SOLVING PROBLEMS IN DIFFERENT WAYS: TEACHERS' KNOWLEDGE SITUATED IN PRACTICE

Anat Levav-Waynberg and Roza Leikin University of Haifa, Israel This study is part of a larger longitudinal research concerned with teachers' knowledge development in the context of connecting tasks. We find this study important in the view of the existing gap between practice and theory of employing connecting tasks in school mathematics. In this paper we focus on the development of teachers' problem-solving performance as a result of systematic learning and their teaching practice associated with connecting tasks. This study is aimed to deepen our understanding of the role of both systematic (though learning) and craft (through teaching) teachers' knowledge development. We argue for the necessity of combining systematic and craft modes of development in teacher education. THEORETICAL BACKGROUND Connecting tasks: The gap between theory and practice In this study a task is called a multiple-solution connecting task (connecting task for short) when it can be solved in different ways using: (a) different definitions or representations of a mathematical concept; (b) hierarchy, which is expressed in seeing an idea as a special case of a more general idea; (c) different mathematics tools and theorems from a particular mathematical topic; and (d) different mathematics tools and theorems from different branches of mathematics (Leikin Levav-Waynberg, Gurevich & Mednikov, 2006). Figure 1 presents an example of a connecting task in our study. The rational for the implementation of connecting tasks in school mathematics is rooted in the view that connections form an essential part of mathematical understanding (e.g., Skemp, 1987; Hiebert & Carpenter, 1992). One of the well recognized ways for developing connectedness of ones mathematical knowledge is solving problems in different ways (e.g., House & Coxford, 1995; NCTM, 2000). Stigler and Hiebert (1999), showed that multiple solutions to problems increased the quality of mathematical lessons. Consequently, we consider integrating connecting tasks in ones teaching practice critical for fostering the connectedness of students mathematical knowledge both by teaching students multiple solutions to problems and recognizing that in a class of students there are multiple ways in which pupils do solve problems. Despite the importance of implementing connecting tasks as stressed in the research literature, teachers seldom solve problems in different ways either for themselves or in their classes. Stigler and Hiebert (1999) as well as Ma (1999) showed that in US classrooms teachers rarely introduce their students to multiple-solution tasks. Our study has shown that Israeli teachers as well rarely employ connecting tasks
2006. In Novotn, J., Moraov, H., Krtk, M. & Stehlkov, N. (Eds.). Proceedings 30th Conference of the 4 - 57 International Group for the Psychology of Mathematics Education, Vol. 4, pp. 57-64. Prague: PME.

Levav-Waynberg & Leikin systematically in their classes (Leikin et al., 2006). We conjecture that this gap between theory and practice turns connecting tasks into the powerful environment for the development of teachers' knowledge.
Quadrilateral ABCD is inscribed in a circle. The diagonal AC is the diameter of the circle. The angle A is 60 degrees. We mark angle BAC as . Which values of give the maximal area of quadrilateral ABCD? S ABCD = S ABC + S ADC = R (h + h ) SABCD is maximal when h1+h2 is maximal B 1 2 Solution 1: Calculus-based solution of the minima-maxima problem:
R

A
60o-

2 120o-2 R

h1

C
h2

h1 + h2 = R 3 cos ( 2),0 < < ; (h1 + h2 )' = 0 sin( 2) = 0 = 3 3 3 6

Solution 2: Properties of trigonometric functions cos (60o 2) =1 is the maximal value of the function = 30o

Solution 3: Geometric solution (circle properties) Segment BD is constant (cord BD is 120o) for all values of . Thus the two diagonals of the quadrilateral are constant. The area of such a quadrilateral is maximal when the diagonals are perpendicular: when the AC bisects angle BAD. Thus = 30 o Solution 4: Symmetry considerations Segment BD is constant. In a symmetric situation, BD is perpendicular to AC and the area of the quadrilateral equals 1/2BDAC. If we break the symmetrical situation we decrease the sum of the heights (h1+h2) and therefore reduce the area of the quadrilateral. Thus = 30 o .

Figure 1: Example of the task in the study Teachers' knowledge In our study we address epistemological complexity of teacher's knowledge by integrating two well recognized theoretical perspectives on teachers knowledge (e.g., Shulman, 1986; Kennedy, 2002; for elaboration of the model see Leikin, 2006). We briefly explain two dimensions of teachers' knowledge in the context of connecting tasks. Dimension 1 'KINDS OF TEACHERS' KNOWLEDGE' is based on Shulman's (1986) components of knowledge: Teachers subject-matter knowledge (SMK) comprises their own knowledge of mathematical connections of different types, their ability to solve problems in multiple ways and to hold a rich collection of examples of connecting tasks. Along with Ma's (1999) definition of profound understanding of mathematics we consider problem solving in different ways as an integral part of teachers' subject matter knowledge. Teachers pedagogical content knowledge (PCK) includes knowledge of how students cope with connecting tasks, as well as knowledge of appropriate learning setting. Teachers curricular content knowledge includes knowledge of different types of curricula, connections between different curricular topics and understanding different approaches to teaching connecting tasks. Dimension 2 'SOURCES OF TEACHERS' KNOWLEDGE' is based on Kennedy's (2002) classification of teachers' knowledge according to the sources of its development: Teachers' craft knowledge related to connecting tasks is largely
4 - 58 PME30 2006

Levav-Waynberg & Leikin developed through classroom experiences with connecting tasks. Systematic knowledge is acquired mainly through systematic studies of mathematics and pedagogy in colleges and universities. Prescriptive knowledge is acquired through institutional policies, which are transparent in tests, accountability systems, and texts of diverse nature. Along with Kennedy's definitions we incorporate in our study two modes of teacher knowledge development: the systematic (through learning) mode and the craft (through teaching) mode. At the same time we acknowledge the irreducibility of the prescriptive sources into the process of knowledge development THE STUDY Our longitudinal study is based on Teacher Development Experiment (TDE: Simon, 2002) using interview and observation research methods. It is aimed at developing a model of Teachers' Knowledge Development (TKD) which describes and characterizes development of teachers' SMK and PCK in systematic and craft modes. For the analysis of the development of teacher systematic knowledge, during the first year of the study, we observed 12 secondary school mathematics teachers who volunteered to take part in a 56-hour professional development course focusing connecting tasks (Course A). Ten of them were interviewed before course (int-A) and all 12 teachers were interviewed at the end of the course (int-B). In order to analyze the development of teachers' craft knowledge the teachers were asked to teach connecting tasks in their classes during the second year of the intervention. Seven of the 12 teachers fulfilled this requirement and six of them further participated in the whole-group discussions focusing on teaching connecting tasks (Course B). Nine teachers were interviewed at the end of the second year of the research intervention (int-C). We report here data regarding five teachers who consistently took part in all the stages of the study. This sample is a representative of the whole group of teachers from the perspective of the teachers' educational background and teaching experience. This paper is focused on the question: How does teachers' problem-solving performance change (a) in systematic (through learning) mode and (b) in craft (through teaching) mode? We demonstrate that teachers' problem-solving performance is situated in their practice of different kinds (Lave, 1996). The mathematical problems in the interview: In order to answer research questions presented in this paper we asked teachers to solve problems in different ways. The problems in all three interviews were chosen based on the following considerations: (1) To explore teachers' subject matter knowledge problems should belong to different topics and include different types of connections. (2) To explore the sources of teachers' knowledge, some of the problems we used were solved in different ways in the textbooks (e.g., #1, Figure 2), others were not (##2, 3, 4, Figure 2). (3) To address the teachers curricular knowledge, some solutions to the problems belonged to the school curriculum and are often presented in the textbooks (e.g., prescribed solutions in Figure 2), whereas other solutions are rarely found in school textbooks (non-prescribed solutions in Figure 2).
PME30 2006 4 - 59

Levav-Waynberg & Leikin

# Problems Types of solutions: Connections Connecting task with Curricular-prescribed Multiple Solutions (CMS tasks)
1 Solve the system of equations:
2 x + y = 6 x + 2 y = 6

Prescribed solutions: Substitution, linear combination, graphing. Non-prescribed solutions: Symmetry, matrices

Different representations, techniques, topics in algebra, fields in mathematics

Connecting tasks with no curricular-prescribed (Unconventional) Multiple Solutions (UMS tasks)

2 Max-Min Problem (see Problem 1 Figure 1) Prescribed solution: Solution 1, Figure 1 Non-prescribed solutions: Solutions 2, 3, 4 Figure 1 Prescribed solution: Algebraic solution Non-prescribed solutions: Logical solution , Pictorial solution (1D), Graphic solution (2D), Area-based solution Prescribed solution: Midline in a trapezoid theorem and right-angle isosceles triangle Non-prescribed solution: Four midlines in the quadrilateral theorem and diagonals in a square Different fields in mathematics, different properties of mathematical objects. Different representations of the same concept

3 Word (motion) problem Dan and Moshe walk from the train station to the hotel. They start out at the same time. Dan walks half the time at speed v1and half the time at speed v2. Moshe walks half way at speed v1 and half way at speed v2. Who gets to the hotel first: Dan or Moshe? 4 Geometry problem In an isosceles trapezoid ABCD the diagonals are perpendicular. Prove that the height of the trapezoid equals its midline.

Different concepts, theorems and definitions

Figure 2: Examples of mathematical problems used in the interviews As mentioned earlier, to simplify the presentation of the results we show only the data related to 5 teachers who equally participated in all the stages of the intervention. Our report focuses on 4 types of tasks, which were included in all the three interviews (see Figure 2). RESULTS Knowledge situated in practice Table 1 summarizes teachers' problem-solving performance with different types of tasks in the three interviews. Curricular orientation of teachers' SMK Data from int-A show that teachers' reasoning was pretty much curricular-prescribed. We draw this conclusion from the analysis of the number and the nature of the solutions the teachers gave for the problems of different kinds in int-A. For the system of equations (CMS task, Figure 2) most of the teachers gave three different solutions all of which were curriculum-prescribed. In contrast, for UMS tasks in most of the cases the teachers suggested only one solution, which matched the place in the
4 - 60 PME30 2006

Levav-Waynberg & Leikin textbook where the task appeared. Clearly the teachers had difficulties in suggesting multiple solutions for UMS tasks. For example, T2 expressed this difficulty in thinking about a different solution when solving a minima-maxima problem (presented in Figure 1).
Interviewer: [After presentation of calculus-based solution] Can you, please, think of another solution?

T2: I dont know, because of the word "maximal". I keep thinking about the derivative. I have never thought about this type of problems from a different perspective. Table 1: Teachers' problem-solving performance on Connecting tasks
Interview
Total no of different solution given by Average no of solutions per teacher

Type of task
CMS

The task

T1

T2

T3

T4

T5

System of linear equations Min-Max problem UMS Word (motion) problem Geometry Average no of solutions per problem CMS Simple quadratic inequality Min-Max problem UMS Word (motion) problem Geometry Average no of solutions per problem CMS Absolute value inequality Min-Max problem UMS Word (work) problem Geometry Average no of solutions per problem

3 2* 1 2 2 6** 3** 3** 2 3.5 3 3 1 3 2.5

2 1 1 1 1.25 3 3** 2 2 2.5 3 5* 3 3 3.5

3 2* 1 1 1.5 6** 3** 2 2 3.25 3 4** 2 4 3.25

3 1 2* 0 1.5 3 2* 3** 1 2.25 3 2* 3* 3 2.75

3 0 0 0 0.75 3 2** 4* 2 2.75 2 0 1 2 1.25

2.8 1.2 1 0.8 1.45 4.2 2.6 2.8 1.8 2.85 2.8 2.8 2 3 2.65

5 2 2 2 7 4 5 3 5 6 7 5

* - non-prescribed solution the teacher taught tasks of this kind during the 2nd year of experiment according to the curriculum the tasks of this kind was incorporated in teaching as multiple-solution connecting tasks during the 2nd year of experiment

Another evidence for the curricular-prescribed nature of teachers' knowledge in the field of connecting tasks may be seen in T3's reply to the maxima-minima problem (see Figure 1). After suggesting calculus-based solution for this problem, when stimulated by the interviewer, this teacher intuitively suggested symmetry-based solution. However, she considered this solution mathematically insufficient. In her opinion it was a tool for the development of [students] intuition rather than formal mathematical proof:
T3: If I increased till it's 60 degrees then I lose the quadrilateral and get a triangle. One triangle of 30, 60, 90 so the area of the reduced quadrilateral is R 2 3 . It's the same situation if I decrease to zero. But if I keep the "intermediate" situation I get two triangles with the area of R 2 3 . Intuitively it's the best situation Many times you see something intuitively but you can't prove it, but then you are in a much better position than someone who can't see it intuitively If you succeed in proving what you are seeing, you prove the problem. PME30 2006 4 - 61

Levav-Waynberg & Leikin Two other teachers who suggested non-prescribed solutions (T1 and T4, Table 1: intA) clearly indicated systematic sources for these solutions: T1 suggested symmetrybased solution for maxima-minima problem as learned in a teacher-development workshop while T4 suggested graphical solution for a word problem as studied in the teacher certificate program she had recently completed. Nonetheless, in general, the data from int-A demonstrates that teachers had hardly met different approaches to solving problems in systematic framework before their participation in our study. Knowledge situated in problem-solving practice Analysis of teachers' learning in systematic mode was performed by comparison between teachers' solutions in int-A and int-B as well as the analysis of teachers' problem-solving discourse during the course. We found that (rather naturally) the teachers significantly improved their performance on connecting tasks through their participation in course A: The number of solutions they gave to all the problems as well as the number of non-prescribed solutions they suggested increased. This is evident from the study of each teachers and the whole groups performance (see Table 1). On the average there were twice as many solutions per problem per teacher provided in int-B as solutions provided in int-A (2.85 vs. 1.45). For the CMS task in int-B the teachers provided 7 different solutions, some of which non-prescribed (T1 and T3). We observed changes in teachers' reaction to the request to find multiple solutions: the replies became more fluent and positive. This change is evident in T2's reply when solving max-min problem during int-B. In contrast to her response in intA (as shown earlier) after the course she enjoyed having several ideas for approaching a problem of this kind:
T2: It's not the only way [using derivative] we may build a table and show the students that moving the X from 1 to 5 increases the area and then it starts to decrease. This is a quadratic function, so the principle of continuity should work here. We can also find the minimal point of a parabola without using the derivative If x=0 then we get 100, and of x=10 then we also get 100. It gets the largest value in the two edges. It is not just the largest but equal. This is parabola. Then because of the symmetry axis it has to be in the middle: 5. Oh it's beautiful!

We argue that through their participation in the course the teachers developed both their "feeling of different solutions" and positive position with respect to this kind of mathematical activities, they became more creative and confident. After participation in our course all the teachers provided multiple solutions for all the problems (except T4 for geometry problem) and each teacher suggested non-prescribed solutions for at least two problems in the interview. Knowledge situated in teaching The number of solutions suggested per problem by each teacher changed from average of 4.2 in int-B to 2.8 in int-C, and the number of non-curricular solutions also decreased. We found clear relationship between the teachers' problem-solving performance in int-C and the topics they taught during the period of time between interviews B and C in general and the incorporation of the connecting tasks in their lessons in particular.
4 - 62 PME30 2006

Levav-Waynberg & Leikin All the teachers taught CMS task (Table 1: int. C, absolute value inequality) between the int-B and int-C. We learned that when teachers taught this task they used curricular-prescribed solutions only. This practice was reflected in teachers solutions during the interviews: all the teachers suggested curricular prescribed multiple solutions for the CMS task. Analysis of teachers' problem solving performance on UMS tasks revealed several phenomena: First we saw that teachers improved their results from int-B to int-C on the tasks that were incorporated in teaching as multiplesolution connecting tasks (e.g., Table 1: T2 and T3 on min-max problem; T1 on geometric problem). Moreover, non-prescribed solutions suggested by the teachers during int-C appeared for this kind of task in most cases. Second, teachers at least maintained their previous success on UMS tasks, which they did teach without special attention to multiple solutions (e.g., Table 1: T1 on min-max problem, T2 and T3 for word problem). Teachers' problem solving performance on UMS tasks that belonged to the topics that teachers did not teach during the year in many cases wasnt as good as it appeared in int-B (e.g., Table 1: T1 on word problem, T5 on min-max and word problem). A specific tendency was found for geometric problems. All the teachers, except T2, taught geometry problems during the year. For all the teachers the number of solutions they suggested for geometric problems increased. We connect this phenomena with the fact that the teachers (according to their multiple reports at the end of course A and in int. B) became more attentive to students' solutions, started collecting them and allowed students "always present all the solutions they found" without saying "this is good but we do not have enough time". We hypothesize that this combination of awareness and flexibility allowed teachers learn multiple solutions in geometry from their students. DISCUSSION AND CONCLUSIONS By analyzing teachers' knowledge at the beginning of the intervention we suggest some explanations for the gap between theory and practice in the field of connecting tasks: We find teachers' mathematical and pedagogical knowledge of connecting tasks curricular oriented and prescribed. Our data demonstrates that teachers associate responsibility for the success of their students with institutional policies, which are evident in the tests system that "proscribes" implementation of connecting tasks in school. On the positive side we show that implementation of connecting tasks in systematic mode meaningfully develops teachers' SMK and their problem-solving performance on multiple-solution connecting tasks with further improvement in craft mode whereas implementation in teaching is a necessary condition for the maintenance of this development. In this way our analysis highlights the situatedness of teachers' knowledge (Lave, 1996). Based on this study results we argue that combination of the systematic and craft mode are most effective for teachers knowledge development. The development of the related instructional materials by curricular designers and mathematics educators as well as curricular changes may foster changes in teachers' disposition towards multiple-solution connecting tasks.

PME30 2006

4 - 63

Levav-Waynberg & Leikin References:

Hiebert, J. & Carpenter, T. P. (1992). Learning and Teaching with Understanding. In D. A. Grouws (Ed.): Handbook of research on mathematics teaching and learning (pp. 65-97). New York, NY: Macmillan. House, P.A. & Coxford, A. F. (1995). Connecting Mathematics across the Curriculum: 1995 Yearbook. Reston, Va.: NCTM. Kennedy, M. M. (2002). Knowledge and teaching. Teacher and Teaching: theory and practice, 8, 355-370 Lave, J. (1996). Teaching, as learning, in practice. Mind, Culture & Activity, 3 (3), 149-164. Leikin, R. (2006). Learning by teaching: The case of Sieve of Eratosthenes and one elementary school teacher. In R. Zazkis & S. Campbell (Eds.), Number Theory in Mathematics Education: Perspectives and Prospects. (pp. 115-140). Mahwah, NJ: Erlbaum. Leikin, R., Levav-Waynberg, A., Gurevich, I, & Mednikov, L. (2006). Implementation of multiple solution connecting tasks: Do students attitudes support teachers reluctance? FOCUS on Learning Problems in Mathematics, 28,1-22. Ma L. (1999). Knowing and teaching elementary mathematics: Teachers understanding of fundamental mathematics in China and the United States. Hillsdale, New Jersey: Lawrence Erlbaum Associates. National Council of Teachers of Mathematics. (NCTM) (2000). Principles and standards for school mathematics. Reston, VA: NCTM. Shulman, L. S. (1986). Those who understand: Knowing growth in teaching. Educational Researcher. 5 (2), 4-14. Simon, M. A. (2000). Research on the development of mathematics teachers: The Teacher Development Experiment. In A. E. Kelly & R. A. Lesh (Eds.), Research design in mathematics and science education. (pp. 335- 359). New Jersey: Lawrence Erlbaum Associates. Skemp, R. R. (1987). The psychology of learning mathematics. Lawrence Erlbaum Associates, Publishers: Hillsdale, NJ. Stigler J. W., & Hiebert J. (1999). The teaching gap: Best ideas from the worlds teachers for improving education in the classroom. New York: The Free Press.

4 - 64

PME30 2006