Sunteți pe pagina 1din 41

Teachers' Pedagogical Content Beliefs in Mathematics

Author(s): Penelope L. Peterson, Elizabeth Fennema, Thomas P. Carpenter and Megan Loef
Source: Cognition and Instruction, Vol. 6, No. 1 (1989), pp. 1-40
Published by: Taylor & Francis, Ltd.
Stable URL: http://www.jstor.org/stable/3233461
Accessed: 06-09-2017 01:39 UTC

REFERENCES
Linked references are available on JSTOR for this article:
http://www.jstor.org/stable/3233461?seq=1&cid=pdf-reference#references_tab_contents
You may need to log in to JSTOR to access the linked references.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
http://about.jstor.org/terms

Taylor & Francis, Ltd. is collaborating with JSTOR to digitize, preserve and extend access to
Cognition and Instruction

This content downloaded from 136.167.3.36 on Wed, 06 Sep 2017 01:39:13 UTC
All use subject to http://about.jstor.org/terms
COGNITION AND INSTRUCTION, 1989, 6(1), 1-40
Copyright ? 1989, Lawrence Erlbaum Associates, Inc.

Teachers' Pedagogical Content Beliefs in


Mathematics

Penelope L. Peterson
Michigan State University

Elizabeth Fennema, Thomas P. Carpenter,


and Megan Loef
University of Wisconsin-Madison

This study examined relationships among first-grade teachers' pedagogical


content beliefs, teachers' pedagogical content knowledge, and students'
achievement in mathematics. Teachers (N = 39) completed structured
questionnaires and interviews on their beliefs and knowledge about instruc-
tion, children's learning, and the mathematics content in addition and
subtraction. Results indicated significant positive relationships among teach-
ers' beliefs, teachers' knowledge, and students' problem-solving achievement.
Compared to teachers with a less cognitively based perspective (LCB teach-
ers), teachers with a more cognitively based perspective (CB teachers) made
extensive use of word problems in introducing and teaching addition and
subtraction. They also spent time developing children's counting strategies
before teaching number facts. CB teachers had greater knowledge of word-
problem types and greater knowledge of their children's problem-solving
strategies than did LCB teachers. Furthermore, CB teachers obtained this
latter knowledge by observing their children in problem situations rather than
by relying on tests or formal assessments. Children with CB teachers scored
higher on word problem-solving achievement than did children with LCB
teachers, but children from both types of classes did equally well on
addition/subtraction number facts.

Over the past decade or so, research on teaching has increasingly suggested
that, to gain a more complete understanding of how classroom instruction
affects students, researchers need to view the teacher not only as one who
engages in certain classroom behaviors, but also as an active processor of

Requests for reprints should be sent to Penelope L. Peterson, College of Education,


Erickson Hall, Michigan State University, East Lansing, MI 48824-1034.

This content downloaded from 136.167.3.36 on Wed, 06 Sep 2017 01:39:13 UTC
All use subject to http://about.jstor.org/terms
2 PETERSON, FENNEMA, CARPENTER, LOEF

information before, during, and after classroom instruction (Clark &


Peterson, 1986; Shavelson & Stern, 1981; Shulman & Elstein, 1975). In the
most recent review of research on teachers' thought processes, Clark and
Peterson (1986) concluded that the research substantiates the view of the
teacher as a reflective, thoughtful individual. Moreover, the research
documents that teaching is a complex and cognitively demanding process.
Teachers' beliefs, knowledge, judgments, thoughts, and decisions have a
profound effect on the way they teach as well as on students' learning in
their classrooms. In addition, teachers' beliefs, thoughts, judgments,
knowledge, and decisions affect how teachers perceive and think about
teaching a new curriculum that they receive and to what extent they
implement the training or curriculum as intended by the developers.

THE SUBJECT-MATTER CONTENT IN RESEARCH ON


TRAINING

Recently, Shulman (1986) pointed out that a major limitation of this new
cognitive research on training is the absence of a focus on the subject matter
to be taught. He referred to this blind spot with respect to content as the
"missing paradigm" among the various research paradigms for the study of
teaching (Shulman, 1986, pp. 6-7). Similarly, in a recent review of research
on mathematics teaching, Romberg and Carpenter (1986) noted that
research on teaching and research on learning mathematics have been
conducted as two separate disciplines of scientific inquiry. They suggested
that a pressing need exists for researchers from these two research traditions
to interact and collaborate in their research on teaching mathematics and,
in particular, for researchers on teaching to incorporate an analysis of the
mathematics content into their studies of teaching.
In showing how the subject-matter content might be integrated into the
study of teachers' thought processes, Shulman provided the example of
teachers' "pedagogical content knowledge" as an important category of
teacher cognitions that are related to teachers' thoughts and behavior in the
classroom, as well as to their students' learning and achievement. He
defined teachers' pedagogical content knowledge as including the following:

The ways of representing and formulating the subject that make it compre-
hensible to others ... alternate forms of representation, some of which derive
from research where others originate in the wisdom of practice ... an
understanding of what makes the learning of specific topics easier or difficult:
The conceptions and preconceptions that students of different ages and
backgrounds bring with them to the learning of those most frequently taught
topics and lessons. (Shulman, 1986, p. 9)

This content downloaded from 136.167.3.36 on Wed, 06 Sep 2017 01:39:13 UTC
All use subject to http://about.jstor.org/terms
TEACHERS' BELIEFS IN MATHEMATICS 3

Researchers might bring a similar analysis and incorporation of the


subject-matter content into the study of other categories of teachers'
thought processes, including teachers' cognitions (Peterson, 1988) and
teachers' beliefs.

TEACHERS' PEDAGOGICAL CONTENT BELIEFS

As persons whose daily task is to understand and interpret the rapi


events in a classroom, and to make decisions and act on their in
tions, teachers obviously rely on their knowledge and beliefs. In
their review of the small number of studies on teachers' beliefs
Peterson (1986) concluded that "a teacher's cognitive and other
are guided by and make sense in relation to a personally held
beliefs" (p. 207). Unfortunately, research on teachers' beliefs ha
concerned with subject-matter content; rather, it has focused o
general conceptions of their roles (Janesick, 1977; Munby, 198
general beliefs about curriculum (Bussis, Chittenden, & Amarel,
the general principles they use to explain their own interactive
(Connors, 1978; Marland, 1977). Thus, these studies are limited
they report findings on teachers' beliefs across a wide range of
areas and grade levels. A few researchers have attempted to ta
matter into account in analyzing teachers beliefs. For example
researchers have examined teachers' beliefs about mathematics curricula
(Collier, 1972; Meyerson, 1977; Shirk, 1972; Skemp, 1978; Thompson,
1982). However, these researchers tended to examine teachers' beliefs abou
mathematics or about curriculum and instruction in general. They did no
analyze teachers' beliefs within a specific topic area and grade level, such
teaching and learning of addition and subtraction in first grade.
In contrast, cognitive-science researchers have frequently focused on
learning within a specific content domain. They have demonstrated that a
analysis of performance within a specific content domain leads to greate
prediction and understanding of the learner's performance than does
analysis not focused specifically on the content domain (Carpenter &
Peterson, 1988; Newell & Simon, 1972; Romberg & Carpenter, 1986).
In our study, we attempted to (a) conceptualize teachers' pedagogical
content beliefs; (b) provide an example of how teachers' pedagogical
content beliefs in mathematics might be analyzed, specifically, in teaching
and learning addition and subtraction in first-grade mathematics; and (c)
understand and describe how teachers' pedagogical content beliefs in
addition and subtraction influence teachers' thinking, decision making,
teaching, and, ultimately, their students' learning and achievement. We
began by developing a conceptual framework for analyzing teachers'
pedagogical content beliefs in addition and subtraction.

This content downloaded from 136.167.3.36 on Wed, 06 Sep 2017 01:39:13 UTC
All use subject to http://about.jstor.org/terms
4 PETERSON, FENNEMA, CARPENTER, LOEF

A Framework for Analyzing Teachers' Pedagogical


Content Beliefs

From our review of the scholarly research, we derived four separat


interrelated constructs that represent fundamental assumptions underly
much of the contemporary cognitive research on children's learning
Case, 1983; Cobb, 1988; Resnick, 1981, Resnick & Ford, 1981).

1. Children construct their own mathematical knowledge. Cognit


science researchers have shown that children develop informal syste
mathematics outside the classroom. Children do not simply absorb
they are taught; they structure and interpret the presented mathem
curriculum and instruction in light of their existing knowledge (see,
Cobb, 1988; Resnick & Ford, 1981). The assumption might be represe
as a continuum, ranging from children actively construct their ow
knowledge at one end to children passively receive mathematical knowled
from the teacher or others at the other end.
2. Mathematics instruction should be organized to facilitate child
construction of knowledge. A second assumption related to the first
that mathematics instruction should be organized to facilitate childr
knowledge construction. A contrasting assumption might be that m
matics instruction should be organized to facilitate the teacher's cl
presentation of knowledge. Attempts to draw instructional implica
from recent cognitive science research on children's learning of mathem
favor the former assumption (e.g., Cobb, 1988; Confrey, 1986).
3. Children's development of mathematical ideas should provide t
basis for sequencing topics for instruction. A third assumption rela
the first two is that children's naturalistic development of mathem
ideas should provide the basis for sequencing topics for instruction
e.g., Case, 1983; Case & Bereiter, 1982). A contrasting notion might be
the structure of mathematics should provide the basis for sequencing top
for instruction. If the structure of mathematics were to provide the bas
sequencing topics for instruction, it follows that instruction would
with teaching formal mathematical symbolism and procedures. How
basing sequence on the naturalistic development of ideas acknowledge
children use direct modeling strategies and counting strategies before th
can come to understand and use the more abstract number facts and written
symbolism that accompany number facts.
4. Mathematical skills should be taught in relation to understanding and
problem solving. A fourth assumption is that mathematical skills should be
taught in relation to understanding and problem solving. A contrasting
perspective would be that mathematical skills should be taught as discrete
components in isolation from understanding and problem solving. Both
perspectives assume that the learning of mathematics skills, understanding,

This content downloaded from 136.167.3.36 on Wed, 06 Sep 2017 01:39:13 UTC
All use subject to http://about.jstor.org/terms
TEACHERS' BELIEFS IN MATHEMATICS 5

and problem solving are all important goals of mathematics instructio


However, the two propositions make very different statements both abo
their relationships and about the most effective ways to achieve these goals.

RESEARCH QUESTIONS

In our study, we used the above framework to examine and desc


first-grade teachers' pedagogical content beliefs in addition and subtra
We also explored the possibility of important links among teachers'
pedagogical content beliefs, teachers' pedagogical content knowledge in
addition and subtraction, and students' achievement. We were guided by
the following research questions:

1. What are teachers' pedagogical beliefs about mathematics, curricu-


lum, and instruction in addition and subtraction?
2. How are teachers' pedagogical content beliefs reflected in their reports
of their approaches to teaching, their choices of mathematics content,
their concepts of the roles of the teacher and the learner, and their
goals for instruction?
3. Is there a relationship between teachers' pedagogical content beliefs
and their students' achievement, including achievement of computa-
tional skills (number facts) as well as problem solving?
4. What is the relationship between teachers' pedagogical content beliefs
and their pedagogical content knowledge?

METHOD

Participants

The participants in the study were 39 first-grade teachers and their stude
in 27 schools located in Madison, Wisconsin, and four small rural comm
nities near Madison. The schools included 3 Catholic schools and 24 pu
schools. Six teachers taught first and second graders; the others taught on
first graders. All teachers in the sample volunteered to participate
month-long in-service program in mathematics the following summer
to be included in a study of classroom instruction the following year.
mean number of years teaching elementary school was 10.90 in our sample
and the mean number of years teaching first grade was 5.62. Two of
teachers were in their first year of teaching. Thirty-four of the teachers
participated in in-service courses in the last 3 years, and 9 of them
participated in courses dealing with mathematics. Twenty-two of th
teachers indicated that they were familiar with activity-based programs s
as Comprehensive School Mathematics Project, Developing Mathematic

This content downloaded from 136.167.3.36 on Wed, 06 Sep 2017 01:39:13 UTC
All use subject to http://about.jstor.org/terms
6 PETERSON, FENNEMA, CARPENTER, LOEF

Processes, or Mathematics Their Way; and 6 teachers used one of these


programs for instruction. The other 33 teachers used eight different
textbooks.

Instrumentation

We used two techniques for assessing teachers' beliefs. First, we developed


a 48-item belief questionnaire to assess teachers' beliefs according to the
four constructs described above. Second, we designed a structured interview
to obtain more specific information on teachers' conceptions of mathemat-
ics, curriculum, the roles of teacher and student, and their goals for
instruction in addition and subtraction.

Belief questionnaire. The belief questionnaire was designed to assess


teachers' pedagogical content beliefs in addition and subtraction. The
questionnaire consisted of 48 items designed to assess teachers' beliefs on
four subscales or constructs. Each subscale consisted of 12 items to which
the teachers responded on a 5-point Likert scale by indicating strongly
agree, agree, undecided, disagree, or strongly disagree.
Each of the four subscales on the belief questionnaire was designed to
measure interrelated but separate constructs. The first subscale was con-
cerned with how children learn mathematics. A high score on this subscale
indicated a belief that children construct their own knowledge, whereas a
low score on the construct reflected a belief that children receive knowledge.
Half the items on each subscale were worded so that agreement with the
statement indicated agreement with a cognitively based perspective, whereas
the other half of the subscale items were worded so that agreement with
them indicated less agreement with a cognitively based perspective. Exam-
ples of items on Subscale 1 include:

Most young children can figure out a way to solve simple word
problems.
It is important for a child to discover how to solve simple word
problems for him/herself.
It is important for a child to know how to follow directions to be a
good problem solver.
Children learn math best by attending to the teacher's explanations.

The second subscale measured a construct that was concerned with the
relationship between mathematical skills and understanding and problem
solving. A high score on this construct indicated a belief that skills should
be taught in relationship to understanding and problem solving, whereas a
lower score indicated an agreement with the belief that skills should be

This content downloaded from 136.167.3.36 on Wed, 06 Sep 2017 01:39:13 UTC
All use subject to http://about.jstor.org/terms
TEACHERS' BELIEFS IN MATHEMATICS 7

taught in isolation from understanding and problem solving. Examples


items on Subscale 2 include:

Time should be spent solving simple word problems before children


spend much time practicing computational procedures.
Children will not understand addition and subtraction until they have
mastered some basic number facts.
Children should not solve simple word problems until they have
mastered some basic number facts.
Children should understand computational procedures before they
spend much time practicing them.

The third subscale assessed teachers' beliefs about what should provide
the basis for sequencing topics in addition and subtraction instruction. A
high score on this construct indicated a belief that children's natural
development of mathematical ideas should provide the basis for sequencing
topics for instruction, whereas a low score on this construct indicated a
belief that formal mathematics should provide the basis for sequencing
topics for instruction. Examples of Subscale 3 items include:

The natural development of children's mathematics ideas should


determine the sequence of topics used for instruction.
The mathematical sequence must be considered in planning for
instruction.
It is more important to use children's concept development in planning
an instructional sequence than to use a mathematically determined
sequence.
When selecting the next topic to be taught, one must consider the
logical organization of mathematics.

The fourth subscale was designed to assess teachers' beliefs about h


addition and subtraction should be taught. A high score on this constr
reflected a belief that mathematics instruction should facilitate children's
construction of knowledge, whereas a low score reflected a belief that
instruction should be organized to facilitate teachers' presentation of
knowledge. Examples of Subscale 4 items include:

Teachers should allow children to figure out their own ways to solve
simple word problems.
Teachers should teach exact procedures for solving word problems.
Teachers should tell children who are having difficulty solving a word
problem how to solve the problem.

This content downloaded from 136.167.3.36 on Wed, 06 Sep 2017 01:39:13 UTC
All use subject to http://about.jstor.org/terms
8 PETERSON, FENNEMA, CARPENTER, LOEF

The teacher should demonstrate how to solve simple word problems


before children are allowed to solve problems.
It is better to teach children how to solve one kind of word problem
at a time.

The internal consistency of teachers' scores on each subscale was esti-


mated using Cronbach's alpha with the scores from the sample of 39
teachers. The internal consistency estimates for each subscale are presented
in Table 2. The internal consistency of teachers' scores on the total belief
scale was .93.

Belief interview. Teachers' beliefs related to the four constructs were


also assessed through a structured interview. The interview was designed to
obtain two kinds of information on teachers' beliefs about mathematics,
curriculum, and instruction in teaching addition and subtraction. First, we
wanted to obtain specific information on the content and techniques that
teachers used to teach addition and subtraction in their classrooms. We also
wanted teachers to describe specifically their own curriculum objectives and
their conceptions of the roles of student and teacher in their mathematics
classes. Second, we wanted to use another method to obtain information on
teachers' beliefs about the constructs. In other words, we wanted to use an
alternative assessment to demonstrate construct validity through a
multimethod approach. For this reason, we often had teachers provide
rationales for many of their statements throughout the interview. Then we
asked interviewers to rate teachers on the four constructs following the
interview, and we also had trained coders rate each teacher on the four
constructs after reading the written protocols of the interviews.
The interview followed a structured interview format. The interviewer
first asked the question as written and then followed the teacher's answer
with a probe only once if necessary. Probes included, "Anything else?",
"Can you tell me more?", or repeating or paraphrasing the question.
Throughout the interview, if the interviewer thought that the teacher had
already answered a question, the interviewer would say, "The next question
is [reads next question]. I think that you already answered that question. Do
you think that you answered it? Is there anything else you want to add?"
The interviewer began the interview with the following directions:

We are interested in how you teach addition in your classroom and


why you teach it the way that you do. We are going to ask you some
specific questions about the way you teach addition and why you
teach as you do. There are no right or wrong answers to these
questions. We are interested in your opinions and ideas. We would
like you to respond to the specific questions. However, after you

This content downloaded from 136.167.3.36 on Wed, 06 Sep 2017 01:39:13 UTC
All use subject to http://about.jstor.org/terms
TEACHERS' BELIEFS IN MATHEMATICS 9

answer each question, please feel free to depart from the original
question if you have additional comments on how or why you teach
addition as you do. As you can see, we are audiotaping your
responses.

After giving the directions, the interviewer then proceeded with th


structured questions on the interview. All teachers' responses to the
interview were audiotaped and later transcribed. Interview questions and
procedures included the following:

1. "Describe as specifically as you can the lesson in which you introduc


addition to your class. We are interested in the way you organize and
present the mathematics content as well as the specific teaching method
and strategies that you use."
2. The interviewer then selected three things that the teacher said that he
or she did, such as the way he or she sequenced the lesson, organized th
lesson, explained certain concepts, or taught a specific skill. Then for eac
of these things, the interviewer asked, "Why did you decide to do that?"
3. "What do you think the role of the teacher should be in introducin
addition to first graders?"
4. "What do you think the role of the learner should be in an introduc
tory addition lesson?"
5. The interviewer repeated Questions 1 through 4, except that the
interviewer asked the teacher to describe as specifically as possible a typical
lesson on addition in his or her class, how a typical lesson differedfrom the
introductory lesson on addition, and whether and how the typical lesson
might change over the course of the school year.
6. "Are there certain kinds of word problems in addition that you believe
all children should learn to solve? If so, what are they?"
7. "In teaching addition in first grade, what do you believe should be th
relative emphasis on fact knowledge versus understanding versus solving of
word problems? Why?"
8. "What do you see as the relationship between learning of addition
facts, understanding addition, and solving addition word problems?"
9. "Children have different abilities and knowledge about addition. Ho
do you find out about these differences?"

Following the interview, a score for each teacher on each of the fou
belief constructs was obtained by having the interviewer rate the teacher on
a 5-point scale for each construct. The interviewer was directed to thin
back over the teacher's responses and judge where the teacher fell on th
continuum for each of the four belief constructs. If the interviewer was
unable to make a judgment for one or more beliefs, she was told to ask

This content downloaded from 136.167.3.36 on Wed, 06 Sep 2017 01:39:13 UTC
All use subject to http://about.jstor.org/terms
10 PETERSON, FENNEMA, CARPENTER, LOEF

additional questions to get the information needed to make a judgment.


Interviewers did not know the teachers prior to interviewing them and had
not seen the teachers' responses to the belief questionnaire.
Teachers' responses to the belief interview were audiotaped and later
transcribed. Two trained coders read the written protocols and prepared
brief summaries of the teachers' responses. Each coder read and summa-
rized half the teachers' protocols. In addition, they read and rated a small
common set to demonstrate interrater agreement. After reading the written
protocol and preparing a summary of the teachers' responses, each coder
coded the teacher on the four belief constructs using the same 5-point rating
scale used by the interviewers. While preparing their summaries and rating
the teachers, the coders were blind both to the teacher's identity and to the
teacher's scores on the belief questionnaire.

Assessment of pedagogical content knowledge. We used several


techniques to measure a wide range of teachers' pedagogical content
knowledge in addition and subtraction (see Carpenter, Fennema, Peterson,
& Carey, 1988, for a complete description of these measures). Here we
focus on three measures:

1. Relative problem difficulty. To test teachers' ability to distinguish


the relative difficulty of different word-problem types, teachers were given
16 pairs of word problems. They were asked to identify which of the two
problems was more difficult for first-grade children, or they could respond
that the problems were of equal difficulty. Pairs of problems were drawn
from classes of known difficulty based on a body of cognitive research on
problem solving (Carpenter et al., 1988).
2. General knowledge of strategies. Teachers' general knowledge of
strategies that children use to solve addition and subtraction problems was
assessed by showing them a videotape of three children solving different
word problems and asking the teachers to show how these children would
solve related problems. Each videotaped episode was repeated twice before
the teachers were asked to demonstrate how the child would solve other
problems, and the teachers responded to all questions about a given child
before the next episode was played.
3. Knowledge of individual students' strategies. Teachers' knowledge
of their own students was tested by asking them to demonstrate how si
randomly selected students in their classes would solve six different addition
and subtraction word problems. The same six problems were administere
to the six students 1 or 2 days before the teacher interview. Two scores were
generated from this test. A correct-answer score was based on the teachers'
success in predicting whether a student could solve a given problem
correctly irrespective of whether the teacher predicted the strategy that the

This content downloaded from 136.167.3.36 on Wed, 06 Sep 2017 01:39:13 UTC
All use subject to http://about.jstor.org/terms
TEACHERS' BELIEFS IN MATHEMATICS 11

student used. A strategy score was based on the teachers' success in


accurately predicting the strategy the given child would use to solve each
problem.

Number-fact (computation) test. To assess students' achievement of


number-fact knowledge, we developed a test of 20 addition and subtraction
computation items utilizing basic number facts. Ten items involved sums
less than 10, and 10 items involved sums between 10 and 19. Sixteen items
were written vertically, and 4 items were written horizontally. Addition and
subtraction items were ordered randomly in the test. Students were given 2
min to complete the test. The internal consistency of the number-fact test
was estimated at .90 using Cronbach's alpha (N = 710 students).

Problem-solving test. Student achievement was also assessed using a


problem-solving test consisting of 17 word-problems. The test included 9
addition and subtraction problems representing a range of word-problem
types as described by Carpenter et al. (1988). The other 11 problems
involved several operations, included extraneous numbers, or required
grouping and partitioning. All numbers in the problems were less than 20.
Each problem in the problem-solving test was printed on a separate page of
the test booklet. The test administrator read the problems to the students
and instructed the students not to turn to the next problem until directed to
do so. Cronbach's alpha for the problem-solving test was .84 (N = 710).

Procedure

Teachers completed the belief questionnaire and the structured belief


interview in spring of 1986. Each teacher completed the belief intervi
followed by the belief questionnaire during a 1- to 2-hr individual ses
with a trained interviewer. The interviewers were three graduate stude
Prior to beginning the interviews, the graduate students were trained in an
8-hr training course extending over a period of a week. During the we
each interviewer practiced the interview by administering it to seve
different teachers. Following each practice interview, one of the princ
investigators listened to and critiqued the audiotaped interviews. In ad
tion, the graduate students listened to one another's interviews. The
interviewers practiced until they reached "criterion" by administering
interview in a similar fashion.
Student achievement was also assessed during spring of 1986. Trained
testers administered the computation and problem-solving tests to all
students in the 39 teachers' classrooms. The testers followed written
protocols in administering the tests.
The audiotapes of teacher interviews were transcribed verbatim. Wr

This content downloaded from 136.167.3.36 on Wed, 06 Sep 2017 01:39:13 UTC
All use subject to http://about.jstor.org/terms
12 PETERSON, FENNEMA, CARPENTER, LOEF

transcripts were then summarized and rated by trained coders as described


earlier. Both complete written protocols and summaries of the teacher
interviews were used in later analyses.
Item responses from the teacher belief questionnaire and student achieve-
ment tests were entered by hand onto a microcomputer data base. All
statistical analyses were then performed using the Statistical Package for the
Social Sciences (SPSS; Nie, Hull, Jenkins, Steinbrenner, & Bent, 1975).

RESULTS AND DISCUSSION

In our presentation of the results and discussion, we focus on eac


research questions. We begin by describing teachers' beliefs abou
matics curriculum and instruction in addition and subtraction.

Variation Among Teachers in Their Beliefs About


Mathematics, Curriculum, and Instruction in
Addition and Subtraction

Table 1 presents the means and standard deviations of teachers' sco


the four constructs as measured by the belief questionnaire. A higher s
on the construct indicates greater agreement with a cognitively
perspective. The maximum score on each belief construct was 60;
minimum score was 12.
As can be seen from Table 1, significant variation existed among teachers
in their scores on the four belief constructs. Teachers differed in the degree
to which their beliefs corresponded to a cognitively based perspective.
Constructs 1, 3, and 4 showed significant variation with neither "ceiling"
nor "floor" effects in teachers' scores. Teachers had the highest average
agreement with the perspective that "math skills should be taught in
relationship to problem solving" rather than "in isolation from problem
solving." In contrast, teachers had the lowest average agreement with the
perspective that children "construct" mathematical knowledge rather than
"receive" mathematical knowledge.
Table 1 also presents interviewers' and coders' ratings of the interview
protocols for each of the four constructs. As was the case for the belief
questionnaire, interviewers' ratings and coders' ratings showed significant
variation among all teachers in their scores. A maximum score for both
interviewers' and coders' ratings was 5, and the minimum score was 1.
As shown in Table 1, both interviewers' and coders' ratings tended to fall
around the middle of the scale; the average for all teachers tended to fall
around 3. However, as indicated by the large standard deviations for
interviewers' ratings and coders' ratings, significant variation existed among

This content downloaded from 136.167.3.36 on Wed, 06 Sep 2017 01:39:13 UTC
All use subject to http://about.jstor.org/terms
TABLE 1
Means and Standard Deviations of Teachers' Scores on Constructs as Measu
and by Interviewers' and Coders' Ratings of Belief In

Belief Questionnaire Interviewer's Rati

All CB LCB All CB


Teachers Teachers Teachers Teachers Teachers
Belief Interview (N = 39) (n = 7) (n = 7) (N = 40) (n = 7) (n

Children construct/ M 40.51 47.57 31.57 2.65 4.


receive math knowledge. SD 6.82 5.80 4.61 1.33 0
Math skills should be M 49.59 58.29 42.71 3.67 4
taught in relation to/in SD 6.22 1.97 3.30 1.09 0
isolation from problem
solving.
Instruction should be M 45.36 53.57 38.57 2.27 4
sequenced to build on SD 5.56 3.01 2.37 1.43 1.
children's development of
math ideas/structure
of math.
Instruction should M 45.05 53.00 37.57 2.87 4.4
be organized to facilitate SD 5.89 5.29 2.88 1.50 0
children's construction/teacher
presentation of knowledge.
Total across M 180.51 212.43 150.43 11.48 17.43
all belief constructs. SD 21.13 11.04 6.83 4.71 3

Note. CB Teachers = teachers with a co


1; LCB Teachers = teachers with a less c

This content downloaded from 136.167.3.36 on Wed, 06 Sep 2017 01:39:13 UTC
All use subject to http://about.jstor.org/terms
14 PETERSON, FENNEMA, CARPENTER, LOEF

the teachers in the extent to which they were rated as having beliefs that
corresponded to a cognitively based perspective. As was the case for the
belief questionnaire, the highest ratings by both interviewers and coders
were given on Belief Construct 2, which suggests that, in comparison to the
other three constructs, teachers' views about Construct 2 more closely
corresponded to a cognitively based perspective than did their views on the
other constructs.
We turn now to an examination of the differences between teachers
whose beliefs corresponded more closely to a cognitively based perspectiv
and those whose beliefs corresponded less closely to that perspective.

Comparison of Teachers Whose Beliefs About


Mathematics Instruction Differed According to a
Cognitively Based Perspective

As we have seen, the 40 teachers in this study varied significantly in the


degree to which their beliefs about mathematics, mathematics instruction,
and curriculum reflected a cognitively based perspective. On each construct,
we found significant variation among teachers in their questionnaire scores;
however, teachers' scores on the four belief constructs were significantly
positively correlated. Table 2 presents the intercorrelations among teachers'
scores on the four cognitive-science belief constructs (range = .57 to .76).
Because teachers' scores on the four belief constructs were highly
intercorrelated, teachers' implicit beliefs about mathematics instruction and
curriculum might be considered either as one global construct or four
related constructs on which we might array teachers. Moreover, we might
rank teachers from high to low on the constructs and examine differences
between teachers in their approach to first-grade mathematics, curriculum,
and instruction.

TABLE 2
Correlations Between Teachers' Construct Scores on the
Mathematics Belief Questionnaire

Belief Construct 1 2 3 4

1. Children construct/receive math knowledge. .86 .60 .57 .76


2. Math skills should be taught in relation to/in isolation from
problem solving. .75 .71 .59
3. Instruction should be sequenced
development of math ideas/structure of math. .86 .75
4. Instruction should be organized to facilitate children's
construction/teacher presentation of math knowledge. .81

Note. Numbers on the diagonal are estimates of internal consistency usi


alpha (N = 39 teachers; 12 items on each belief scale).

This content downloaded from 136.167.3.36 on Wed, 06 Sep 2017 01:39:13 UTC
All use subject to http://about.jstor.org/terms
TEACHERS' BELIEFS IN MATHEMATICS 15

To explore these ideas further, we began by ranking teachers from hig


to low on each construct. We also calculated the mean, median, and
standard deviation for each construct. Figure 1 presents the rankings of
teachers by ID numbers on the four belief constructs.
Consistent with the high intercorrelations found between the belief
constructs, Figure 1 demonstrates that teachers who scored high on a given
belief construct also tended to score high on the other three constructs. In
contrast, teachers who scored low on a construct tended to score low on the
other three constructs. For purposes of further analyses, we selected 7
teachers who scored consistently high on all four constructs as measured by
the questionnaire and 7 teachers who scored consistently low on all four
constructs as measured by the questionnaire. The selected teachers are
shown in Figure 1. In the following discussion, we refer to the teachers who
were high on the belief constructs as teachers with a cognitively based
perspective (CB teachers) and the teachers who were low as teachers with a
less cognitively based perspective (LCB teachers). We computed t tests to
determine if these two groups of teachers differed significantly in their
average scores on each of the four constructs as well as on their total
construct score. Table 1 shows the mean belief scores for the two groups
and indicates that CB teachers scored significantly higher than LCB
teachers on all the belief constructs from the questionnaire.
Descriptive information on the teachers showed that CB teachers had
more years of teaching experience (M = 14.57) than did LCB teachers (M
= 8.00), and this difference was statistically significant, t(12) = 2.15, p <
.05. Thus, the differences that we discuss between CB and LCB teachers
may reflect differences in experience alone, or they may be more appropri-
ately considered as differences in the accumulated "wisdom of practice" that
underlies teachers' pedagogical content knowledge and beliefs (Shulman,
1986, 1987). We found no other systematic demographic differences
between teachers. For example, we found no pattern of differences between
CB and LCB teachers in socioeconomic status of their children or in type of
school in which they taught. Indeed, two schools had one CB teacher and
one LCB teacher in the same school, thus suggesting that differences
between the LCB and CB groups were not due to school philosophy. Also,
no differences were noted between CB and LCB teachers in their partici-
pation in in-service courses and workshops or in reported use of activity-
based mathematics curricula.
To describe and examine further the differences between teachers on the
belief constructs, we turned to the data obtained from interviews with
teachers. In support of the validity of our constructs, we found significant
positive intercorrelations among teachers' total scores on the belief con-
structs as measured by the belief questionnaire, interviewers' ratings, and
coders' ratings of the interview protocols. In addition, the CB teachers and

This content downloaded from 136.167.3.36 on Wed, 06 Sep 2017 01:39:13 UTC
All use subject to http://about.jstor.org/terms
Scale 1: Children Construct/Receive Math Knowledge Scale 2: Sk
from
C)

57-60 57-60

53-56 (Q 0 53-56
30
3@@
49-52 49-52

45-48 0(00000 45-48


41-44 ? 0000000median 41-44

37-40 )E)e0000000 37-40

33-36 @ (00 33-36

29-32 (i)? 29-32


25-28
25-28 ()
21-24 ( 21-24

median = 41.0, mean = 40.5, s.d. = 6.8 median

Scale 3: Instructional Scale


Sequence 4: Instr
Based
Structure Presen

57-60 ( 57-60 ()
53-56 (g)? ? 53-56 @()
49-52 Q?000 49-52 @00
45-48 0000 median median
45-48 ?
41-44 0@000000000000 41-44 }QQ
37-40 @??)@? 37-40 ?( )
33-36 ( 33-36 ()@(
29-32 29-32
25-28 25-28
21-24 21-24

median = 44.0, mean = 45.36, s.d. = 5.59 median = 45.

FIGURE 1 Distribution of scores on four belief construc


including Teachers 2, 4, 12, 16, 51, 56, and 63 and the LCB subgroup includin

This content downloaded from 136.167.3.36 on Wed, 06 Sep 2017 01:39:13 UTC
All use subject to http://about.jstor.org/terms
TEACHERS' BELIEFS IN MATHEMATICS 17

the LCB teachers differed significantly in their scores on the belief


constructs as measured by both interviewers' and coders' ratings of proto-
cols. These comparisons are shown in Table 1.
In examining the verbal interview protocols of the CB and LCB teachers,
we found several important differences between the groups in their pur-
ported approaches to teaching addition and subtraction in first grade, their
conceptions of the mathematics curriculum in addition and subtraction,
their conceptions of the roles of the teacher and the learner, and their goals
for instruction. We turn now to a discussion of each of these differences.

Early Emphasis and Use of Word Problems in


Teaching Addition and Subtraction

We examined the interview protocols of teachers to determine if teachers


whose implicit beliefs about mathematics instruction in first grade agreed
with a cognitively based perspective would also agree with the early
introduction of word problems in teaching addition/subtraction. We found
that this was the case. When asked the question, "Do you have your
children work on addition word problems any time during the first few
months of the school year?" the CB teachers uniformly responded "Yes."
In contrast, 6 of the LCB teachers responded "No" to the question, and only
1 LCB teacher responded "Yes."
Moreover, in providing a rationale for their use of word problems in
teaching addition from the very beginning, CB teachers' responses were
particularly revealing. For example, one CB teacher gave the following
rationale for the early use of story problems:

Teacher 16: Because I think the kids can solve the problems if they
relate it to something in their life, rather than if I put an algorithm up
they would say, "I have no idea what that means," but they sure do
know what that means if I put it in a story context.

The following two CB teachers also provided similar rationales for their use
of story problems:

Teacher 2: I guess the same, again, the relation to their lives. I really
think that little children, whatever we do in school, have a very
difficult time relating . . . so, I really think you have to make it
[mathematics] relevant to their little lives, and you have to point it
out.

Teacher 56: I guess it's just part of me that life is reality, that math is
life. . . .

This content downloaded from 136.167.3.36 on Wed, 06 Sep 2017 01:39:13 UTC
All use subject to http://about.jstor.org/terms
18 PETERSON, FENNEMA, CARPENTER, LOEF

Thus, the CB teachers tended to express the belief that, in teaching


addition and subtraction concepts and skills, they needed to make these
concepts relevant to students' lives to enhance their understanding. Also,
these teachers seemed to realize that first graders had the capabilities to
solve word problems, at least orally, as soon as they entered their
classrooms. Further, these teachers believed that building on students'
capabilities and interests in solving real-world problems would motivate as
well as promote cognitive knowledge growth in children by building on
children's existing informal knowledge structures. Moreover, CB teachers
indicated that they initially presented word problems orally because their
children did not have the necessary language skills to read the word
problems at the beginning of first grade. However, the CB teachers realized
that, despite their lack of reading abilities, beginning first graders are quite
capable of conceptualizing the language in addition word-problems when
presented orally.

Teachers' Rationales for Not Using Addition


Word-Problems During the First Few Months

Interestingly, the primary reason that LCB teachers did not introduce
addition/subtraction word problems during the first few months seemed to
be because the teachers followed the mathematics textbook, and the
textbook did not do so. Five of the 6 LCB teachers who reported not using
word problems early in the year indicated that they did not do so because
they followed the textbook. The LCB teachers' rationales for not using
word problems early in the year to teach addition seemed to reflect a belief
that the sequence of mathematics instruction should be based on the
underlying structure of mathematics as reflected in the mathematics text-
book rather than on the sequence of children's development of mathemat-
ical ideas. This construct was one of the four belief constructs measured on
the belief questionnaire, and as noted previously, all the LCB teachers
obtained scores below the median on this construct.
Although reliance on the textbook seemed to be the LCB teachers'
primary reason for not introducing addition word problems in the first few
months, two other reasons were mentioned. Three of the 6 teachers
indicated that first-grade students needed to have some reading skills prior
to working on addition/subtraction word problems (Teachers 6, 8, and 17).
In addition, 2 of the 6 teachers mentioned that knowledge of some number
facts was a prerequisite for students' understanding mathematics word
problems (Teacher 6 and 66). This perspective is reflected in the following
teacher's response:

Teacher 6: I give them a few months to begin reading. It's hard to


throw them right into it. ... I just think that they need to have some

This content downloaded from 136.167.3.36 on Wed, 06 Sep 2017 01:39:13 UTC
All use subject to http://about.jstor.org/terms
TEACHERS' BELIEFS IN MATHEMATICS 19

successful experiences at reading and start to get to know the math


facts, what the signs mean, and that type of thing before you can
really start.

Similarly, another LCB teacher (Teacher 66) commented:

Teacher 66: As far as the children reading it [the word problem] alone,
they are not capable.

Interviewer: How about reading them aloud? Do you do that?

Teacher 66: Yes, but I find that sometimes they are not real good at
a structured word problem early. But if you present that type of a
word problem too early, they just don't want to do it; they find it
hard. And after they've learned some of their number facts, they seem
a little bit more willing.

These responses seem to reflect the underlying belief that number facts
should be taught as discrete components rather than in relationship to
understanding and problem solving. Consistent with this perspective of
teaching word problems separately from teaching math facts, 2 of the 6
LCB teachers who did not use word problems early indicated that when they
did teach problems, they taught them using a discrete-skill approach in
which they taught step-by-step algorithmic procedures to solving the word
problems by focusing on key words in the problem. An example of this
approach to solving word problems is the following:

Teacher 17: What I do first is I introduce what a word problem is. We


go through a step-by-step procedure: Learning how to take a story
apart, learning how to read the story problem, what we are looking
for in a story problem, how to figure out if we are going to add or
subtract in a story problem. So, it's just a step-by-step procedure.

A similar reliance on a step-by-step procedure to teach addition/


subtraction word-problem solving as a discrete skill is represented by the
following LCB teacher:

Teacher 11: I start out reading them [word problems] to the class and
see where they are on the page. And I teach the key words, "What does
'in all' mean?" "In all" means that we need to add it, and, there again,
direct teaching and how you do it: You put it - the things - over here
on the side, and you add it up, and, of course, you know the answer
to this, and then you put it over on to the blank.

This content downloaded from 136.167.3.36 on Wed, 06 Sep 2017 01:39:13 UTC
All use subject to http://about.jstor.org/terms
20 PETERSON, FENNEMA, CARPENTER, LOEF

The responses of these 2 teachers might be examined together with the


responses of the 3 teachers who mentioned that students need either
prerequisite reading skills or number-fact skills to begin working on
addition/subtraction word problems. Taken together, the responses dem-
onstrated that 5 of the 6 LCB teachers who indicated that they did not use
word problems seemed to do so because they viewed teaching addition/
subtraction word problems as teaching a discrete skill that should be
introduced later, after first graders have acquired the necessary skills for
learning to solve addition/subtraction word problems.

Relative Emphasis on Number Facts, Mathematical


Understanding, and Word Problems

During the interview, teachers were asked, "In teaching addition in first
grade, what do you believe should be the relative emphasis on fact
knowledge versus understanding versus solving a word problem?" A
cognitively based perspective would put greater emphasis on teaching
understanding and problem solving with less emphasis on memorizing and
teaching number facts. In keeping with this perspective, we found that, in
their interview responses, CB teachers definitively rated the learning of
number facts as least important when compared with understanding and
problem solving. Five of the 7 CB teachers clearly stated that number fact
learning would be the lowest priority by indicating "facts are last," "fact
knowledge is the least of my concerns," "fact is low on my list," "the very
last would be memorization," and "memorizing facts thirdly." The re-
maining 2 CB teachers clearly implied that fact knowledge was lowest on the
list of priorities by stressing that their highest priorities and emphases in
instruction were on understanding and using word problems. Thus, by
stressing their emphases on understanding and problem solving, they clearly
indicated that fact knowledge was of much less concern to them. CB
teachers' perspective on the relative emphasis that should be given to
understanding, fact knowledge, and word-problem solving is illustrated by
the following teachers' response:

Teacher 16: I guess understanding is what I would go with-that they


understand numbers and what they are doing, and how they are using
it, and how they can use numbers to get from where they are to where
they were in the beginning. So I guess I would have to say under-
standing [is more important].

Interviewer: Okay. Next?

This content downloaded from 136.167.3.36 on Wed, 06 Sep 2017 01:39:13 UTC
All use subject to http://about.jstor.org/terms
TEACHERS' BELIEFS IN MATHEMATICS 21

Teacher 16: It would be word problems. The very last would be


memorization.

Interviewer: Why?

Teacher 16: Because I find that they learn the facts by doing the other,
but it doesn't work in reverse.

In contrast to the CB teachers, the LCB teachers were more varied in their
rankings of the relative emphases that they placed on number facts,
mathematical understanding, and word problems. Although 2 of the LCB
teachers (Teachers 11 and 66) indicated that they placed the least emphasis
on knowledge of number facts, the remaining 5 teachers either placed
number facts as first or second in order of priority. Two teachers (Teachers
6 and 8) indicated that they put the highest priority on both number facts
and understanding; the remaining 3 teachers (Teachers 17, 54, and 64)
indicated that they emphasized understanding first, number facts second,
and word problems last. Thus, given that all 14 teachers placed mathemat-
ical understanding first in order of importance, the important difference
between LCB and CB teachers seemed to be in their relative emphases on
number facts versus word problems. CB teachers emphasized word prob-
lems as the basis for teaching mathematics and assumed that fact knowledge
would follow, whereas LCB teachers emphasized fact knowledge and
assumed that they would teach word problems after students had learned
their number facts.
The LCB teachers' conception of the relationships among fact knowl-
edge, understanding, and word problems might best be typified by the
following response:

Teacher 8: I guess you need understanding of the concepts before you


can do the facts, and you need facts before you can do the word
problems because you might understand the word problems, but if
you can't add, it's useless. So, I would say that understanding is
needed for facts, and then facts are needed for word problems.
Understanding obviously is needed, though, for word problems, too.
Understanding what addition means, just that old concept of putting
together. Understanding, I guess, would include, though, a lot of
concepts other than facts or even word problems, more than, less
than-that type of situation.

Similarly, an LCB teacher who responded that she would put the most
emphasis on facts and understanding provided the following rationale:

This content downloaded from 136.167.3.36 on Wed, 06 Sep 2017 01:39:13 UTC
All use subject to http://about.jstor.org/terms
22 PETERSON, FENNEMA, CARPENTER, LOEF

Teacher 6: Because if they don't know the facts, then they can't do the
word problem, plain and simple. I mean, if they can't add the facts
together, then they can't even, you know, when they have a word
problem, they need to know the facts to do a word problem.

All 14 teachers indicated that they emphasized mathematical under-


standing first. These findings may indicate that the interview question was
"loaded" in the direction of social desirability so that most teachers tended
to provide what they thought was the desirable response-that they taught
for mathematical understanding by their students. These findings might
also suggest that the word understanding is so global and all inclusive that
it can have very different meanings to very different types of teachers and,
thus, does not differentiate clearly between teachers in terms of their
content beliefs and the specific content and instruction that they deem
important in mathematics.

Specific Strategies and Methods for Teaching


Addition and Subtraction

Teachers' underlying beliefs about mathematics and instruction we


reflected clearly in their interview descriptions of the specific strateg
methods that they used for teaching addition and subtraction. For exam
teachers were asked to describe as specifically as possible the lesson in w
they introduce addition to their class. CB teachers and LCB diffe
significantly in both the kind of mathematics content they repo
presenting and the type of mathematics strategies that they rep
employing to teach introductory lessons in addition. The followin
protocol (Teacher 16) is representative of the CB teachers and serv
basis for contrasting the protocols of the LCB teachers. When ask
describe her introductory lesson on addition/subtraction, CB Teach
reported:

I guess the very beginning of addition is way back to counting object


And they can count the objects, and we do a lot to make sure that th
have one-to-one correspondence. We use manipulatives. We use k
of story problems along with it to make it interesting to listen to. I
it in small groups. I try to get a feel for where the children are
From there we go on to lots of story problems with manipulatives an
showing me before we ever do any algorithms or anything like t
We also do a lot of front counting and skip count so they feel r
good about what number goes next. . . . The next step would be
had one, and I count one or add one, and then what do I have? So, I
just show the relationship between front counting and adding. ... We

This content downloaded from 136.167.3.36 on Wed, 06 Sep 2017 01:39:13 UTC
All use subject to http://about.jstor.org/terms
TEACHERS' BELIEFS IN MATHEMATICS 23

do back counting, too, so we go forward and backward and forward


and backward and forward and backward before we ever do any
algorithms. Then we do the story problems with manipulatives. We do
more than and less than before. ... I don't use a lot of worksheets
and things at the beginning. We just do things.

In contrast, LCB Teacher 6 provided a rather different description of h


introductory lesson:

I guess I would use manipulatives at the beginning of the year to teach


the concept of adding two groups together and then counting in-all
and also take kids in front of the room here, and you add two and
show the class, as a large group, by using the children themselves.
Then alongside of any manipulatives, I would have the written
numbers or have the numbers written on the board so that they can
associate the manipulatives with the written numbers. ... I guess you
would want to make sure that they knew what the signs were-the
addition sign, what that meant. That meant that you were adding it.
And just the concept of "in all" for addition or "altogether." Just
using blocks or manipulatives or the Unifix cubes to show them how
to bring them together. ... From the beginning of the year, I showed
everyone how to use a number line ... they know how to add on the
number line and which direction to go.

A second perspective is provided by the following LCB teacher (Teacher


64):

You've taught all the numerals, the concept of the numerals, and what
a numeral is based on, so you start teaching the concepts of adding
very simple numbers, you know, 1 + 1. I use the blackboard to a great
extent. I use manipulatives. I have situations on the blackboard where
I have 1 + 1 = a whole. You talk about the signs, the add-on sign and
the equal sign. You talk about the two parts equaling a whole. Lots of
drills. I work on lots of drill, where they've got pictures or 1 star + 1
star = how many stars? Then they have the picture situation, and then
I do a mathematical sentence with numerals .... Then I work with the
families. The 1 + 1 = 2, turn it around, 1 + 1 = 2, 2 - 1 = 1, so
that they can put those into situations. On those first lessons, when
I'm teaching the math, I work with situations up to 5, sums up to 5.

In comparing these descriptions of the introductory lessons of the CB


teacher with the two LCB teachers, we see several important differences.
One apparent difference is that the two LCB teachers reported using written

This content downloaded from 136.167.3.36 on Wed, 06 Sep 2017 01:39:13 UTC
All use subject to http://about.jstor.org/terms
24 PETERSON, FENNEMA, CARPENTER, LOEF

symbolism in their introduction of addition and subtraction. For example,


they reported teaching the written symbols for the numbers, for the
addition and subtraction signs, and for the equal signs. They also reported
introducing the idea of writing mathematical sentences with numerals to
represent the problem. In contrast, the CB teacher above did not mention
use of written symbols or written symbolization to represent the problem.
As a second apparent difference, although both the CB and LCB teachers
reported using manipulatives and counters in their introductory lessons,
they seemed to use manipulatives and counters for very different reasons.
LCB teachers seemed to use manipulatives to provide a concrete represen-
tation of the written symbols; for example, a certain number of counters
would represent the written number, or two groups of counters put together
would represent the process of adding together as represented by the
addition sign. In contrast, the CB teacher had the children use counters as
an integral part of developing children's counting strategies for solving
addition word problems.
A third apparent difference between the CB and LCB teachers is that in
her introductory lesson the CB teacher built on children's informal strate-
gies for solving addition problems. She seemed to realize that children
naturally begin by using counting strategies to solve addition problems.
Through her instruction and working on these strategies, she showed how
counting is related to addition, and she facilitated children's development of
counting strategies to solve addition problems. Finally, an obvious differ-
ence appeared between the CB and the LCB teachers in their reports of
explicit teaching of number facts in introductory addition/subtraction
lessons. The LCB teachers implied that they began teaching number facts
early on whereas the CB teacher clearly indicated that she did not emphasize
"algorithms" until much later.

The roles of teacher and learner. Several other differences between


the groups of teachers were apparent in their descriptions of the role of the
learner and the teacher in the introductory addition lessons. Some of these
differences seemed to be an extension of those already described and
appeared to reflect underlying beliefs about mathematics and children's
mathematical knowledge. CB Teacher 16 described the role of the teacher
as:

. . providing them [first graders] opportunities to use it [addition


If we take attendance, how many are we supposed to count? If
someone is gone, we back-count 1 and that kind of thing, so I should
be able to count that many. Put it to use so they [the children] use it,
besides making sure conceptually, they know what they are doing.

In describing the role of the learner, she continued:

This content downloaded from 136.167.3.36 on Wed, 06 Sep 2017 01:39:13 UTC
All use subject to http://about.jstor.org/terms
TEACHERS' BELIEFS IN MATHEMATICS 25

... like the teacher also has to be the learner. She has to pay attention
to where the kids are, learn from them where they are, and dictate
what her next step is because there are a lot of different learners and
learners' styles. I guess learners must be actively involved in doing
work and think about what they are doing and verbalizing what they
are doing.

In contrast, LCB Teacher 6 described the role of the teacher as follows:

It's a big role. I have taught first graders that when they first came in,
they didn't have any concept of what adding was. ... The teacher has
to do it step-by-step, and you have to explain it daily; you have to go
over it until they start getting the concept. Subtraction was even
harder for them to understand. It took a long time. But they just
didn't understand that "take away" was sending a couple of kids out
of the room or in a take away you would take two blocks away or give
them to someone else. You would have to verbalize it and talk through
it almost every day, so they would start getting that concept.

She described the role of the learners in her classroom as follows:

They should have some manipulatives so that they have an under-


standing of addition and subtraction .... They should also be ex-
pected to write the number out, not just do it as manipulatives, but
also have the numbers alongside so that they associate the two all the
time. ... Do the seatwork, review the concepts, and get an under-
standing of it.

In sum, the LCB and CB teachers' views of the roles of teacher and
learner expressed very different perspectives. For example, the CB teacher
seemed to view the teacher and the learner as actively engaged with one
another in construction of mathematical knowledge and understanding. In
contrast, the LCB teachers seemed to view the teacher's role as one of
organizing and presenting mathematical knowledge and the child's role as
receiving the mathematical knowledge presented by the teacher. Together,
teachers' descriptions of the roles of teacher and learner and of the specific
teaching strategies and mathematics content seemed to reflect strongly the
belief constructs that we measured on the belief scale. The protocol of CB
Teacher 16, for example, showed clearly that in her introductory addition
lesson she was taking a view of children as constructing mathematical
knowledge and a view that instruction should be sequenced to build on
children's development of mathematics ideas and organized to facilitate
children's construction of mathematical knowledge. In contrast, the de-
scriptions of the lessons taught by the two LCB teachers (6 and 64) reflected

This content downloaded from 136.167.3.36 on Wed, 06 Sep 2017 01:39:13 UTC
All use subject to http://about.jstor.org/terms
26 PETERSON, FENNEMA, CARPENTER, LOEF

an idea that instruction should be sequenced to build on the structure of


mathematics as embodied in mathematics textbooks, and that instruction
should be organized to facilitate teachers' presentation of mathematical
knowledge. The LCB teachers' early reliance on symbolism, for example,
reflected a notion that symbolism is integrally related to the underlying
structure of mathematics, and thus, the LCB teachers seemed to be guided
by the implicit question "What are the important dimensions in the
underlying structure of mathematics, as reflected in mathematics textbooks,
that need to be taught?" rather than the question "How shall I build on
children's existing knowledge and informal mathematical strategies?"

Teachers' Beliefs and Students' Achievement

We have suggested that teachers' implicit beliefs about mathematics in


tion should be significantly related to their students' mathematics lear
and achievement. We have seen, for example, that when compare
teachers whose beliefs did not correspond closely to a cognitivel
perspective, teachers whose implicit mathematics did correspo
cognitively based perspective differed in their reports of their tea
addition in important ways. First, CB teachers reported usin
problems as a basis for introducing addition, and they reported co
emphasis on solving word problems throughout the year. Second
reported placing least emphasis on students' learning number fa
greater relative emphasis on solving word problems and gaining ma
ical understanding. These latter findings suggest important diff
between CB teachers and LCB teachers in their content selection and their
relative emphases in teaching addition. Presumably, if the content and
quality of instruction in addition and subtraction affect students' learning,
the content and quality of students' learning and achievement will differ
between CB and LCB teachers.

Correlations between teachers' beliefs and student achievement.


We explored the relationship between teachers' cognitive-science beli
and students' mathematics achievement in two ways. First, we comput
correlations between teachers' scores on the belief constructs and their
students' achievement in addition and subtraction. Table 3 presents thes
correlations. Teachers' total scores on the belief questionnaire were signif
icantly positively related to their students' problem-solving scores. Simila
findings appeared for the interviewers' and coders' ratings of teachers on th
belief constructs. However, teachers' scores on the belief questionnaire
interviewers' ratings, and coders' ratings of teachers on the belief constructs
were unrelated to students' number-fact knowledge.
The positive correlation between teachers' cognitive beliefs and student

This content downloaded from 136.167.3.36 on Wed, 06 Sep 2017 01:39:13 UTC
All use subject to http://about.jstor.org/terms
TEACHERS' BELIEFS IN MATHEMATICS 27

TABLE 3
Correlations Between Teachers' Scores on
Belief Constructs and Students' Achievement in Addition/Subtraction

Measure 2 3 4 5 6

Total scores on belief constructs


1. Belief questionnaire. .65* .59* -.10 .32* .42*
2. Interviewer's rating. .60* -.08 .28* .37*
3. Coder's rating of interview. -.07 .33* .42*
Students' achievement score
4. Addition/subtraction number-fact
knowledge. .48* -
5. Problem solving. .88*
6. Problem solving adjusted f
knowledge.

Note. N = 37 teachers.
*p < .05, one-tailed.

problem solving may arise from a chain of causality whereby teachers'


beliefs affected students' problem solving indirectly. Thus, the effects of
teachers' beliefs on students' achievement may have been mediated through
problem-solving practice in the following way:

1. Teachers believed that their children knew how to solve problems


informally.
2. Therefore, these teachers included more problem solving in their
instructions.
3. Thus, their children did better on problem-solving tests.

Because we did not have an initial measure of students' problem solving, we


do not know how and whether teachers' beliefs that children know how to
solve problems informally may have resulted from the teachers' actually
having students who were more able problem solvers when they entered first
grade.

Comparison of student achievement of CB and LCB teachers. In


addition to examining these correlations, we used a second technique: We
compared student achievement scores from those with the 7 CB teachers
versus those with the 7 LCB teachers. Across all 39 teachers, students' mean
achievement was 11.05 (SD = 1.66) for problem solving and 11.43 (SD =
2.94) for number facts. A comparison of the mean achievement of students
in CB and LCB teachers' classes showed that these classes differed
significantly on the problem-solving test (M = 12.04, SD = 1.6
teachers; M = 10.53, SD = 1.45, for LCB teachers). Students in CB

This content downloaded from 136.167.3.36 on Wed, 06 Sep 2017 01:39:13 UTC
All use subject to http://about.jstor.org/terms
28 PETERSON, FENNEMA, CARPENTER, LOEF

teachers' classes scored significantly higher than students in LCB teachers'


classes on problem solving, t(12) = 1.81, p < .05, one-tailed. Although this
difference was small, this comparison remained statistically significant even
when students' scores were adjusted for students' number-fact knowledge,
t(12) = 2.19, p < .05, one-tailed. Interestingly, students did not differ
significantly in their mean achievement on addition/subtraction number
facts in the CB teachers' classes compared with the LCB teachers' classes (M
= 11.32, SD = 2.82, for CB teachers; M = 12.00, SD = 2.48, for LCB
teachers), t(12) = -0.48, p > .05.
Together, our correlational analyses and our comparisons of CB and
LCB teachers yielded similar findings for student achievement. Teachers
whose beliefs reflected a cognitively based perspective had teachers who
scored higher in problem solving than did teachers whose beliefs corre-
sponded to a less cognitively based perspective. On the other hand, teachers'
beliefs were unrelated to students' learning of addition/subtraction number
facts. Students in CB and LCB teachers' classes knew addition/subtraction
number facts equally well. This finding is particularly interesting in view of
the fact that LCB teachers indicated in the interview that they placed greater
emphasis on the teaching of addition/subtraction number facts than did CB
teachers. CB teachers reported that they placed least emphasis on teaching
their students addition/subtraction number facts and much greater em-
phasis on instruction aimed at facilitating students' problem solving and
understanding.
The achievement data suggest an important relationship between (a)
teachers' beliefs, teachers' reported emphases, and the use of word problems
initially and throughout the year and (b) their students' problem-solving
achievement. However, the direction of the relationship is uncertain.
Because our data are only correlational and because we did not measure
student achievement at the beginning of the academic year, we cannot infer
a causal relationship between teachers' beliefs and their students' problem-
solving achievement at the end of the year. When compared with LCB
teachers, CB teachers may have developed beliefs corresponding to a
cognitively based perspective and a related emphasis on students' word-
problem solving as a result of having students who were higher in
problem-solving ability than those students in LCB teachers' classes. In
other words, CB teachers may have developed their beliefs about problem
solving, such as using word problems as a basis for instruction, as a result
of having students who were good at problem solving from the beginning
and throughout the year. On the other hand, the analyses of the verbal
protocols of CB teachers compared with LCB teachers provide some
evidence of a causal relationship, with teachers' beliefs affecting their
selection, organization, and presentation of mathematics content and their

This content downloaded from 136.167.3.36 on Wed, 06 Sep 2017 01:39:13 UTC
All use subject to http://about.jstor.org/terms
TEACHERS' BELIEFS IN MATHEMATICS 29

pedagogical strategies, which, in turn, may have affected the nature and
quality of students' learning and achievement in addition and subtraction.

Teachers' Pedagogical Content Beliefs and


Knowledge

Teachers' pedagogical content beliefs in addition and subtraction should be


significantly related to their pedagogical content knowledge in addition and
subtraction. In our analyses of teachers' pedagogical content beliefs, for
example, we found that teachers whose beliefs were more cognitively based
reported using word problems as a basis for introducing addition and
subtraction, and they reported continued emphases on word-problem
solving throughout the year. In addition, CB teachers and LCB teachers
tended to use addition and subtraction word problems very differently in
their teaching of addition and subtraction. Thus, one might also expect CB
and LCB teachers to differ in their knowledge of addition and subtraction
word problems.
A second important finding for teachers' pedagogical content beliefs was
that, compared with all CB teachers, LCB teachers in their teaching of
addition and subtraction tended to place relatively greater emphases on
students' learning of addition and subtraction facts from the beginning of
the year and throughout the year. Findings from cognitive-science research
have shown that children use both physical modeling and counting strate-
gies to solve addition and subtraction word problems for a long time before
they are able to master number facts at a recall level. Number facts are
learned over an extended period of time, and recall of number facts and
counting are used concurrently by children.
In our analysis of teachers' pedagogical content knowledge, we found
that the teachers in our sample tended to underestimate their students' use
of counting strategies to solve addition and subtraction word problems and
to overestimate their use of recalled number facts to solve the problems
(Carpenter et al., 1988). Teachers whose content beliefs are more
cognitively based, however, might be more accurate in predicting their
students' use of counting strategies to solve addition and subtraction word
problems, because they have greater underlying knowledge that children use
counting strategies for an extended period of time before they master
number facts. In contrast, teachers whose beliefs are less cognitively based
might be less likely to know of children's persistent use of counting
strategies and might be more likely to overestimate their students' use of
number facts to solve such problems. We explored these two hypotheses
dealing with teachers' knowledge of word problems and teachers' knowl-

This content downloaded from 136.167.3.36 on Wed, 06 Sep 2017 01:39:13 UTC
All use subject to http://about.jstor.org/terms
30 PETERSON, FENNEMA, CARPENTER, LOEF

edge of students' strategies for solving word problems. Let us consider these
results in detail.

Teachers' knowledge of addition and subtraction word problems.


We found evidence indicating that teachers' pedagogical content beliefs in
addition and subtraction were significantly related to teachers' knowledge
of classes of addition and subtraction word-problem types and the exact
distinctions between different addition/subtraction word-problem types.
As reported by Carpenter et al. (1988), 18 of the teachers in our study
focused on the language in the problem or "key words" in evaluating the
relative difficulty of pairs of addition and subtraction word problems. The
following teachers' responses represented a key-word approach to analyzing
the relative difficulty of addition and subtraction word problems:

Teacher 6: [The problem] doesn't ask "How many left?", which clues
the kids into subtraction.

Teacher 64: "Have left" is easy for children to do. ... "Have left" is
easier to understand [than "How many more"].

Further analyses showed that teachers' key-word approach to evaluating


the relative difficulty of addition and subtraction word problems was
significantly related to their scores on the belief questionnaire. The teachers
who used key words to analyze problems had significantly lower scores on
the total belief construct as measured by the belief questionnaire (M =
173.82) compared with teachers who did not take a key-word approach (M
= 185.68), t(37) = 1.79, p < .05, one-tailed. Similarly, the key-word
teachers had significantly lower scores on the belief construct as measured
by interviewers' ratings of teachers (M = 9.78) when compared with
teachers who did not use a key-word approach (M = 12.86), t(38) = 2.16,
p < .05, one-tailed. Similar findings appeared when we compared the 7 CB
teachers with the 7 LCB teachers. Six of the 7 LCB teachers focused on key
words in the addition and subtraction word problems in assessing the
relative difficulty of the problems. In contrast, only 1 of the 7 CB teachers
focused on key words.
The prominence of key words in LCB teachers' analyses of addition and
subtraction word problems is consistent with the focus of several of these
teachers on key words in their description of the way they taught such
problems in their classrooms. The key-word approach stresses surface
features or words in the problem and seems to reflect pedagogical content
knowledge and pedagogical beliefs that are not consistent with a cognitively
based perspective. Researchers have found that, for children, addition and
subtraction word-problem difficulty is a function of the semantic features

This content downloaded from 136.167.3.36 on Wed, 06 Sep 2017 01:39:13 UTC
All use subject to http://about.jstor.org/terms
TEACHERS' BELIEFS IN MATHEMATICS 31

of the problem, together with processes and strategies that children


generally use to solve different problems (Carpenter & Moser, 1984; Riley
& Greeno, 1988). Analysis of word problems in terms of specific words they
use, such as altogether or more, addresses only superficial features of the
problems. It does not address the underlying processes that children need to
use in solving the problems or the strategies that tend to make the problems
more or less difficult for the child.

Knowledge of a variety of word-problem types. Teachers' pedagog-


ical content knowledge of more sophisticated word-problem types in
addition and subtraction might also be related to their underlying pedagog-
ical content beliefs. For example, researchers have distinguished between
different classes of addition and subtraction word problems based on
semantic characteristics of the problems as well as the processes that
students use to solve the problems (Carpenter & Moser, 1983; Riley &
Greeno, 1988). This analysis goes beyond a simple analysis of problems in
terms of addition or subtraction or in terms of "join" or "separate." The
analysis also goes beyond the basic categroy of "result unknown" problems,
which is the major type of word problem presented in most first-grade
mathematics textbooks. Thus, researchers have distinguished more difficult
problem classes including "start unknown," "change unknown," and "com-
pare" problems.
Comparing the protocols of the 7 CB teachers' analyses of the difficulty
of word problems with the LCB teachers' analyses, we found that signifi-
cantly more CB teachers knew and understood the more complex distinc-
tions between word-problem types than did LCB teachers. For example,
teachers were given the following two word problems to compare and
discuss in terms of relative difficulty:

Kim had 5 pencils. How many more pencils does she need to buy to
have 13 pencils altogether?

Mike had some books.


He lost 5 of them.
Now he has 8 books left.
How many books did Mike have to start with?

Six of the 7 CB teachers referred to the "start unknown" and/or "change


unknown" distinction in discussing the difficulty of the problems. Although
the interviewer here did not record the teachers' responses verbatim, the
interviewer did make notes on selected aspects of the teachers' responses as
follows:

This content downloaded from 136.167.3.36 on Wed, 06 Sep 2017 01:39:13 UTC
All use subject to http://about.jstor.org/terms
32 PETERSON, FENNEMA, CARPENTER, LOEF

Teacher 2: The second is more difficult because it is difficult to figure


out "What did Mike start with." The child would have to understand
two sets. For the second problem, you need to put both sets back
together. This takes it out of the child's "familiar experience" [with
result-unknown problems]. This requires a level above basic under-
standing of addition and subtraction.

Teacher 63: You need to begin with the unknown. Children don't see
this as an addition problem.

Teacher 12: These are "off beat" problems. The children are used to a
format in which the unknown is as follows: 5 + 8 = ?. To figure out
these two problems, they would have to understand that the unknown
in the problem is in a different place: 5 + ? = 13 for the first problem
and ? - 5 = 8 for the second problem.

In contrast, in discussing and comparing these two problems, only 1 of


the 7 LCB teachers focused on the start-unknown distinction. LCB teachers
tended to focus on other aspects of the problem, such as that the first
problem was more difficult because it was subtraction, whereas the second
problem was easier, because it involved adding (LCB Teacher 64). A similar
inattention to the start-unknown distinction is shown in the following LCB
teacher's very global description of reasons for the relative difficulty of the
two problems:

Teacher 11: The second problem would be harder because it sounded


harder. Even if the problem was read over and over, it would be
harder to sort through the information. It would be harder for the
child to get this salient information.

These descriptions suggest that CB teachers, but not LCB teachers, in our
study had a principled framework for thinking about and distinguishing
between addition and subtraction word problem types.

Beliefs about use of a variety of word problem types. These


findings suggest a connection between teachers' knowledge of types of
addition and subtraction word problems and their underlying beliefs and
use of word problems in their classrooms. In the belief interview, teachers
were asked, "Are there certain kinds of word problems in addition that you
believe all children should learn to solve? If so, what are they?" In their
responses, 6 of the 7 LCB teachers indicated that the only kinds of word
problems that they focused on and used in their teaching were addition and
subtraction problems with the result unknown. Thus, they did not use or

This content downloaded from 136.167.3.36 on Wed, 06 Sep 2017 01:39:13 UTC
All use subject to http://about.jstor.org/terms
TEACHERS' BELIEFS IN MATHEMATICS 33

introduce more sophisticated or complex problems such as start-unknown


or change-unknown problems. The following exemplify the LCB teachers'
responses:

Teacher 54: Yes, the simple word problems, you know if you want to
go buy something in the store, if you want to like buy a piece of candy
or pencil or something, you can add the two together and then you
know if you have enough money. Just simple things like that.

Teacher 6: I guess just finding the answer. You know, they're given
two numbers, and they have to figure it out if it's "in all" or if it's "less
than."

Teacher 11: I guess the one that I'm talking about: That John had 3
cards and Joe had 4, and how many cards did they have altogether? I
teach them the most common addition story problem.

In contrast, the 7 CB teachers' responses indicated knowledge and use of


a variety of addition/subtraction word-problem types in their teaching, as
well as an acknowledgment of the importance of presenting a variety of
problem types. The following CB teacher's response illustrates this perspec-
tive:

Teacher 16: I try to do a big variety, yes, so it isn't always "how much
is it altogether?" When it's like, "We should have four in our group
and three are here. How many more do we need to make our complete
group?" That kind . . . trying to get away from just "How many
altogether?"

Teachers' knowledge of students' strategies for solving word


problems. In our analyses, we explored whether teachers whose content
beliefs were more cognitively based might be more knowledgeable about
children's strategies for solving addition and subtraction word problems.
Teachers' general knowledge of children's strategies was significantly posi-
tively related to teachers' scores on the total belief construct as measured by
the questionnaire (r = .28), the interviewers' ratings (r = .25), and the
coders' ratings (r = .26). Furthermore, although the difference was small,
CB teachers had significantly greater knowledge of children's strategies (M
= 12.86, SD = 0.90) than did LCB teachers (M= 11.71, SD = 1.11), t(12)
= 2.11, p < .05. Teachers who were more cognitively based in their beliefs,
however, were not necessarily better able to predict their own students' use
of particular strategies compared with teachers who were less cognitively
oriented in their beliefs (e.g., r = .13 between knowledge of individual

This content downloaded from 136.167.3.36 on Wed, 06 Sep 2017 01:39:13 UTC
All use subject to http://about.jstor.org/terms
34 PETERSON, FENNEMA, CARPENTER, LOEF

students' strategies and total belief construct on questionnaire). We also


found no significant relationship between teachers' content beliefs and their
ability to predict their own students' success in solving word problems (r =
.18 for total belief construct on questionnaire; r = .11 for interviewers'
ratings; r = .10 for coders' ratings). To explore further the possible
relationship between teachers' general knowledge of children's strategies
and their content beliefs, we used the data from the "knowledge of
individual students' strategies" measure to construct an additional score
representing the teachers' ability to accurately predict across a group of 6
children the number of times those children actually used those strategies.
In the interview on knowledge of individual students' strategies, teachers
were asked to predict for each of 6 students in their class the specific
strategies that the student would use to solve each of five word problems.
For each of the 40 teachers in the study, we computed a score for the total
number of times the teacher expected those students in her class to use a
given type of strategy minus the number of times the students actually used
that type of strategy to solve the problem. We then computed the absolute
value of this score for each teacher. This resulting score was similar to the
"general knowledge of strategies" score, because it examined the accuracy
of teachers' strategy predictions across the 6 children in her class. We then
correlated this score with teachers' scores on the belief construct from the
belief questionnaire. We found that teachers who were more cognitively
based in their beliefs were somewhat better predictors of their students' use
of solution strategies than were teachers who were less cognitively based in
their beliefs (r = -.24).
To further explore this finding, we compared the CB teachers with the
LCB teachers. We examined individual teachers' accuracy of predictions for
students' use of advanced counting strategies. Advanced counting strategies
included "counting on," using a forward counting sequence, or using a
backward counting sequence, but did not include "counting all." We found
evidence that CB teachers were more accurate in their prediction of
advanced counting strategies than were LCB teachers. For example, in
predicting the use of advanced counting strategies to solve the three word
problems in which students could make use of counters, 3 of the 7 CB
teachers accurately predicted the exact number of advanced counting
strategies that their students would use. A 4th CB teacher underestimated
use by only one. In contrast, only one LCB teacher was completely accurate
in predicting her students' actual use of advanced counting strategies (not
including "counting all"). The remaining LCB teachers all underestimated
students' use of advanced counting strategies. The underestimates of LCB
teachers ranged from a minimum of 2 to a maximum of 14 uses of advanced
counting strategies by students. The students in Teacher 17's class, for

This content downloaded from 136.167.3.36 on Wed, 06 Sep 2017 01:39:13 UTC
All use subject to http://about.jstor.org/terms
TEACHERS' BELIEFS IN MATHEMATICS 35

example, actually used advanced counting strategies 14 times more than this
LCB teacher estimated they would.
After predicting the strategies that the 6 individual target students in their
class would use, teachers were asked to identify additional strategies for the
class as a whole that students might use to solve word problems. We
compared CB with LCB teachers on the total number of advanced counti
strategies that were identified either for the 6 target students in their class o
for additional class members. Again, the 7 CB teachers correctly identifi
significantly more advanced counting strategies (M = 3.14) than did al
LCB teachers (M = 1.86), t(12) = 2.25, p < .05. In contrast, CB and LCB
teachers did not differ in their reported knowledge of students' use of direc
modeling strategies to solve addition and subtraction word problems.
Together, the findings for direct modeling and advanced counting
strategies seemed to indicate that, although both LCB and CB teachers
knew of students' use of direct modeling strategies, CB teachers were more
aware of children's extensive use of counting strategies. Thus, both grou
of teachers understood the need to provide an opportunity for students
directly model addition word problems using counters in the introducto
lessons. In contrast to CB teachers, however, LCB teachers moved more
quickly in their teaching from an emphasis on direct modeling of the word
problem to an emphasis on students' use of number facts, because the
lacked knowledge of children's persistent use of counting strategies eve
while they are developing number-fact skills.

Teachers' techniques for assessing students' abilities. Why did


CB teachers have more accurate knowledge than LCB teachers of their
students' use of strategies for solving addition and subtraction word
problems and how did CB teachers obtain this knowledge? To answer this
question, we obtained some information from teachers' responses to
interview questions on how teachers found out about differences in their
children's abilities and knowledge about addition. The CB teachers re-
sponded uniformly that they found out about differences in a child's
abilities by giving him or her a problem and then observing and listening to
him or her work the problem. In contrast, CB teachers never mentioned
using tests, records, or written seatwork, whereas 4 of the 7 LCB teachers
mentioned using these more "formal" techniques. Two LCB teachers
indicated that they obtained information on children's abilities incidentally
(e.g., "If you are just looking, the things become very obvious"). Interest-
ingly, the 7th CB teacher (Teacher 66) used observations, but she focused
on "how quickly they [the students] respond to a question you've asked,"
rather than on the problem-solving process.
These data suggest that CB teachers' assessment strategies were signifi-

This content downloaded from 136.167.3.36 on Wed, 06 Sep 2017 01:39:13 UTC
All use subject to http://about.jstor.org/terms
36 PETERSON, FENNEMA, CARPENTER, LOEF

cantly related to their knowledge of their children's problem-solving


strategies. CB teachers seemed to be more knowledgeable about their
children's problem-solving strategies because they actively assessed their
students' thinking strategies-by putting children in problem-solving situa-
tions and then observing and listening to them. Thus, their assessment
techniques were more clinical and adapted to "getting inside children's
heads" by diagnosing students' thought processes than were the assessment
techniques of the LCB teachers. LCB teachers either did not assess
systematically or relied on formal assessment techniques that would give
them information only on whether their students could obtain the correct
answer to a problem and not on how they solved the problem.

Summary. These data suggest that CB teachers had a greater under-


lying knowledge than LCB teachers that children use counting strategies for
an extended period of time before they master (and while they master)
number facts. Furthermore, in contrast to LCB teachers, CB teachers may
have developed greater knowledge of their students' use of strategies
because of their techniques for assessing and diagnosing differences in
children's abilities in addition. Rather than relying only on tests of students'
mathematical knowledge and abilities, CB teachers made their own obser-
vations of students in problem-solving situations to find out what their
students knew about addition and subtraction. Based on teachers' reported
descriptions of their introductory and typical lessons in addition, we found
that CB and LCB teachers differed in the amount of time they devoted to
children's counting strategies before teaching number-fact symbolism. CB
teachers reported spending significantly more time on development and
facilitation of children's counting strategies. Taken together, teachers'
knowledge of individual students' use of advanced counting strategies and
teachers' reports of how they developed and worked on advanced counting
strategies in their classroom suggested strong connections among teachers'
pedagogical content knowledge in addition and subtraction, teachers'
pedagogical content beliefs, their teaching strategies and presentation of
content in addition and subtraction, and students' learning of addition and
subtraction.

CONCLUSIONS

Although this study represents a beginning attempt to examine te


pedagogical content beliefs in mathematics, the findings are provo
They suggest that teachers' pedagogical content beliefs and teach
pedagogical knowledge may be importantly linked to teachers' clas
actions and, ultimately, to students' classroom learning in mathe

This content downloaded from 136.167.3.36 on Wed, 06 Sep 2017 01:39:13 UTC
All use subject to http://about.jstor.org/terms
TEACHERS' BELIEFS IN MATHEMATICS 37

Although our study focused on teachers' pedagogical content beliefs in the


specific content area of first-grade addition and subtraction, the theoretical
constructs and methodology used here apply to other specific content areas
in mathematics that may also be analyzed from the perspective of recent
cognitive theory and research in children's mathematics learning. From our
research, we draw five important conclusions.
First, teachers vary widely in their pedagogical content beliefs about
mathematics, curriculum, and instruction in addition and subtraction.
Second, first-grade teachers differ significantly in the extent to which
their pedagogical content beliefs about mathematics, curriculum, and
instruction in addition and subtraction correspond to a cognitively based
perspective. Moreover, these differences are reflected in their pedagogical
content knowledge about specific mathematics content, such as addition
and subtraction word problems, as well as in their knowledge of their
students' individual strategies to solve word problems. Furthermore, these
differences are reflected in the kind and quality of content and strategies
that teachers report selecting to teach addition and subtraction and in their
goals for instruction in addition and subtraction. The extent to which
teachers' beliefs correspond to a cognitive-science perspective also seems to
be reflected in the content and type of their students' learning of addition
and subtraction.
Third, teachers' pedagogical content beliefs in addition and subtraction
show up in self-reports of their approaches to teaching addition and
subtraction, their choices of mathematics content, their conceptions of the
roles of teacher and learner, and their goals for instruction. Specifically,
when compared with a teacher subgroup having a less cognitively based
perspective (LCB teachers), teachers with a cognitively based perspective
(CB teachers) reported making extensive use of word problems in intro-
ducing addition and subtraction to children and in teaching addition and
subtraction throughout the year. In their introductory and typical lessons in
addition and subtraction, CB teachers also reported spending a lot of time
developing students' use of counting strategies to solve addition and
subtraction word problems before introducing and teaching number facts.
Consistent with this notion, CB teachers placed least emphasis on teaching
number facts and most emphasis on mathematical understanding and
problem solving. LCB teachers placed most emphasis on number facts and
understanding and least emphasis on problem solving. Moreover, CB
teachers view the teacher and learner as actively engaged with one another
in construction of mathematical knowledge and understanding. Thus, the
teacher's role is one of facilitating the construction of student under-
standing and knowledge, and the student's role is one of engagement in
active cognitive learning.
Fourth, teachers' pedagogical content beliefs in addition and subtraction

This content downloaded from 136.167.3.36 on Wed, 06 Sep 2017 01:39:13 UTC
All use subject to http://about.jstor.org/terms
38 PETERSON, FENNEMA, CARPENTER, LOEF

are related to students' mathematics learning and achievement, specifically,


achievement of problem-solving abilities in addition and subtraction. CB
teachers tended to have children who were higher in word problem-solving
ability in addition and subtraction than did LCB teachers. However,
teachers' pedagogical content beliefs were unrelated to students' achieve-
ments of computational skills (number facts) in addition and subtraction.
These differences in problem-solving achievement seemed to be related to
CB teachers' emphases on word problems as a basis for introducing
addition and subtraction and to their reported emphasis on solving word-
problems throughout the year. Interestingly, when compared with LCB
teachers, CB teachers reported placing least emphasis on students' learning
of addition and subtraction number facts and relatively greater emphasis on
students' word-problem solving and understanding. Although the findings
are only correlational, the finding of no difference in number-fact knowl-
edge between CB and LCB teachers suggests the possibility that students in
CB teachers' classes are able to master computational skills and number-
fact knowledge concurrently with their development of word problem-
solving skills. Moreover, their students master number facts as well as the
students in the LCB teachers' classes, where greater emphasis was placed on
mastering computational skills than on problem solving.
Finally, teachers' pedagogical content beliefs and their pedagogical
content knowledge seem to be interrelated. In the future, researchers need
to conceptualize further teachers' pedagogical content knowledge and
beliefs and to gather empirical data to explore the interconnections between
teachers' pedagogical content beliefs and pedagogical content knowledge.
Further conceptualization and additional empirical research should lead
researchers to a greater understanding of how teachers' pedagogical content
beliefs underlie their classroom actions and how important differences
between teachers in their pedagogical content beliefs affect teachers'
educational practice and, ultimately, influence students' learning and
achievement.

ACKNOWLEDGM ENTS

The research reported in this article was funded in part by National Sci
Foundation Grant MDR-8550236). The opinions expressed herein do
necessarily reflect the position, policy, or endorsement of the Na
Science Foundation.
This article was presented at the meeting of the American Educational
Research Association, Washington, DC, April 1987.

This content downloaded from 136.167.3.36 on Wed, 06 Sep 2017 01:39:13 UTC
All use subject to http://about.jstor.org/terms
TEACHERS' BELIEFS IN MATHEMATICS 39

REFERENCES

Bussis, A. M., Chittenden, F., & Amarel, M. (1976). Beyond surface curriculum.
CO: Westview.
Carpenter, T. P., Fennema, E., Peterson, P. L., & Carey, D. (1988). Teachers' pedagogical
content knowledge in mathematics. Journal for Research in Mathematics Education, 19,
385-401.

Carpenter, T. P., & Moser, J. M. (1983). The acquisition of addition and subtraction concep
In R. Lesh & M. Landau (Eds.), The acquisition of mathematical concepts and process
(pp. 7-14). New York: Academic.
Carpenter, T. P., & Moser, J. M. (1984). The acquisition of addition and subtraction concep
in grades one through three. Journalfor Research in Mathematics Education, 15, 179-20
Carpenter, T. P., & Peterson, P. L. (Eds.). (1988). Learning mathematics through instructio
[Special issue]. Educational Psychologist, 23(2).
Case, R. (1983). Intellectual development: A systematic reinterpretation. New York: A
demic.
Case, R., & Bereiter, C. (1982). From behaviorism to cognitive behaviorism: Steps in the
evolution of instructional design. Paper presented at the Conference for Educational
Psychologists, Caracus, Venezuela.
Clark, C. M., & Peterson, P. L. (1986). Teachers' thought processes. In M. C. Wittrock (Ed.),
Handbook of research on teaching (3rd ed., pp. 255-296). New York: Macmillan.
Cobb, P. (1988). The tension between theories of learning and theories of instruction in
mathematics education. Educational Psychologist, 23(2), 87-104.
Collier, C. P. (1972). Three studies of teacher planning (Research Series No. 55). East Lansing:
Michigan State University.
Confrey, J. (1986). A critique of teacher effectiveness research in mathematics education.
Journal of Research in Mathematics Education, 17, 347-360.
Conners, R. D. (1978). An analysis of teacher thought processes, beliefs, and principles during
instruction. Unpublished doctoral dissertation, University of Alberta, Edmonton, Canada.
Janesick, V. (1977). An ethnographic study of a teacher's classroom perspective. Unpublished
doctoral dissertation, Michigan State University, East Lansing.
Marland, P. W. (1977). A study of teachers' interactive thoughts. Unpublished doctoral
dissertation, University of Alberta, Edmonton, Canada.
Munby, H. (1983, April). A qualitative study of teachers' beliefs and principles. Paper
presented at the meeting of the American Educational Research Association, Montreal.
Meyerson, L. N. (1977). Conceptions of knowledge in mathematics: Interactions with and
applications to a teaching methods course. Unpublished doctoral dissertation, State
University of New York at Buffalo.
Newell, A., & Simon, H. A. (1972). Human problem solving. Englewood Cliffs, NJ:
Prentice-Hall.
Nie, N. H., Hull, C. H., Jenkins, J. G., Steinbrenner, K., & Bent, D. H. (1975). Statistical
package for the social sciences. New York: McGraw-Hill.
Peterson, P. L. (1988). Teachers' and students' cognitional knowledge for classroom teaching
and learning. Educational Researcher, 17(5), 5-14.
Resnick, L. B. (1981). Instructional psychology. Annual Review of Psychology, 32, 659-704.
Resnick, L. B., & Ford, W. W. (1981). The psychology of mathematics for instruction.
Hillsdale, NJ: Lawrence Erlbaum Associates, Inc.
Riley, M. S., & Greeno, J. G. (1988). Developmental analysis of understanding language about
quantities and of solving problems. Cognition and Instruction, 5, 49-101.
Romberg, T. A., & Carpenter, T. P. (1986). Research on teaching and learning mathematics:

This content downloaded from 136.167.3.36 on Wed, 06 Sep 2017 01:39:13 UTC
All use subject to http://about.jstor.org/terms
40 PETERSON, FENNEMA, CARPENTER, LOEF

Two disciplines of scientific inquiry. In M. Wittrock (Ed.), Handbook of research on


teaching (3rd ed., pp. 850-873). New York: Macmillan.
Shavelson, R. J., & Stern, P. (1981). Research on teachers' pedagogical thoughts, judgments,
decisions, and behavior. Review of Educational Research, 51, 455-498.
Shirk, G. B. (1972). An examination of conceptual frameworks of beginning mathematics
teachers. Unpublished doctoral dissertation, University of Illinois, Urbana-Champaign.
Shulman, L. S. (1986). Those who understand: Knowledge growth in teaching. Educational
Researcher, 15(2), 4-14.
Shulman, L. S. (1987). Knowledge and teaching: Foundations of the new reform. Harvard
Educational Review, 51, 1-22.
Shulman, L. S., & Elstein, A. S. (1975). Studies of problem-solving, judgment and decision-
making: Implications for educational research. In F. N. Kerlinger (Ed.), Review of research
in education (Vol. 3, pp. 3-42). Itasca, IL: Peacock.
Skemp, R. R. (1978). Relational understanding and instrumental understanding. Arithmetic
Teacher, 26, 9-15.
Thompson, A. G. (1982). Teachers' conceptions of mathematics and mathematics teaching:
Three case studies. Unpublished doctoral dissertation, University of Georgia, Athens.

This content downloaded from 136.167.3.36 on Wed, 06 Sep 2017 01:39:13 UTC
All use subject to http://about.jstor.org/terms

S-ar putea să vă placă și