LEARNERS' DIFFICULTIES WITH QUANTITATIVE UNITS IN ALGEBRAIC WORD PROBLEMS AND THE TEACHER'S INTERPRETATION OF THOSE DIFFICULTIES Abstract This study examines 8th grade students' coordination of quantitative units arising from word problems that can be solved via a set of equations that are reducible to a single equation with a single unknown. Along with Unit-Coordination, Quantitative Unit Conservation also emerges as a necessary construct in dealing with such problems. We base our analysis within a framework of quantitative reasoning (Thompson, 1988, 1989, 1993, 1995) and a theory of children’s unitscoordination with different levels of units (Steffe, 1994) that both encompass and are extended by these two constructs. Our data consist of videotaped classroom lessons, student interviews and teacher interviews. On-going analyses of these data were conducted during the teaching sequence. A retrospective analysis, using constant comparison methodology, was then undertaken during which the classroom video, related student interviews and teacher interviews were revisited many times in order to generate a thematic analysis. Our results indicate that the identification and coordination of the units involved in the problem situation are critical aspects of quantitative reasoning and need to be emphasized in the teaching-learning process. We also concluded that unit coordination and unit conservation are cognitive prerequisites for constructing a meaningful algebraic equation when reasoning quantitatively about a situation. Introduction The research presented in this paper indicates that the ability to coordinate different units in a quantitative situation is an important skill for students to develop in order to be successful in both representing and solving algebraic word problems. Whether this be coordinating different levels of units in a whole number multiplicative situation (e.g. Steffe, 1994) or in a fraction situation (e.g. Lamon, 1994; Olive, 1999; Olive and Steffe, 2002; Steffe, 2002) or in dealing with intensive (e.g., miles per hour) as well as extensive (e.g., number of hours) quantities (Schwartz, 1988) the crucial point is to understand what is being done with the varying quantities in these situations and how the units involved can be related (Thompson, 1988, 1989, 1993, 1995). Theoretical Framework This study is part of Project CoSTAR (Coordinating Students’ and Teachers’ Algebraic Reasoning)1 that has as its main purpose the coordination of research on students’ understandings and teachers’ practices and interpretations of students’ actions relative to algebraic reasoning. This particular study is informed by recent research on students’ understanding of algebraic symbols (Kieran and Sfard, 1999), and students’ construction and coordination of quantitative units (Lamon, 1994; Olive, 1999; Smith and Thompson, (in press); Steffe, 1994, 2002; Thompson, 1988, 1989, 1993, 1995). The main theoretical framework for this study is based on the construct of quantitative reasoning. Thompson (1995) states that “Quantitative reasoning is not reasoning about numbers, it is about reasoning about objects and their measurements (i.e., quantities) and relationships among quantities” (p. 206). Thompson (1988) defined a quantity as follows: “A quantity is a measurable quality of something. A magnitude of a quantity is the quantity’s measure in some unit.” The important distinction that Thompson makes in this definition is between the quality of something and the magnitude of that same thing. A quantity has to be both named (by its quality) and measured by some identified unit. We concur with Thompson on this necessity for forming or identifying a unit by which the named quality may be measured or quantified. We need to delve further into the nature of these quantities in order to uncover the units associated with them. An ordering of the form (name, unit) is helpful for the sake of proper coordination. For instance, the coordination (dime, number of dimes) is not the same as (dime, value of a dime) or (dimes, value per dime). Smith and Thompson (in press) formalize this “name—unit” coordination in their representation of quantitative relationships throughout their article. They “use ovals to represent quantities and place inside those ovals all relevant information about those quantities—their ‘name,’ their units of measure, and any numerical value or expression that is given or can be inferred.” (footnote 4 on p. 19) Schwartz (1988) used the term referent in a way similar to how we are using identified unit of measure and called such quantities adjectival quantities. (p. 41) He stated that all quantities have referents and that the “composing of two mathematical quantities to yield a third derived quantity can take either of two forms, referent preserving composition or referent transforming composition.” (p. 41) The referent transforming composition, Schwartz claims, forces us to distinguish between two different kinds of quantity: extensive quantity and intensive quantity. An extensive quantity can be counted or measured directly, whereas an intensive quantity is derived from the multiplicative combination of two like or unlike quantities, and is usually recognized by the use of “per” in its referent unit (e.g., miles per hour, price per pound). Schwartz (1988) also distinguished the name of the quantity from its referent unit. For instance, the intensive quantity speed could have referent miles per hour or feet per second. In word problems involving extensive and intensive quantities, one further step is needed, beyond coordination of each one of those quantities. We somehow would need to reconcile all these quantities, each of which can be coordinated in the form (name of quantity, unit of the quantity). In other words, we not only look at each coordinated quantity separately, but also look at all these quantities together as a whole. This coherence of the whole requires that we meaningfully combine each coordinated quantity: A coordination of coordinated quantities. We refer to this second level of coordination as “quantitative unit conservation”. Introducing a second level of coordination of quantitative units necessitates a view of different levels of quantitative units. Behr, Harel, Post, & Lesh, (1994) conducted a conceptual analysis of different levels of quantitative units. Steffe (1994) provided a psychological analysis of children’s construction of composite units at different levels of composition (a singleton unit, a unit of units, and a unit of units of units) involved in multiplicative reasoning. Olive (1999), Olive and Steffe (2002), and Steffe (2002) extended the idea of multiple levels of units to children’s reasoning with fractional quantities. We shall show, in the analysis of the coins problem in this study, that students need to operate with three levels of units in order to successfully make a coordination of coordinated quantities necessary for quantitative unit conservation. Thompson (1988, 1995) provides another way of thinking about this second level of coordination through his description of quantitative operations and relationships. Thompson (1995) states: “quantitative operations and numerical operations should not be thought as being the same.” (p. 212) In his 1988 paper he describes four types of quantitative operations: combining quantities either additively or multiplicatively, and comparing quantities either additively or multiplicatively. He goes on to state that: Complex quantitative reasoning entails relating groups of quantitative mental operations, such as in forming a multiplicative comparison of an additive comparison and an additive combination (i.e., “How many times bigger is this difference than is this combination?”). Quantitative reasoning also entails reasoning relationally about quantitative structures, entails the constitutive mental operations for comprehending a quantity situationally, and entails the constitutive mental operations which allow one to recognize a quantity as one whose value varies or can vary. (p. 165) We consider this elaboration to be supportive of our construct of quantitative unit conservation. With this construct, we are opening one more theoretical perspective that covers a range of mathematical practices associated with solving word problems. Such mathematical practices include, but are not limited to, taking care of priority of operations, using parentheses appropriately (in order to combine quantities), and substituting literal expressions for other literal symbols. All these mathematical practices serve one crucial idea, and that is to maintain the equality of expressions on both sides of an equation (Chazan and Yerulshalmy, 2003), while being aware of what's happening on both sides: Things we are adding or subtracting have to be like terms while those we multiply or divide do not necessarily have to be so. The simplified expressions on both sides of the equation must be “like-terms” in the sense that they both have the same “simplified” unit. The simplified unit throughout the process of obtaining equivalent equations must be conserved. In this paper we explore the units coordination arising from situations that can be represented by linear equations involving more than one unknown or variable but that can be reduced to an equation in a single variable; that is, a system of linear equations that can be solved by substitution. We consider these situations to be ones that require complex quantitative reasoning (Thompson, 1988) in that they entail relating groups of quantitative relationships; they also involve the use of algebraic notation that adds another layer of symbolic complexity to students’ quantitative reasoning. Students often associate the algebraic symbol with the name of the quantity rather than its magnitude, as Thompson (1995) pointed out: When we reasoned symbolically, we needed to remind ourselves continually that W stood for the number of women and that M stood for the number of men. When students fail to keep in mind that letters represent numerical values, they will think of an expression like W=8/9M as saying “one woman is eight-ninths of a man” instead of thinking “the number of women is eight-ninths the number of men.” Also, students will often read the (equivalent) equation 9W=8M as “There are 9 women for every 8 men” instead of as “9 times the number of women equals 8 times the number of men.” Students’ thinking of letters as standing for objects is well researched, and it has pernicious consequences for students’ understanding of algebra... (p. 209) Several students in this study made similar associations between the letters in an algebraic expression as standing for objects (names of quantities) rather than numbers (magnitudes of the quantities). Through our analysis of the classroom discussions, students’ explanations and responses to interview tasks, along with interviews with the classroom teacher, we have come to realize that the identification and coordination of the units involved in the problem situation (the name-unit coordination) are critical aspects of quantitative reasoning that need to be emphasized in the teaching-learning process. Context and Methodology This study took place in an 8th-grade classroom in a rural middle school in the southeastern United States. The 24 students were between 13 and 14 years old and had been placed in the algebra class based on their success in 7th-grade mathematics. The students were racially, socially and economically diverse, with an approximately equal distribution of gender. All eight class lessons on a unit that focused on writing and solving algebraic equations from word problems were videotaped using two cameras, one focused on the teacher and the other on the students. Four students were interviewed twice in pairs (a pair of girls and a pair of boys) during the three weeks of the study. The classroom teacher was also interviewed twice during the three weeks. All interviews were videotaped. Excerpts from the classroom videotapes were used during both student and teacher interviews to initiate discussion of the learning that was taking place in the classroom. Excerpts from the videotapes of student interviews were also used in the teacher interviews. The first author conducted all of the interviews. This paper focuses on problems arising from a particular contextual situation (the Coins Problem). The data for the Coins Problem were collected during the two class lessons and subsequent student and teacher interviews that dealt with the following word problem from UNIT 4 of College Preparatory Mathematics (CPM) Algebra 1, 2nd edition (2002): Mrs. Speedy keeps coins for paying the toll crossing on her commute to and from work. She presently has three more dimes than nickels and two fewer quarters than nickels. The total value of the coins is $5.40. Find the number of each type of coin that she has. Students were first asked to create a “guess and check” table to find possible solutions to the problem. They were then challenged to write an equation to represent the problem. This Coins Problem gave rise to student difficulties that can be explained in terms of unit identification and coordination. Analysis Process Each day the classroom video data from the two cameras were viewed and digitally mixed using a picture-in-picture technology. A written summary of the lesson with time-stamps for video reference was created from the mixed video. This written summary also contained comments about any significant events and screen shots from the video when needed for clarification or highlight. These written “lesson graphs” were then used to select excerpts from the classroom video to be used in the student or teacher interviews, and to plan questions and related problems to pose to the interviewees in an effort to understand how the students (and teacher) had interpreted the problem and the classroom discussions that followed from different students’ attempts to address the problem. CPM ALGEBRA I UNIT 4, Choosing a Phone Plan (CP) CPM UNIT 4 CLASS CP 0,1 10/25/04 CP 1,4-8 10/26/04 CP 15,16 10/27/04 Students P&M Students B&G TEACHER 10/28/04 10/29/04 CP 17,18 11/01/04 CP 27, 28, 38 11/02/04 CP 45, 40, 48(a), 39(c), 65(b) 11/03/04 11/03/04 11/03/04 11/05/04 CP 92 11/08/04 CP 92 11/09/04 11/10/04 Figure 1: Connections among class lessons, student interviews & teacher interviews After the end of the three weeks of data collection, the corpus of classroom video data was reviewed, along with the associated lesson graphs to generate possible themes for a more detailed analysis. All student and teacher interviews were transcribed from audio files created from the videotapes of the interviews. A chart of relationships among class lessons, student interviews and teacher interviews was then created. This chart indicated which class lessons (including the specific activities from the CPM unit) were used or referenced in which student and teacher interviews, and which student interviews were used or referenced in the teacher interviews (see figure 1). A retrospective analysis, using constant comparison methodology, was then undertaken during which the classroom video, related student interviews and teacher interviews were revisited many times in order to generate a thematic analysis from which the results emerged. Results of the Analysis The protocols we are presenting below came from important events in class lessons, student and teacher interviews. By important events, we mean critical instances that are meaningfully related to our two theoretical constructs (quantitative unit coordination and conservation). For the coins problem, we started with a detailed description of the situation from the related class lesson. We then looked for where else the coins problem was used among class videos, student interviews, and teacher interviews. We revisited the transcripts of these video-episodes many times, and refined them further. The protocols presented below do not necessarily follow a chronological order; rather, critical incidents that occurred during class lessons are followed up through excerpts from both student and teacher interviews before returning to the protocols excerpted from the class lessons. This sequence of evidence follows the thematic analysis that emerged from the retrospective analysis of the total set of video data. The Coins Problem Mrs. Speedy keeps coins for paying the toll crossing on her commute to and from work. She presently has three more dimes than nickels and two fewer quarters than nickels. The total value is $5.40. Find the number of each type of coin she has. (from CPM Algebra 1, UNIT 4, CP-16, 2002) This problem was introduced during the third class period in Unit 4 on 10/27/04 (see Figure 1). Problem CP-3 in the Unit was very similar, also dealing with combinations of nickels, dimes and quarters, but the teacher had skipped over that problem in a previous class period, thus problem CP-16 was the first one of this type that the students had encountered. In this problem situation, when trying to calculate the total value of all coins, the monetary values of specific coins are intensive quantities (they are the values per coin) and the numbers of each coin and total value are extensive quantities. Distinguishing between these two different types of quantities surfaced as a problem during the classroom discussions. Associating appropriate units with the different quantities and combining unknown quantities emerged as further problems during the student interviews. A major confusion arose during the class lesson on 10/27/04 in naming the quantities in the situation. Students had chosen the letter N to represent the nickels in the problem, however, it became apparent from the discussion that, while N stood for the number of nickels for the teacher and for some of the students, for others it either represented the value of all the nickels together or just stood for the coin (a nickel). When the teacher, Ms. Jennings2 asked the students “What are we gonna call dimes?” (immediately after writing “n=nickels” on the classroom board) some students answered “two N”, and this could be a corroboration that those students saw N as the value, and not the number of coins under consideration. The following dialogue between Ms. Jennings and a student, Cathy, taken from the classroom video illustrates the confusion: Protocol I: Student's confusion about naming coins (from classroom video on 10/27/04) Ms. Jennings: We are not done... We are just naming our variables right now. We haven't begun to make an equation yet. We have to know what we are naming, before we put in an equation. Cathy: So why can't we just put them all with their first letter? Like N equals nickels, just keep doing, D dimes, Q quarters. Ms. Jennings: Let me ask you this question and see if you can solve it: “N plus D plus Q equals 5 dollars and forty cents. How many of each one do I have?” By challenging Cathy with the statement “N plus D plus Q equals five dollars and forty cents. How many of each one do I have?” in the above dialogue, Ms. Jennings may have added to the confusion (over what the letters represented in the situation). Ms. Jennings actually wrote on the white-board during the lesson: “n=nickels, d=dimes, q=quarters” following Cathy’s suggestion. In her interview with the first author a few days later, Ms. Jennings commented that students name a coin by its first letter to make it easy to identify in the equation but that later confuses them. Protocol II: Students' confusion (from Teacher interview on 11/03/04) Ms. Jennings: But I also know that they will not be able to solve that equation with N and D and Q because they have no value at that point. Interviewer: Okay. Ms. Jennings: And even though they know that they’re adding 3 different kinds of coins together to make an amount of money, without consolidating that variable in some way, they won’t figure out the number of each one that they need to find. The source of this confusion partly comes from what Ms. Jennings had written on the board: “n=nickels, d=dimes, q=quarters”. Ms. Jennings’ pedagogical approach in the classroom is to accept students’ suggestions without evaluation from her, with the intent of having her students evaluate and discuss what is said during the lesson. This approach leads to rich discussions and productive arguments among the students but can also leave some students confused as to what is mathematically acceptable and what is not. Ms. Jennings’ introductory question “What are we gonna call nickels, dimes, quarters?” could have been misleading (as she did not specifically say number of nickels, dimes and quarters). In the interview with students Pam and Maria, on the morning following the classroom lesson, the interviewer (first author) showed the classroom video episode from Protocol I above. While Pam and Maria appear to have understood the situation and did not appear to have been mislead by the confusion evidenced in the classroom episode, this was not the case with the two boys who were interviewed later in the week. We begin with Pam and Maria’s interview as their responses to the interviewer’s questions framed the tasks for the later interview with the two boys. Protocol III begins after the interviewer has shown Pam and Maria the video of the classroom episode: Protocol III: Students' interpretation of Cathy's remark (from student interview on 10/28/04) Interviewer: Okay. What do you think Cathy means by N for nickels, D for dimes, and Q for quarters? Pam: That represents how much you have, that’s what she’s talking about. Interviewer: How much? Maria: No, she was thinking about the value of each one. Pam: The value, yeah. Interviewer: Oh, rather than what? Maria: The number of coins. Later in the same interview, Maria distinguished the differences among three different types of quantities: the value of a coin, the number of that coin and the total value of all the coins of that type. She was then able to combine her total values for each type of coin to produce the total of all coins ($5.40). Protocol IV: Unitizing quantities (from student interview on 10/28/04) Maria: Yes. Okay. Okay, this is the value of the nickel, so it would be… Interviewer: What is “this?” Maria: .05. And, in any number, let’s say 5, so it’ll be .05 times 5 will give the amount of nickels and you do… Interviewer: The amount of nickels? Maria: No, the value of the whole nickels that you have. Interviewer: Does that make sense? Maria: Yeah. And, then you do the same for D and Q and it comes out to $5.40. Interviewer: Do the same for D and Q for me. Maria: Okay, so it will be, let’s say, 10 dimes. So, it’ll be 10 times .1 will give you the value and the same for Q. If you do times any number, so Q… the letters mean any number you can think of. Interviewer: Well, what in the terms of the problem what do those letters stand for? Maria: The number of coins you need to get $5.40. Maria’s statement “the letters mean any number you can think of” is evidence that she knows she is dealing with the quantity “number of a coin”. Moreover, she separately calculates the total value for each coin, and this could be seen as her coordination of units before adding them together. In fact, during this interview, by explaining this unit coordination, Maria makes sure that the “terms” she is adding are like terms, and then she concludes the addition and writes the first equation 0.05n+0.1d+0.25q=5.40; and after substitution, the second equation 0.05n+0.1(n+3)+0.25(n-2)=5.40. Both Maria and Pam correctly write and explain these equations and they agree it makes sense, as the following protocol indicates: Protocol V: Student's Explanation of How the Second Equation comes from the First One (from interview on 10/28/04) Interviewer: Yes. Now how does the second one come from that first one. Pam: Because it said that we had 3 more dimes than nickels, right? So, that’s why it says N plus 3. Interviewer: And, what’s the N plus 3 instead of? Pam: The… Maria: D. Pam: The T? Interviewer: D. Maria: D. Pam: D, oh. The D. That’s right ‘cause it gets you the dime… the value of the dime and how much coins and how much dimes we have. Interviewer: How many dimes we have. Okay. Pam: And, then nickels, the N minus 2 gives you how much quarters you have. Interviewer: Why? Pam: Because it said you have 2 fewer quarters than you do nickels. Interviewer: Okay. Does this make sense now? Maria: Mm-hmm. Interviewer: Is that how you would explain it, Maria? Okay. Okay, good During the process of obtaining her equivalent form 0.05n+0.1(n+3)+0.25(n-2)=5.40, Maria performed several notable mathematical practices. First was her correct substitution of expressions for literal symbols, as in this case, n+3 for d, and n-2 for q. Second, her placing of parentheses around those expressions appropriately. This way of writing the expressions had a purpose. Each product on the left hand side represented a composed quantity, and had to possess a unit inherent in its structure. Moreover, each product, having the same unit: value, was connected meaningfully via the addition operation. This was when she identified these products as monetary values. In this whole process of obtaining the equivalent form, there is another meaningful mathematical practice, which we call quantitative unit conservation: Not only did each product on the left hand side of the equation have the same unit as the quantity on the right hand side, but their combination in the form of a sum – they could be combined because they were like terms – had the same unit as the quantity on the right hand side of the equation. In this way, there is this notion of coherence between each term on the left with the term on the right, as well as the coherence of the combined expression on the left with the expression on the right. In Thompson’s (1988, 1993) terms, Maria and Pam had constructed a quantitative structure – a network of quantitative relationships that were embodied in this second equation. We can also analyze Maria’s solution in terms of different levels of units: a single coin is the first level, the value of the single coin and the number of those single coins are units established at a second level (a composite unit of units), whereas establishing the value of all the coins requires a third level of units (a composed unit of units of units). The ease with which Maria established the second equation, with correct parentheses, indicates that she had these three levels of units available to her prior to operating. Even though these two students eventually wrote both equations correctly, Ms. Jennings had some doubts about other students’ realization that they would not be able to solve the first equation in terms of the three different variables: Protocol VI: Ms. Jennings’ Comments on Students' Setting up the Equations (from Teacher Interview on 11/03/04) Ms. Jennings: I know that they understand what the N, D and Q stand for and that they do know the value of each coin, obviously, but I’m not, I’m not sure how they will reach a conclusion that they can’t solve the first equation. I don’t know how they will decide they can’t do it. Interviewer: Yeah. Ms. Jennings: But if they can’t do it, then I guess my job is to re-route them into naming a dime in terms of a nickel by the information from the problem. I mean I really liked the first equation. Interviewer: Which? Ms. Jennings: The first one that both of them wrote. Interviewer: Oh. In terms of N, D and Q? Ms. Jennings: In terms of just making sense of how many you have and the value of each coin and then turning it more into each variable in terms of the first ‘cause the N + 3 and the N – 2 were implied in the problem's information, but they weren’t getting that the first time or the first couple of times. In this last remark, Ms. Jennings is referring to many students “not getting that” during the class lesson rather than Pam and Maria during the interview. Her reflection on her role as teacher indicates that she is aware of students’ needs to first represent the situation in terms of nickels, dimes and quarters and it is then up to her to help them use the relations among the numbers of the different coins to create an equation in one unknown that they can solve. In contrast to Maria’s success in setting up both equations in Protocol V above, the two boys who were interviewed (Ben and Greg), while able to make the unit coordination to produce the first equation similarly to Maria, were not able to produce the second equation through substitution. In searching for a possible explanation for this behavior, we found that in class, in the process of naming quantities, when the teacher challenged students by her statement “I have three more dimes than nickels”, Greg had commented “n plus 6”. In the interview, when he was questioned about this, he said “I was thinking about the value... because it takes six nickels to get three dimes”. Therefore, in class, it appears that Greg was aware of the two units concerning the coin – its number and its value – his answer, however, did not reflect an appropriate coordination between these two units. Greg’s partner, Ben, had given a very clear explanation in class for coordinating the value of each coin with the number of each coin. The interviewer asked him to repeat what he had said in class. Protocol VII begins with Ben’s explanation. Protocol VII: Ben's Explanation for Setting up the First Equation (from student interview on 10/29/04) Ben: Okay. Before you find the answer, I said whatever the values are for the numbers, like, a nickel is .05 and you have to multiply that times the number of nickels and that gives you the value and then the dimes, .1, I think, yeah, .1 times the number of dimes that you have and that’ll give you the value of that one and then the quarters… then .25 (inaudible) you have to multiply that times the number of quarters you have. Interviewer: And, what will that give you? Ben: It should give you $5.40. Ben appears to have reasoned with unit coordination, like Maria. He knew that when he multiplied the value of a coin by the number of that coin, the result was a monetary value. Following this explanation, the interviewer asked Ben to write down an equation based on his explanation, using the symbols N for number of nickels, D for number of dimes and Q for number of quarters. Ben eventually ended up with the following equation: (.05N)+(.1D)+(.25Q)=$5.40. The interviewer then asked the boys what they knew about D and Q. Greg responded that you have to have the nickels to find D and Q. The interviewer asked him to write down what D equals in terms of “that number of nickels” (pointing to the N). Greg wrote the expression: (.05n+3). Protocol VIII picks up at this point in the interview: Protocol VIII: Finding D in terms of N (from student interview on 10/29/04) Interviewer: Do you agree, Ben? Ben: Yeah. Interviewer: That’s… what about… what is that about the dimes? Is that the value of the dimes, the number of the dimes, the picture of the dimes? Ben: I think it’s the value. Interviewer: The color of the dimes? Ben: I think it’s the value. Interviewer: You think that’s the value of the dimes. Ben: Because the… ‘cause the… .05 is 5 cents and that’s a nickel and then the N is the number of nickels, so that gives you the value, plus 3 should give you another value. Ben's interpretation of Greg’s expression suggests that he was aware of the different types of quantities involved (value and number of coins) but may have had problems coordinating the quantitative units meaningfully. Ben was aware that 0.05n and 3 both must have the same unit, value, in order to be added. There is also the possibility that Ben interpreted the statement in the problem “three more dimes than nickels” to mean that the value of the dimes was three more than the value of the nickels. This would explain the acceptance of 0.05 (5 cents) as the unit value used to find the value of the dimes. In her interview Ms. Jennings interpreted Ben’s problem as stemming from the use of an unknown value for N in the expression, as indicated in the following protocol: Protocol IX: Ms. Jennings’ Interpretation of Ben's Explanation (from Teacher Interview on 11/03/04) Ms. Jennings: I think he really knows what he means, but he’s really having trouble with interpreting how to write it as an expression. I think, I think that if he knew a number of nickels and he knows the order of operations, if he knew the number of nickels and multiplied by the 5 cents and then added a 3, he would see his mistake, but I don’t think he would do it without knowing a number of coins. Interviewer: Okay. So what does that imply about his understanding of the role of N there? Ms. Jennings: I don’t think he sees N + 3 as one number. I think he only sees the N as a number and whatever it is, you’re gonna add 3 to it. In the continuation of the student interview, the interviewer also believed that Ben would realize his mistake if he was to work with an actual value for N. He suggested a little experiment. He asked the two boys to assume that they have 2 nickels, and evaluate their conjecture for this value of N. The students accepted that if they have 2 nickels, they must have 5 dimes and the value of 5 dimes is 50 cents. When they were asked to evaluate the expression 0.05n+3 for n=2, they realized that it would exceed three whole dollars, and realized that their conjecture must be false. The interviewer then shifted the focus back to what D was in terms of N. Protocol X begins at this point in the interview: Protocol X: Creating Expressions for D and Q (from student interview on 10/29/04) Interviewer: But, I don’t want the value, I want D. What’s D in terms of N? Greg: N plus 3. Interviewer: Can you write that down? Put D equals. [Greg writes: D=N+3] Now, what do you know about Q? Ben: Q is 2 less then N. Interviewer: Yeah. So, write down the equation for me. Ben: N minus 2. [Greg writes Q=N-2] Interviewer: Okay. So, now can you rewrite this equation [pointing to the first equation: (0.05 n) + (0.1 d) + (0.25 q) = 5.40] just using N? Okay, can you do that for me? Greg: To get the value? Greg started by writing .05n, then put a plus sign, and then put n+3 and stopped. He hesitated for a while in this step, and he asked himself (audibly) “How can you get the value?” He erased the n+3, and replaced it with .1 n + 3, without parentheses. He did the same thing for the last expression and wrote .25 n – 2 . His complete expression for the second equation was 0.05n+0.1n+3+0.25n–2=5.40. Upon the interviewer's question whether he agrees, Ben said that 0.05n was correct. They both hesitated for the remaining terms on the left hand side. They tried to compare this expression, namely their second equation with their first equation (0.05 n) + (0.1 d) + (0.25 q) = 5.40 (note that Ben had placed parentheses around each of the expressions that indicated the value of each set of coins in this first equation). Protocol XI continues from this point: Protocol XI: Finding the Second Equation by Substituting for D and Q (from student interview on 10/29/04) Interviewer: Let’s look at this equation that you all agreed was correct [points to first equation]. Ben: 5 cents times the number of nickels plus 10 cents… Greg: (inaudible), but you’re trying to get D equals N plus 3? Interviewer: Yeah. Read what this says. Ben: 0.1 times D or times the number of dimes. Interviewer: Keep that. 0.1 times the number of dimes. Is that what this says [pointing to the second part of Greg’s equation: 0.1n+3]? Ben: This says, .1 times the number of nickels. So… Greg: Plus 3. Interviewer: Plus 3. Greg: It gives you the number of dimes. Ben: Oh, if you say it like that then that means the nickels are the same as … the two N’s are the same? Interviewer: Yeah, the two N’s are going to be the same. Once you’ve picked the number for the number of nickels that stays the same throughout the equation. Okay. This says, .1 times the number of dimes [pointing to (.10D) in the first equation]. Does this say .1 times the number of dimes [pointing to.10n+3 in the second equation]? What did it say? Ben: .1 times the number of nickels… Interviewer: Plus 3. [Greg draws parentheses around the (n+3)] Okay… what about the other one…[Greg draws parentheses around the (n-2)] Now, does it say, .1 times the number of dimes? Greg: Yes. Ben: Yeah. Interviewer: What does this say [pointing to .25(n-2)]? Greg: That says, .25 times (overtalking) Ben: Times the number of nickels. Greg: Quarters. Ben: Yeah, quarters. The big surprise in the above protocol is Ben’s question about the nickels – the two N’s – having to be the same in the second equation. His question and surprise indicate that Ben (like several other students in the class) had not realized what we, as adults, take for granted when reading an algebraic expression: that the same literal symbols stand for the same values throughout the expression. Greg’s question after writing n+3: “How can you get the value?” indicates that he was making a distinction between the two different quantitative units (number of coins and value of coins). His lack of use of parentheses when writing the expression 0.1n+3 caused a perturbation for Ben, whereas Greg knew that the n+3 gave the number of coins. Greg eventually realized that he needed to add parentheses to distinguish this quantity (number of coins) from the value of the coin (0.1), and produced the second equation: .05n+.10(n+3)+.25(n-2)=5.40. The lack of appropriate parentheses in Greg’s first expression for D (.05n+3) and in his first attempt at the second equation led to Ben and Greg’s difficulties in coordinating units within their quantities and seeing the coherence of the whole equation, namely that each product on the left hand side must be consistent in units with the term on the right hand side, and that they could then be added because they were in terms of the same unit: monetary value. In her interview, Ms. Jennings commented about the boys' mistakes concerning parentheses: Protocol XII: Ms. Jennings’ comment on boys' omitting parentheses (from Teacher Interview on 11/03/04) Ms. Jennings: I wonder how they know that parentheses make that difference? Interviewer: It helps unitize that expression (N + 3) is my interpretation. Ms. Jennings: And I know that and you know that, but I’m not sure how they realize that. Interviewer: Yeah. For the girls, it was there immediately. For Pam and Maria, they wrote their equations with the parentheses. Ms. Jennings: And this took quite a while. In our opinion, the use of parentheses is one demonstration of how much a student is able to understand and to see the relation between the quantities and their units, and the different levels of units that are available to them prior to operating. At the beginning, both in the classroom video and students’ interviews, students hesitated a lot about whether they were dealing with values or numbers of coins. For the time being, this problem has not been resolved entirely, and it seems to us that the boys' omitting parentheses is in a way related to their misunderstanding and misinterpreting the quantities arising from the problem. In contrast to the girls, Ben and Greg were not able to construct a meaningful quantitative structure using the literal symbols in these equations. They were able to work with units at the second level (number of coins and value of a coin), but did not have available a third level of units that would enable them to envision the quantitative structure required to meaningfully combine and find the value of all the coins. They can construct this third level structure through their activity but it was not available to them prior to activity (as it was for Maria). As stated in the above protocol, the use of parentheses helped the boys to unitize the unknown quantities (numbers of each coin) so that they could then combine them with the known quantities (values of each coin) to produce a value, but this only required operating with two levels of units at one time. The third level of unit (value of combined coins) only emerged after their struggles with these two levels of units. The teacher interview on 11/03/04 started with asking Ms. Jennings her perception of what was going on with the students. Her initial comments are worth noting: Well, my first perception is probably that I jumped in too fast and that they weren’t ready to think about three different variables in terms of one. And, also, the fact that with coins we have decimals to play with didn’t make the problem any easier for them to think about. Even though they know what value is represented by coins, to multiply that value to a variable that they don’t understand yet was a leap. Ms. Jennings recognizes the complexity in the coins problem but focuses on there being more than one “variable” and the use of decimals rather than the difficulty we perceived as distinguishing between names of coins and the different quantitative units, and levels of units. Throughout the class lesson Ms. Jennings used the name of the coin, (nickel, dime, quarter) to stand for the number of coins. There are several instances from the interview where this issue becomes apparent. For instance, her comments “I guess my job is to re-route them into naming a dime in terms of a nickel by the information from the problem.” could imply that for the teacher, a dime is a variable that can be expressed in terms of the other variable, nickel. She probably used the names of coins assuming that number of the coins was to be understood. We believe that this practice became ambiguous for students; they were unsure as to whether the number or the value of a dime was the referent, when using its name (a dime). In fact, at the beginning of the class session, we believe that it was this communication problem that lead students in an unproductive direction. Ms. Jennings commented about the boys' construction and confusion over the expression (0.05n+3); she said that Ben really knew what he meant, but he was having trouble with how to write it as an expression. She thought that if he knew the number of nickels and multiplied by the 5 cents and then added a 3, he would see his mistake. Ms. Jennings commented, however, that Ben would not realize his error without knowing a specific number of coins. In other words, when focused on the nature of these coins, Ms. Jennings realized that the coin must be associated with the units inherent in its structure. When questioned about Ben's understanding of the role of n in (0.05n+3), she said she does not think Ben sees n + 3 as one number. She added “I think he only sees the n as a number and whatever it is, you’re gonna add 3 to it”. Ms. Jennings is aware of students’ difficulties with interpreting algebraic expressions as a “process” to be carried out rather than as an “object” on which to operate (Tall et al., 2001). In another instance from her interview, Ms. Jennings was asked to compare the behavior of the boys with that of Maria. She commented that Maria understood that the “n + 3” was an expression meaning d and that Maria had an understanding of what an expression means: just to rename one thing by another name. The boys, however, did not really have that understanding. According to Ms. Jennings, Ben and Greg understood how to name the new unknown quantity, but they didn’t know how to use it to coordinate an amount of money with an amount of coins. In other words, they could not make an appropriate “name-unit” coordination. Conclusions & Discussion With this study, we have attempted to extend Thompson’s (1988, 1989, 1993, 1995) theoretical framework that encompasses a way to look at word problems that can be solved via a set of equations that are reducible to a single equation with a single unknown. We claim that through this lens, one could interpret all the variables arising from word problems as not simply ordinary quantities named in the problem, but, rather, as mathematical objects with names, values and associated units. We tried to rely on students' judgments essentially on this coordination of different quantities’ names and their units. Hence the theoretical construct “quantitative unit coordination” is suggested as the main extension to Thompson’s theoretical model of quantitative reasoning that we want to bring forth through this study. In applying a quantitative unit coordination construct to the coins problem, a coin was not to be understood simply as its name (e.g. “dime”). Rather we needed to associate or coordinate a coin with a unit present in its monetary nature. The Coins Problem helped us focus on each coin in the problem as an object with a name and the unit associated with the specific coin. A dime is not a quantity; dime is only the name of the coin stated in the problem. Once a dime is associated with its value, for instance, that's when a quantity is born. Therefore, with this quantitative unit coordination, “value of a dime” or “value per dime” becomes an intensive quantity. Similarly, as our analysis also shows, neither n, nor d, nor q stands for the names of the coins: nickel, dime, and quarter, respectively. It was essential to coordinate once again the names of these coins with their “number”, hence to make of them extensive quantities with referent units “number of nickels”, “number of dimes”, and “number of quarters”. We needed not only to coordinate each mathematical object with its name and unit, but also to do a coordination of the big picture that consisted of six different values (three of which were unknown) enabling us to write the linear equation 0.05n+0.1d+0.25q=5.4. Such complex coordinations are necessary to form what Thompson (1993) calls a quantitative structure – a network of quantitative relationships. We relied on evidence of students' judgments concerning this coordination of the quantitative referents and their units. Maria’s successful coordination of the quantitative referent and its unit, and the difficulties that Ben and Greg experienced in interpreting the various quantitative referents and combining them with appropriate units to produce quantities that conserved units within an expression (the quantitative structure), support our theoretical conjectures. From an analysis of units point of view (Behr, Harel, Post and Lesh, 1994; Olive, 1999; Olive & Steffe, 2002; Steffe, 1994, 2002) the boys appeared to be limited to operating with only two levels of units (singletons and a composite unit of units), whereas the girls appeared to be able to operate with three levels of units (a composite of composite units). In order to construct a network of quantitative relationships, the students need to be able to form a composite unit of the separate quantitative relationships, each of which is a composite unit (a coordination of different unit quantities). Thus, the ability to work with three levels of units is necessary to form a quantitative structure of this complexity. Another key issue arising from the above analysis is the “consistency” of quantities on the left and right hand sides of an equation. In the coins problem and various other problems, students looked for “likeness” of terms on each side of an equation. This desire for consistency emphasizes another crucial issue, namely the conservation of quantities. The reduced equation for the coins problem 0.05n+0.1d+0.25q=5.4 had to have a consistency, therefore students wanted to look at this equation as representing “value” on each side. As our analysis shows, transforming this first equation in three unknowns to the second equation in only one unknown, 0.05n+0.1(n+3)+0.25(n-2)=5.4 was not very easy for the students. Some students had trouble with parentheses and they overcame this difficulty by relying on quantitative unit conservation: They wanted to have “value” on each side. Therefore here, we hypothesize that quantitative unit coordination necessitates quantitative unit conservation. In Thompson’s 1988 study he found cognitive obstacles to students’ quantitative reasoning that he believed constituted obstacles to algebraic reasoning. One major obstacle was that students’ “failure to distinguish between a quantity and its measure hindered their ability to explicate relationships.” (p. 168) Another major obstacle was that “Multiplicative quantities of any sort (products, ratios, rates) were commonly mis-identified or given an inappropriate unit.” (p. 168) Our study suggests that unit coordination and unit conservation are necessary constructs for overcoming these cognitive obstacles when reasoning quantitatively about a situation. Smith and Thompson (in press) also emphasize a necessity to coordinate between two levels of reasoning: Quantities that result from quantitative operations exist in two different senses, as quantities in their own right and as relationships between the two quantities. It can be conceptually demanding to reason and communicate about such quantities because we must distinguish and coordinate these two senses, and, when necessary, shift between them. (p. 23) This necessity to “shift between them” is what we have referred to as our second level of coordination: a coordination of coordinated units that has as its goal unit conservation. The current study suggests that such a coordination, however, requires operating a priori with three levels of units. The particular teacher in this study had developed a pedagogy based on constructivist principles of how students learn. It was her intension to have students discuss and debate their own interpretations of a particular problem. Her awareness and interpretation of her students’ difficulties in solving the coins problem, led her to question her role in the learning process and helped her to realize that there are appropriate times when she needs to lead students in a more productive direction, and appropriate times to make productive use of a student’s (insightful) interpretation. While realizing that her use of her students’ convenient but ambiguous naming of quantities (without clarification) could contribute to her students’ confusions, she also became aware that her goal to have students simplify algebraic expressions could lead to a miss-match (for her students) between the original situation and the simplified equation. This rush to simplify equations can be understood when the goal throughout solving quantitative problems arithmetically has been to get an answer. Teachers’ and students’ goals and processes need to shift, however, for the development of quantity-based algebraic reasoning. As Thompson (1989) has suggested, the shift in goals is from the goal of “getting an answer” to the goal of “laying out a pattern of reasoning.” The shift in cognitive processes is from immediate evaluation of expressions to the postponement of calculations so that one constructs numeric expressions that capture a “history” of a quantity’s value. These shifts in goals and processes amount to students developing the intention of constructing “numeric formulas” for calculating the values (or magnitudes) of quantities. We conclude from the results of this study that unit coordination and unit conservation are key aspects of such numeric formulas, and their construction requires operating with three levels of units. NOTES: 1. CoSTAR is supported by a grant from the National Science Foundation, grant # REC 0231879. The opinions expressed in this paper are those of the authors and do not necessarily reflect the views of NSF. 2. All names are pseudonyms. References Behr, M. J., Harel, G., Post, T. & Lesh, R. (1994). Units of quantity: A conceptual basis common to additive and multiplicative structures. In G. Harel & J. Confrey (Eds.), The development of multiplicative reasoning (pp. 121-176). New York: SUNY Press. Chazan, D. & Yerulshalmy, M. (2003). On appreciating the cognitive complexity of school algebra: Research on algebra learning and directions of curricular change. In J. Kilpatrick, W. G. Martin and D. Schifter (Eds.), A research Companion to Principles and Standards for School Mathematics, 123-135. Reston, VA: National Council of Teachers of Mathematics (NCTM). College Preparatory Mathematics (Algebra 1), Second Edition (2002). T. Salle, J. Kysh, E. Kasimatis and B. Hoey (Program Directors). Sacramento, CA: CPM Educational Program. Kieran, C. & Sfard, A. (1999). Seeing through symbols: The case of equivalent expressions. Focus on Learning problems in Mathematics, 21(1), 1-17. Lamon, S. (1994). Ratio and proportion: Cognitive foundations in unitizing and norming. In G. Harel & J. Confrey (Eds.), The development of multiplicative reasoning (pp. 89-120). New York: SUNY Press. Olive, J. (1999). From fractions to rational numbers of arithmetic: A reorganization hypothesis. Mathematical Thinking and Learning, 1 (4): 279-314. Olive, J. & Steffe, L. P. (2002). The construction of an iterative fractional scheme: The case of Joe. Journal of Mathematical Behavior, 20, 413-437. Steffe, L. P. (1994). Children’s multiplying schemes. In G. Harel & J. Confrey (Eds.), The development of multiplicative reasoning in the learning of mathematics (pp. 3-39). Albany, NY: SUNY Press. Steffe, L. P. (2002). A new hypothesis concerning children's fractional knowledge. Journal of Mathematical Behavior, 20, 1-41. Schwartz, J. L. (1988). Intensive quantity and referent transforming arithmetic operations. In J. Hiebert & M. Behr (eds.) Number concepts and operations in the middle grades, 41-52. Reston, VA: National Council of Teachers of Mathematics (NCTM) and LEA. Smith, J. P. III & Thompson, P. W. (in press). Quantitative Reasoning and the Development of Algebraic Reasoning. In J. Kaput, D. Carraher, & M. Blanton (Eds.), Employing children's natural powers to build algebraic reasoning in the context of elementary mathematics. Tall, D., Gray, E., Ali, M. B., Crowley, L., DeMarois, P., McGowen, M., Pitta, D., Pinto, M., Thomas, M., & Yusof, Y. (2001). Symbols and the bifurcation between procedural and conceptual thinking. Canadian Journal of Science, Mathematics and Technology Education, 1(1), 81-104. Thompson, P. W. (1988). Quantitative concepts as a foundation for algebra. In M. Behr (Ed.). Proceedings of the Annual Meeting of the North American Chapter of the International Group for the Psychology of Mathematics Education Vol. 1 (pp. 163-170). Dekalb, IL. Thompson, P. W. (1989, March). A cognitive model of quantity-based algebraic reasoning. Paper presented at the Annual Meeting of AERA, San Francisco. Thompson, P. W. (1993). Quantitative reasoning, complexity, and additive structures. Educational Studies in Mathematics, 25(3), 165-208. Thompson, P. W. (1995). Notation, convention, and quantity in elementary mathematics. In J. Sowder & B. Schapelle (Eds.), Providing a foundation for teaching middle school mathematics (pp. 199-221). Albany, NY: SUNY Press.