You are on page 1of 16
Algebra Word Problem Solutions: Thought Processes Underlying a Common Misconception John Clement Journal for Research in Mathematics Education, Vol. 13, No. 1 (Jan., 1982), 16-30. Stable URL htp:/ links jstor-org/sici?sici=002 1-825 1%28198201%2913%3A 1%3C 16%3AA WPSTP%3E2,0,CO%3B2-P Journal for Research in Mathematics Education is currently published by National Council of Teachers of Mathematics, ‘Your use of the ISTOR archive indicates your acceptance of JSTOR’s Terms and Conditions of Use, available at hup:/www,jstororglabout/terms.hml. ISTOR’s Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www jstor.org/journals/netm.html Each copy of any part of @ JSTOR transmission must contain the same copyright notice that appears on the sereen or printed page of such transmission. STOR is an independent not-for-profit organization dedicated to creating and preserving a digital archive of scholarly journals. For more information regarding JSTOR, please contact support @ jstor.org, hupswww jstor.org/ Mon Feb 14 14:31:35 2005 ALGEBRA WORD PROBLEM SOLUTIONS: THOUGHT PROCESSES UNDERLYING A COMMON MISCONCEPTION JOHN CLEMENT University of Massachusetts This article describes test data showing that a large proportion of science-oriented college students were unable to solve a very simple kind of algebra word problem. Several thinking-aloud protocols are presented in order to (a) formulate hypotheses for why the errors occurred, (b) examine the phenomenon of shifts between inconsistent approaches within a single solution, and (c) illustrate a method of diagramming thought processes. The data indicate that relatively advanced students can experience serious difficulties in symbolizing certain meaningful relationships with algebraic equations. Paige and Simon (1966) have shown that syntactic methods—that is, ‘methods not dependent on comprehending the meaning of the described problem situation—are adequate for solving some algebra word problems ‘one variable. They point out, however, that these methods can produce incorrect or meaningless results in other problems. They also present evidence for the idea that some students use an internal “physical representation” that encodes qualitative features of a situation in solving word problems. A number of misconceptions concerning the meaning of algebraic equations have also been discussed by Wagner (Note 1), Collis (1978), Herscovics and Kieran (1980), Galvin and Bell (Note 2), Kiche- mann (1978), and Matz (Note 3). Several of these researchers have noted a tendency on the part of some students to treat numerical variables as if they stood for objects rather than numbers. This study concentrates on ‘one particular type of misconception in order to obtain a deeper under- standing of its cognitive sources. Research attempting to describe thought processes in this area must be considered exploratory, since the pro- cesses involved are rather complex. ‘Research reported in this paper was supported by NSF Award No. SED7S-22083 inthe Joint National Institute of Education/National Science Foundation Program ‘of Research on Cognitive Processes and te Structure of Knowledge in Science and Mathematics. It is based in part on a. presentation given at the Amenican Educational Research Association annual meeting, Boston, April 1980, Written Test Data Table 1 shows six problems given on a 45-minute written test to 150 freshman engineering students. The test was administered during a regularly scheduled class period in the first semester. Problems 5 and 6 belong to a class of problems that should be trivial for a scientifically literate person but that were solved incorrectly by large numbers of these science-oriented students. The contrast between the number of students who correctly solved Problems 1-4 and the number who correctly solved 5 and 6 suggests that most errors were due to a difficulty in translating words to algebraic equations rather than a difficulty with simple algebraic manipulation skills or with simple ratio reasoning of the type required for Problem 4. In a sample of 47 non-science majors taking college algebra, ‘Table 1 Performance on Six Questions in the Writen Test ‘Test questions ( i Solve fore Se = 50 50) A Solve for xin terms of a: 9a = 10s 4. Jones sometimes goes ovis hin iad ‘of gas. When he vt his fend Schwert, fre drives 90 mies and uses“? galons of Bay. (Assume the same driving conditions fboth cases) ” % 5. White an equation using the variables Sand P to tepresent the following satement “there are six times as many students as professor af thin university” Use for fre number of students and P forthe number of profesors Cn) 6. Write an equation using the variables C 2nd $f represen teflon tatenen SAAC Mindy's restaurant for every four people who ordered checscoake, there are ve People wig ordered strudel™ Let C Fepresen the numberof cheesecakes and 5 Fepresent the numberof strdels ordered aso acess 43% solved Problem 5 correctly (Clement, Lochhead, & Monk, 1981). In a different sample, Kaput (Note 4) found that performance was not improved (it was slightly worse) when students were asked to use X and Y as variables instead of 5 and P on Problem 5. AL first it was thought that the errors on such simple problems must be due primarily to carelessness. There was, however, a strong pattern in the errors. In both cases 68% of the errors were reversals: 6S = P (or an algebraically equivalent statement) instead of S = 6P, and 4C = 5S instead of SC = 45. Also, roughly half of the students were given the January 198217 following hint with both problems: ““Be careful. Some students put a number in the wrong place in the equation.” The percentage correct for the group given the hint was only 3 points higher on Problem 5 and 5 points higher on Problem 6. This indicated that the difficulty was more deep-seated than expected. Other written test data discussed by Clement, Lochhead, and Soloway (Note 5) have shown that the difficulty is quite robust in the sense that it shows up consistently in different symbol systems. That is, high percent- ages of reversals are also observed in translations from pictures to equations, data tables to equations, and equations to sentences. This demonstrates that the reversal error is not simply a result of ordering the words in a problem in a particular way. The error persists even when the subject works from pictures or tables instead of words. The misconcep- tions causing the reversal error also seem to be fairly resilient. Calculus students who worked through a 15- to 30-minute teaching unit involving a worked example and practice problems still exhibited basic misunder- standings of the meaning of equations for such problems as 6 (Rosnick & Clement, 1980). Something other than carelessness appears to be in- volved in these errors. Protocol Data In order to develop hypotheses concerning the cognitive events causing these results, audiotaped and videotaped clinical interviews were con- ducted with 15 freshmen who were asked to think aloud as they worked ‘on the Students and Professors Problem and similar problems. Two conceptual sources of reversal errors, a syntactic word order matching Process, and a semantic static comparison process (Figure 1), were identified from the protocols. Word order matching approach. In word order matching, the student simply assumes that the order of key words in the problem statement will Swwracric Senanrie Ss Nerwoo Wow Rae pena oer ahetkison | | Fron re DUALITY Asner es=P | les-P S=6 Resut Incorrect} | Incorrect _| | Correct Passiv active As$0EtATION OPERATION Figure 1. Solutions Methods for Students and Professors Problem. 18 Journal for Research in Mathematics Education ‘map directly into the order of symbols appearing in the equation. This is a syntactic strategy in the sense that it is based on rules for arranging symbols in an expression that do not depend on the meaning of the expression. For example, subject S1 immediately wrote 6S = P and said, “Well, the problem states it right off: °6 times students.” So it will be six times S is equal to professors.” There is no evidence for any more complicated strategy than that of mechanically matching the order of symbols in the equation to the order of words in the problem statement. ‘Static comparison approach. Evidence for the second incorrect strat- egy, the static comparison process, is provided by the transcript of S2 solving the England Problem below. The problem is to write an algebraic expression for the statement, "There are 8 times as many people in China as there are in England.” This student’s second attempt to write an expression is correct (Expression 2 in Line 4). He then actually rejected this expression, however, in favor of the incorrect, reversed expression. ‘Transcript 1: $2; England Problem Student's written work () I per E=8pC Q) 8E=1C @) 8C = 1E (Transcript Excerpts) 1 I: This is one where you have to write an equation for the English sentence. 2S: ‘There are 8 times as many people in China as there aren England." (Writes per pc) So let's use the letters C and E. S: (@)OK—for every person in England there ae 8 people in China so—8 equals 1C as an equation. (Writes 8E ~ IC; Expression 2.) (b) Well if I put—there are ‘more people in China than there ae in England, so if | put 8 Chinese equals one Englishman—{(Writes 8C = 16; Expression 3.) 5: 8Cequals 12 —ah-ub 6 S: 8 Englishmen equals one Chinese—am— 7: What are you thinking about now? 8S: I'm thinking about how stupid itis that I just put this wrong. No—this second ‘eatin [RE = IC] here is wrong—the third one don’t know what I just 9: The third one looks good? 10S: Yeah, the thi one—for every 8 Chinamen—but then again, saying for every 8 Chinese—yeah, thats right—for every 8 Chinese there's one Englishman, 11 I: So just explain how the third one looks good. 12S: Allright—it means tht there is larger number (points to 8C) of Chinese for every Englishman (points to 12). So\you're pointing tothe BC and the 1E? ‘Yeah, and there isa larger number of Chines than there are Englishmen; therefore the number of Chinese to Englishmen should be larger—8C = IE. The student did not simply make a careless mistake by guessing quickly, since he actually wrote the correct answer, considered it, and switched to the reversed equation. Why did this happen? January 198219 In Line 12, $2 clearly indicated that he had comprehended the relative sizes of the two groups in the problem—that there are more people in China. This indicates that he had gone beyond a syntactic word order matching approach and was using a semantic approach dependent on the meaning of the problem. But his intuitions about how to symbolize this relationship were to place the multiplier (8) next to the letter associated with the larger group (also indicated in Lines 4b and 14.) There is some semblance of reason in this approach as an intuitive symbolization strategy, but the approach is a very literal attempt to compare the relative sizes of the two groups in a static manner. Thus this is labeled the static ‘comparison approach. Also, S2 reads the equations incorrectly—for example, Line 10: (reading 8C = 1£) “for every 8 Chinese, there's one Englishman.”” As Rosnick (Rosnick & Clement, 1980) has pointed out, C appears to be used as a label standing for “Chinese” rather than as a variable standing for an unspecified quantity. Also, the equals sign appears to symbolize a correspondence or association between unequal ‘groups for the student rather than symbolizing an equivalence, and the coefficient 8 acts as an adjective rather than as an operator or another number. Operative approach. The example of a static comparison approach (above) contrasts with the example of $3 solving the Students and Professors Problem correctly in the following transcript: Transcript 2: $3; Students and Professors Problem Student's written work: S = # Students (Transcript Excerpts) 1S: (Writes: § = # Students; P = # Prof) S equals the numberof students and P equals, the number of professors, OK, and there are six times as many students as professors, s0it would be would equal 6 times the number af professors. (Writes: Se) 2 I: How does it fit withthe sentence just knew to write i that way. 3. S: Wel, Lust looked at the sentence andit says 6 times as many students as professors, soe numberof suena oe he punks of profesors meso and pas it would make sense that there would be more Students than professors, 4: Can you se that inthe equation? 5S: Yes, because if you assigned a number to that [P] and multiplied it by six you'd have ‘S equals some number lager than what you asigned for that one [P} i you can, say a litle more about how you Analysis of protocols from such successful students indicates that the key to understanding correct translations lies in the ability to invent an 20 Journal for Research in Mathematics Education operation (such as increasing the number of professors sixfold) that ‘generates an equivalence, and to realize that it is precisely this action that is symbolized on the right side of the equation, § = 6P, as in Line 5 above. It is difficult for a mathematically literate person to imagine how an equation could be viewed otherwise. However, from a naive point of view, the correct equation is strange in the sense that a comparison of two unequal groups must be “force+it” somehow into the notation of an equation showing two equal groups. This analysis exposes some tacit assumptions and meanings underlying our conventions for algebraic notation: $3°s correct equation, S = 6P, does not describe the situation at hand in a literal or direct manner; it describes an equivalence relation that would occur if one were to perform a particular hypothetical operation, namely, making the group of professors six times larger than it really is. Some of the above characteristics of cognitive structures hypothesized to be involved in each type of response are shown diagrammatically above the horizontal line in Figure 2. In this figure, vertical lines represent assimilatory links between these characteristics and the parts of the ‘equation used to symbolize them in the external world. It is hypothesized that the primary internal representation used for the word order matching process is a verbal one, whereas that used for the static comparison and active operation processes involves spatial images of groups of people. ‘The operative approach shown on the right in Figure 2 involves viewing the equation as representing an operation (e.g., the coefficient 6 as a multiplier) on a variable quantity (e.g., an unspecified number of profes- sors) to produce a number equal to another unspecified quantity (e.g., the NcoRRECT connect PROCESS 2 PROCESS Figure 2. Three Types of Solution Processes for Students and Professors Problem, ‘Characteristics of internal cognitive processes are represented above the horizontal ine and fare connected to the associated external writen symbols below the horizontal line January 1982 21 number of students). In this sense students understand the equation as a function (in this case a proportional function) relating two variable ‘Quantities. In expressions such as Sx = 45, the letter x can simply be treated as an unknown but fixed quantity rather than as a variable ‘quantity. Understanding an equation in two variables appears to require an understanding of the concept of variable at a deeper level than that required for one-variable equations. answer by the viable trial-and-error method of plugging numbers into a conjectured equation, the use of this strategy alone does not imply that the subject understands why the equation is correct. Clement, Lochhead, and Soloway (1980) found that significantly more students were able to solve these types of word problems correctly in the context of writing computer programs than in the context of simply writing an algebraic equation. This finding supports the theory that an operative approach is important, since programming languages emphasize the role of operators ‘on numerical variables in assignment statements. Protocol Analysis Results Fifteen freshmen were interviewed solving the Students and Professors Problem. The behavior patterns observed and the number of students showing the pattern in their final answer were as follows: Word Order Matching. (2 incorrect, 0 correct) Ws bols in same order as key words it statements like: “The problem just says it that way.” “You just get it directly from the words.”” The subject may also write down each specific symbol in the equation immediately after reading a corresponding part of the key sentence. Static Comparison Pattern. (5 incorrect, 0 correct) (1) Subject indicates that larger number is placed with letter for larger group to show that it is the larger group (associative symbolization aspect); or (2) subject reads ‘equation as if 6S stands for 6 students instead of 6 times the number of students, as in “65 = P means six students for every professor,” and indicates comprehension of relative size or ratio relationship between the two groups (“variable as label" aspect). Subjects may also make addi- tional comments fitting the word order matching pattern, Operative Pattern. (0 incorrect, 4 correct) Refers to multiplying or dividing the number in one group by a constant in order to make it the same amount as the number in the other group. Substituting: Followed by Operative. (0 incorrect, 2 correct) Substi- tutes numbers into trial equation in order to test it, then explains meaning ‘of equation in terms of operative pattern 's algebraic sym- problem statement and makes 22 Journal for Research in Mathematics Education Not Clear. (0 incorrect, 2 correct) These observable behavior patterns are thought of as separate method- ological entities from the theoretical cognitive approaches hypothesized to explain them. Students who displayed an operative behavior pattern in their final approach solved the problem correctly, and students who displayed a static comparison or word order matching pattern solved the problem incorrectly. Eight of the students shifted between correct and incorrect approaches. Three of these were among the seven students giving an incorrect final answer. Protocols for Problem 6, the Cheesecake Problem, have been longer and more complex, involving several types of errors, including those discussed above (Clement, Narode, & Rosnick, 1981; Rosnick & Clem- ent, 1980) Detailed Analysis of a Protocol ‘We will next examine the protocol of a student who shifts back and forth even more dramatically between a correct and an incorrect ap- proach to the problem. For the Students and Professors Problem, S4 wrote: (Equation 1) 65 = 1P (incorrect) (Equation 2) Si = P (correct) (Equation 3) S = 6P (correct) However, he then rejected Equation 3 and pointed to the incorrect Equation 1 as his final answer. These errors are more striking when we consider the fact that he is an engineering major and was doing B+ work in a standard calculus course at the time of the interview. He was able during the same interview to find the derivative of the function fix) ‘Vxe + 1, rapidly differentiating it by using the chain rule. His transcript follows: Transcript 3: $4; Students and Professors Problem Student's written work: s P s P 6 + 1 Students Professors (Equation 1) 65 = 1P (Equation 2) Si6 = P (Equation 3) § = 6P (Student points to Equation | as his final answer.) Part A 1 I: The problem isto write an equation forthe fllowing statement, and the statement iss There are sx times as many students as professors a this university." January 1982 23 2S: Sixtimes as many students a professors. (writes 6) 3: Soter's ue S for students and for profesor. s ? 4S: Stor students (writes students professors) 5S: You gota ratio of six to one (writes § [') 6 I: Yeah 1S: StoP. . OK, jus ab, 6S equals ah, P, ah, 1P (writes 68 = 12, Equation 1 8 OK. Any parce reason you write it cat way? 9S: Well there's ah there's sh say there's Toone, sk times as many students, Which means it's six students to ane professor. OK, and ah, you ean just write it ‘out—six times as many students as professors 0K, Part B ILS: Now, let me see here. OK, if you wanted to get ike an equal number between the {wo you'd have lke, ah, How ean we do this—S divided by 6—I'm just ying 10 figure out if, ah like if you wanted ab, to figure out the—what do we want 19 do here OK, write that one down, S: So aver 6 equals P (writes Si6 = P, Equation 2), which means that there would be Six'P*for'S” (rites $= 6P, Equation 3), We sai there's times as many students 8s professors, which means if you want to, ah, even out the number of students to professors you'd have to have 6 times as many professors guess that's what 1 ‘was trying (0 think to mysel 14: Can you rephrase that—show me with the equation? 15S: Even out the numberof students—a one-to-one correspondence like—so you have a5 many studeats as you have professors. Part 16 I: So here you have $ and P— 17S: Unvtulh, ‘This means there are six times (points to 6P in Equation 3) as many professors as there ae students (points to ), and this is sx times as many stodents (points to 6S in Equation 1) as there are professors (points to IP). 1 was just gonna ‘even them out, I guess. . So, you have “em like a one-to-one ratio, 0 YOU get S ‘equals P I: So how sbout—which equation would be true forthe ori ‘This one right here (points to 65 = IP) T: The original statement was that there are sx times as many students as professors atthe university. 21S: Right there. 6 equals IP. 22 OK, and then what would ths (points to § = 6P) be for? 23: Six, ah, six times as many professors there are students. Major approaches used by S4 are diagrammed in Figure 3. In Section A of the transcript he wrote the incorrect equation 6S = IP. His reference in Lines 5 and 9 to a “‘ratio of 6 to 1” indicates that a word order matching approach was not the only cognitive process involved during the inter- view, and that he had a semantic conception in mind of a large group of students and a much smaller group of professors. This relationship would presumably be supported by his practical knowledge of a typical universi- ty. Itis more plausible that he used a static comparison approach to think about Equation 1, and that this approach is compelling enough to “win out" eventually over his correct approach in Part B. The equation 6S = IP appears to be an incorrect but meaningful way for him to symbolize 24 Journal for Research in Mathematics Education 8 © r ee [srarewens| Z + Pe ee ao ae ite Figure 3. Parti Model of Cognitive Processes Used by $4 in Transcript 4. Relationships between titernal processes and external Behavior sre’ shown. The Observed “shit in Approach" in Line 20 is accounted for by a corresponding shift in the activity of internal Sohemes. the relative sizes of the two groups—indicating the “base ratio” of six students associated with each professor. This interpretration is consistent with the somewhat unusual way in which he included the number I in the equation In Part B of the transcript he was able to generate a correct translation, apparently by shifting toa different approach temporarily. In Lines 11-13, he wrote a correct equation, S16 = P, saying, “If you wanted to get like an equal number between the two, you'd have like, ah, how can we do this— S divided by 6 . . . (writes 5/6). . .if you want to even out the number of students to professors.”” We infer that S4 is mentally performing an action of dividing the group of students into six parts here in order to obtain a one-to-one correspondence between a new group of students and the group of professors. Because it involves an internalized action, this operative approach contrasts with his previous static approach (see Figure 2). In Section C, however, after rewriting his correct equation as $= 6P (Equation 3), $4 reverted to his initial incorrect approach. He read the equation § = 6P in Lines 17 and 23 as meaning “six times as many professors as there are students,” and he therefore rejected it. We infer that the static comparison approach was a more powerful idea within him and dominated his second operative approach by “taking over” again just before Line 17. Ina later session, when he was asked to draw a picture to illustrate the equation 65 = IP, he drew, from right to left, one circle with a P in it, an January 1982 28 ‘equal sign, and six circles with S’s in them. (See Transcript 4 discussed below.) These responses are further evidence that S4’s incorrect equation is not simply based on a syntactic or word order matching strategy, but that he sees it as a reasonable way of symbolizing his semantic conception of the situation. (Transcript 3 was collected early in the study, and from hindsight, the interviewer's technique in Line 3, where he says, “Let's use 5 for students and P for professors” can be criticized. Although this is a common way of saying “Use S for the number of students,” it might be interpreted as “Use 5 for one student.”” The latter interpretation might contribute to the reversal error. However, this was judged not to be a predominant factor in this protocol on the basis of the depth of S2's explanations supporting his two views, and the persistence of his views in a second interview (Transcript 4), and on another similar problem.) Transcript 4: $4; Students and Professors Problem, Session 2 Ina follow-up interview with S4, the interviewer decided to try a simple tutoring strategy to see whether the student would easily change his method of writing equations. The interviewer recalled the Students and Professors Problem and had the student write down the equations 65 = P and 5 = 6P. 1 I: Lthink we have to decide—between those [65 = P and 5 = 6P]. Now, Idlke you to raw-—a- picture. So, is there any way you could represent with circles oF Something—the numberof professors and the number of students, jst the relative OK. you said there were 6 times as many students as there was professor. Rise So, say there was one professor (draws circle with Pinsie).'l puta P inside there (rites "=" next to), and we sad there's 6 stidens for every professor, +0 ¥e ‘Would just correspond that by puting 6 circles, all with S'sin them. (Draws circles to lft of equal sign, putting an Sin each.) 5 10K. Student's written work: 00000 -O ©- O00000 6S: And that would seem to say the same thing (as 6S = Py; whereas this one (points to 5 6P) would be just the opposite as—studens and professors. (Draws sesond row" of circled leitrs.) That would be a prety good ratio ina, i9 a school — 7: Because there would be so many (S: Professors, yeah) teachers for each student? (OK, now the question is which one do you think is the correct one? 8S: This one (points tothe irs equation, 65°~ ‘This transcript supports the hypothesis that his incorrect equation 6S 26 Journal for Research in Mathematics Education IP is understood by him in terms of a semantic, static comparison approach. Diagramming thought processes. The diagram in Figure 3 shows a method for representing and tracking the involvement over time of different mental schemes used by S4 in Transcript 3. This method of diagramming is described in Witz and Easley (in press) and in Easley (1979), and developed further in Clement (1981). As time moves from left to right, a static comparison scheme, an active operation scheme, and an algebraic rule scheme each become active in turn for a period of time. Here it is assumed that each of these is an action-oriented scheme in the Piagetian sense, constituting a stable knowledge structure that is capable of assimilating perceptual aspects of a situation, controlling an action, anticipating certain results of the action, and dominating the student's point of view of the situation for a time. Below the horizontal line in Figure 3 are shown external objects perceived by the student and observable responses produced by the internal schemes. The observed answer change in Line 11 is then accounted for by an internal shift where the active operation scheme becomes dominant. However, the static comparison scheme remains subliminally active (shown by the dotted horizontal line) and regains dominance in Line 17 (shown by the solid horizontal line) after being retriggered by the perception of the equation S = 6P. The jagged line over Line 17 represents the disequilibrium condition created by the mismatch between this correct equation and the student’s static comparison scheme, a condition which leads to the rejection of this equation and a return to favoring the incorrect equation. Heuristic value of diagrams. The diagram, then, represents a low- detail, macrolevel model of the sequence of cognitive processes occurring in $4 over time in this protocol. It illustrates a first-order explanation of the major unobservable events occurring in S4’s working memory and ties them explicitly to the major observable events to be explained. Such a diagram undoubtedly paints an oversimplified picture. But it does serve a heuristic function in making cognitive explanations more explicit and visualizable, and more amenable to criticism and elaboration. Although this diagram appears near the end of this report, it played an important role early in this study. Attempts to account for S4’s puzzling protocol through diagrams led to the first hypothesis of the existence of a static comparison strategy as an incorrect but meaning-based approach. Implications of Shifts Between Approaches We have seen in S2 and S4 two examples where students seem to have two different ways of approaching a problem and where they move back and forth between the approaches. We call this evidence for “shifting between approaches” and hypothesize that it reflects an unobservable January 198227 internal process of shifting between cognitive schemes used to deal with the problem. This provides evidence for the notion that human cognition is not always based on consistent processes—schemes that lead to contradictory results can apparently exist fairly autonomously and inde- pendently in the same individual. One scheme may become active and dominate for a time, only to be superseded by another. This is not so difficult to imagine, because we have all had the experience of being “of two minds” on a subject. The latter is a conscious experience, however, and in some protocols it is not clear that the subjects are consciously aware of the inconsistencies between the processes they are using. Evidence for shifts between approaches was observed in 8 of the 15 interviews, and they may be more common in harder problems, such as the Cheesecake Problem. Rosnick and Clement (1981) interviewed six students solving the Cheesecake Problem after they were shown an explanation of a worked example of the Students and Professors Problem, and found that although all six eventually gave a correct answer, four of them shifted between approaches, and five of them showed serious misunderstandings or misgivings about their answer. ‘There is an important implication of the theory that contradictory schemes for viewing problems can exist in people. The theory implies that students who ate successfully taught a standard method for a ‘mathematical skill may still possess intuitive, nonstandard methods that can compete for control. In other words, learning a new scheme or method does not guarantee that old ones have been unlearned or ‘modified. This can be a good or bad thing, depending on whether the new scheme or method is more satisfying and useful to the individual in all situations. Inconsistent, semiautonomous schemes may be an important factor in a wide variety of educational problems (Clement, Narode, & Rosnick, 1981). Summary Data from group testing indicate that a significant number of college students produce reversal errors in formulating algebraic equations of this kind. Protocol data indicate that the error has two main sources: a syntactic, word order matching process, and a semantic, static symboliza- tion process. The second error-producing process appears to be a fairly deep-seated, intuitive symbolization strategy that can cause students to reject a correct solution they have just generated. Such shifts between approaches indicate that contradictory schemes may continue to exist independently in the same individual. This implies that teaching a student a standard method is no guarantee that the student's own intuitive method will not “take over” in a later problem-solving situation. The protocol analyses also indicate that two key ideas in successful solutions are (a) remembering that variables stand for numbers rather than objects in these 28 Journal for Research in Mathematics Education problems and (b) being able to invent a hypothetical operation on the variables that creates an equivalence. ‘Thought process diagrams of mental schemes interacting over time ‘were an important tool for generating hypotheses in this study. Further research is needed to evaluate the results from the exploratory interviews discussed here and to develop a more detailed theory of the cognitive processes occurring in the solutions. ‘The fact that so many engineering majors cannot formulate or read such simple equations leads to concern about the extent to which students understand how equations are used to symbolize meanings. We suspect that such errors are part of a larger pattern in which schools have been more successful in teaching students to manipulate equations than they have in teaching students to formulate equations in a meaningful way. The identification of these conceptual stumbling blocks using protocol analy- sis should make it easier to design instructional strategies to overcome them. [REFERENCE NOTES 1. Wagner, S, Conservation of equation and function and its relationship to formal ‘operational thought. Paper presented at the American Educational Research Assocla- ton, New York, April 1977 2. Galvin, & Bell, A. Aspects of dificultes in the solution of problems involving the {formulation of equations. Unpublished manuscript, Nottingham, U.K: Nottingham Skil (Centre for Mathematical Education, University of Notingham, 1977, 3. Matz, M. Towards a process model for high school algebra errors. (Working Paper 181) ‘Cambridge, Mass. Artificial Intelligence Laboratory, Massachusetts Institute of Tech: nology. 1979, 4. Kaput, J. Personal communication, June 1980. 5. Clement, J., Lochhead, J., & Soloway, E. Translating between symbol systems: Isolating @ common dlifculty in solving algebra word problems. Unpublished man Script, Amberst, Mass.: Cognitive Development Project, Department of Physics and Astronomy, University of Massachusetts at Amherst, 1979. REFERENCES ‘Clement, J. Cognitive microanalysis: An approach to analyzing intuitive mathematical Feasoning processes, Ia 'S. Wagner & W. Geeslin (Eds.), Modeling mathematical cognitive development, Columbus, Ohio: ERIC Center for Science, Mathematics, and Environmental Education, 1981 (Clement, J., Lochhead, J., & Monk, G. Translation dificulies in earning mathematics. “American Mathematical Monthy, 1981, 8, 286-290 Clement, J., Lochhead, J., & Soloway, E; Positive effects of computer programming on ‘students" understanding of variables and equations. Proceedings of the Annual Confer- fence ofthe American Society for Computing Machinery, 1980, 867-75 ‘Clement, J., Narode, R., & Roshi, P. Intuitive misconceptions in algebra asa source of, math anxiety. Focus on Learning Problems in Mathematics, 1981, 3G), in press. Collis, K. Operational thinking in elementary mathematics. In J. Keats, K. Collis, & C. Halford (Eds), Cognitive development. New York: Wiley, 197. Easley, J. A. The structural paradigm in protocol analysis. In J. Lochhead & J. Clement (48), Cognitive process insiruction, Philadelphia: Franklin Institute Press, 1979. January 1982 29 Herseovies, N. Kieran, C. Constructing meaning for the concept of equation. Mathemat- ‘cs Teacher, 1980, 73, 572-580. Kachemann, D. Children's understanding of numerical variables. Mathematics in School, 1978, 7), 23-26. Paige, J., & Simon, H. Cognitive processes in solving algebra word problems. In B. Keinmuntz (E4.), Problem solving research, method, and theory. New York: Wiley, 1966 Rosnick, P. Some misconceptions conceming the concept of variable. Mathematics ‘Teacher, 1981, 74, 418-420, 480 Rosnick, ., & Clement, J. Learning without understanding: The effect of tutoring strategies ‘on algebra misconceptions. Journal of Mathematical Behavior, 1980, 3(1), 329 Witz, K., & Easley, J. A, New approaches to cognition. In L. Van den Daele, J. Paseual- Leone, & K. Witz (Eds. Neo-Plagetian perspectives on cognition and development New York: Academic Press, in press. [Received May 1980; revised May 1981] NCTM EDUCATIONAL MATERIALS 1908 aSéocison Bee Reston, VA 22081 (Fos) 6203840 SUBTOTALS ess 20% JOURNAL FOR RESEARCH IN MATHEMATICS. EDUCATION. TOTALS: [DPaymentioNCTM in US tunasenclosed Mastercard visa Credit Card & Signature Name Address cy 30 Expiration Date Membership # 2. State — or Poa Case Journal for Research in Mathematics Education

You might also like