, 1966.

You might also like