DER-CHING YANG

TEACHING AND LEARNING NUMBER SENSE – AN INTERVENTION STUDY OF FIFTH GRADE STUDENTS IN TAIWAN

ABSTRACT. Two classes (one experimental and one control) in a public elementary school located in southern Taiwan participated in this study. Number sense activities were conducted in the experimental class as supplementary teaching materials, while the control class followed the standard mathematics curriculum. Data indicate that there are statistically significant differences between pretest and posttest (pretest and retention-test) scores for the experimental and control classes at the 0.01 level. The scores for the experimental class increased 44% after instruction (the mean score went from 12.35 to 17.81), while the scores for the control class increased only 10% after instruction (the mean score went from 11.29 to 12.42). Compared to the control class, the experimental class made much more progress on number sense tests. Results indicate that students in the teaching class (not including the students in the low level) advanced in their use of number sense strategies when responding to interview questions. The data demonstrate that the teaching of number sense activities, executed in the experimental class, is effective in developing children’s number sense. Furthermore, the results of retention demonstrate that the students’ learning was meaningful and significant. KEY WORDS: benchmarks, control class, estimation, experimental class, number sense

R ATIONALE AND P URPOSE The teaching and learning of number sense is considered to be a major topic in international mathematics curricula (Anghileri, 2000; Australian Education Council, 1991; Cockcroft, 1982; Japanese Ministry of Education, 1989; Markovits & Sowder, 1994; McIntosh, Reys, Reys, Bana & Farrel, 1997; National Council of Teachers of Mathematics (NCTM), 1989, 2000; National Research Council, 1989). The Curriculum and Evaluation Standards for School Mathematics highlighted that the teaching of number sense is an essential goal of a school’s mathematics curriculum. Furthermore, the Number and Operations Standard of Principles and Standards for School Mathematics (PSSM) (NCTM, 2000) states that “central to this Standard is the development of number sense” (p. 32). Over the past decade, several research studies in Taiwan have focused on fractional concepts and children’s conceptions of numbers (Lin, 1989;
International Journal of Science and Mathematics Education 1: 115–134, 2003. © 2003 National Science Council, Taiwan. Printed in the Netherlands.

116

D.-C. YANG

Ning, 1992), but research related to number sense has received little attention. During the past two years, due to its emphasis in the NCTM Standards, number sense has stimulated a growing amount of attention and research in Taiwan. There is currently an important mathematics educational revolution in Taiwan. The guidelines for a Nine-Year Joint Mathematics Curricula Plan (Ministry of Education in Taiwan, 2000) stresses that the learning and teaching of mathematics should highlight meaningful connections with real life. The new guidelines for the mathematics curricula plan and number sense both focus on meaningful learning. If important reforms of mathematics education in Taiwan are to be made, the development of number sense should be integrated into the mathematics curriculum. The purpose of this study was to report the results of an intervention study conducted on Taiwanese fifth graders. The related research question is: Can number sense be fostered by appropriate teaching?

BACKGROUND What Is Number Sense? Number sense refers to a person’s general understanding of numbers and operations and the ability to handle daily-life situations that include numbers. This includes the ability to develop useful, flexible, and efficient strategies (i.e., mental computation or estimation) for handling numerical problems (Howden, 1989; McIntosh, Reys & Reys, 1992; Reys, 1994; Reys & Yang, 1998; Sowder, 1992a, 1992b; Treffers, 1991; Yang, 2002a, 2002b). What Are the Number Sense Components? Since number sense has been an important topic in mathematics education, it has produced much research and discussion among mathematics educators, cognitive psychologists, researchers, teachers, and mathematics curricula developers (Howden, 1989; Greeno, 1991; Markovits & Sowder, 1994; McIntosh et al., 1992; NCTM, 1989, 2000; Reys, 1994; Reys & Yang, 1998; Sowder, 1992a, 1992b; Yang, 2002a, 2002b). The researcher generalized the above research reports and defined the number sense components as following: (1) Understanding the basic number meanings: This implies making sense of numbers and developing a conceptual understanding of numbers.

TEACHING AND LEARNING NUMBER SENSE

117

(2) Recognizing the magnitude of numbers: It includes the ability to compare numbers (whole numbers, fractions, decimals, and so on), to order numbers correctly, and to recognize the density of numbers. (3) Using benchmarks appropriately: It includes the ability to develop and flexible use the benchmarks, such as 1, 1/2, 100, and so on, in different situations. (4) Understanding the relative effect of operations on numbers: It includes the ability to identify how the different operations affect the result of numerical problems. (5) Developing different strategies appropriately and assessing the reasonableness of an answer: This implies developing different strategies (i.e., estimation, mental computation) to solve problems appropriately and knowing that the result is reasonable. The above five number sense components constitute the basis for this study. The Role of Number Sense in the Taiwanese Mathematics Curriculum A major reform of Taiwanese mathematics curriculum standards from grade 1 to 9 has been underway since 2001. In 2000, the Guidelines for a 9-Year Joint Curricula Plan (Guidelines (Ministry of Education in Taiwan, 2000)) was released and covered all subjects, including mathematics. Number sense has been considered an important topic in school mathematics curriculum (Anghileri, 2000; Japanese Ministry of Education, 1989; NCTM, 1989, 2000; National Research Council, 1989). However, the topic of “number sense” has not been integrated into the new mathematics curricula. If significant improvements in mathematics are to be made in Taiwan, the development of number sense should be considered as an important topic in the new curricula. The new Guidelines publication does not provide a standard national policy for the teaching and learning of number sense. However, the new standards do emphasize that mathematics teaching and learning should focus on making connections between mathematics and daily life situations. Significantly, the goals of the Guidelines for mathematics curricula and the spirit of the development of number sense are consistent. Even though it lacks a clear policy on the teaching and learning of number sense, the Guidelines emphasizes many of the fundamental principles of number sense. Research Studies Related to Teaching and Learning of Number Sense Several research projects ( Cobb, Wood, Yackel, Nicholls, Wheatly, Trigatti & Perlwitz, 1991; Treffers, 1991; Warrington & Kamii, 1998) have

118

D.-C. YANG

demonstrated that students who participate in well-designed activities are more likely to develop number concepts than students who receive instruction focusing on the development of standard written algorithms and computational proficiency. Several studies (Anghileri, 2000; Markovits & Sowder, 1994; Yang, 2002b; Yang & Reys, 2001a, 2001b) have demonstrated the effectiveness of instruction designed to promote number sense. The study of Cramer, Post & delMas (2002) showed that students in the RNP can develop fractional number sense because “they spent their time interacting with fraction ideas in multiple ways and were provided extended periods of time to develop an understanding of the meaning of symbols” (p. 140). Yang’s study (2002a, 2002b) further demonstrated that sixth graders’ number sense can be developed through process-oriented teaching models.

M ETHOD Sample A public school located in a city in southern Taiwan was selected to participate in this study. Two classes (an experimental class and a control class) were studied. Number sense activities were conducted in the experimental class as supplementary teaching materials, while the control class followed the standard mathematics curriculum. The school, which comprised about 3000 students, serves children from diverse areas. The students in the study come from families with a wide range of occupations, incomes, and educational levels. The experimental class consisted of 37 students (20 boys and 17 girls), and the control class consisted of 38 students (20 boys and 18 girls). Based on the students’ performance on the pretest, students in each class were divided into the following three levels: High – top 10%; Middle – 40–60%; Low – bottom 10%. Two students were randomly selected from each level and interviewed before instruction, after instruction, and four months after the study. Therefore, the sample consisted of 12 students, including the low- (TL12, CL12), middle- (TM12, CM12), and high- (TH12, CH12) level students. Instructional Activities This study defined the five number sense components based on the existing number sense related research reports and documents as described earlier. The instructional activities were designed according to the above number sense framework. Five units of number sense activities were designed for

TEACHING AND LEARNING NUMBER SENSE

119

the use of the teacher. Each unit included 5 lessons. For example, one of the estimation lesson was as following: 1. (a) Please estimate the area and perimeter of the playground in our school. Explain how you reached your answer. (b) Estimate how many students can be contained in our school’s playground. Explain how you reached your answer. Assessment Instruments Assessment instruments included paper-and-pencil tests and interviews. Paper-and-Pencil Tests included Pretest, Posttest, and Retention test, all of which used the “Number Sense Rating Scale (NSRS)” designed by Hsu et al. (2001). The test items in the NSRS based on the related research studies McIntosh et al. (1992, 1997). The NSRS was given in multiple choice format and included 37 items. The reliability test of internal consistency for the NSRS was 0.889. The test-retest reliability coefficient for the NSRS was 0.894 at the 0.01 level. The NSRS also has high content validity, specialist validity, and construct validity (p. 368). The test was given for a period of about 35 min. There are clear explanation about NSRS in the study of Hsu, Yang & Li (2001). Interview Instruments. Pre-interview, Post-interview, and Retention-interview questions were designed by the researcher. All of the three tests used the same items. Each interview included 15 questions designed to investigate whether or not students could effectively use the five number sense components as described in the number sense framework. Students were given as much time as they needed to respond to each question, and each interview lasted about 60 min. All interviews were video-taped and later transcribed. The schedule for instructional units and assessment instruments is presented in Table I. Analysis The NSRS included 37 items. Each item was given 1 point if correct, no point if incorrect. No partial credit was awarded. Therefore, the total possible score for the test is 37 points. Interview. The students’ responses were examined and sorted. In an effort to identify the different strategies used by students, each response was coded (as correct or incorrect) according to one of the following three categories:

120

D.-C. YANG

TABLE I Schedule for instructions and assessments Date 02/01/01–08/30/01 08/30/01–09/08/01 09/10/01–12/31/01 01/01/02–01/18/02 01/19/02–02/17/02 02/18/02–06/01/02 06/01/02–06/12/02 Review the related papers and design the number sense project Pretest and pre-interview for both Exp- and Con-class Number sense activities integrated into Exp-class Posttest and post-interview for both Exp- and Con-class Winter break No number sense activities were integrated into classes Retention-test and retention interview for Exp- and Con-class

• Number sense based – strategies that utilized one or more components of number sense (i.e., benchmarks, number magnitude, relative effect of operations on numbers). • Rule based – applied the rules of standard written algorithms but was unable to go beyond the direct application rule in the explanation. • Could not explain – despite probes and queries by the interviewer, clear explanations could not be obtained. The researcher and the teacher used the transcripts to categorize responses to each correct or incorrect answer. Another teacher served as second coder by using the same transcripts and categorizing the responses independently. These initial reviews produced categorization agreement on more than 90% of the students’ responses. The remaining responses were reexamined and discussed by both coders until a consensus was reached.

H OW W ERE THE N UMBER S ENSE ACTIVITIES I NTEGRATED INTO THE M ATHEMATICS C LASS ? Reys (1994) advocated providing a class with process-oriented activities and establishing a classroom environment that encourages meaningful discussion, exploration, thinking, and reasoning is the best way for students to develop number sense. Several research studies (Anghileri, 2000; Reys, 1994; Fraivillig, 2001) further supported the idea that the effective teaching of mathematics should focus on the learning process. This process not only highlights small-group problem solving and whole class discussions to develop mathematical thinking and reasoning, but also encourages students to defend, query, and prove their ideas. In order to effectively help children

TEACHING AND LEARNING NUMBER SENSE

121

Figure 1. Process-oriented teaching model.

develop number sense, the researcher designed the process-oriented teaching model (Figure 1) to help the teacher effectively integrate number sense activities into mathematics class.

122

D.-C. YANG

R ESULT Table II reports the results of mean scores (correct percent), standard deviations, and t-test results for the experimental class and the control class. The data indicate that there are statistically significant differences between pretest and posttest (pretest and retention-test) scores for both classes. However, the scores for the experimental class increased 44% after instruction (the mean score went from 12.35 to 17.81), while the scores for the control class increased only 10% after instruction (the mean score went from 11.29 to 12.42). Compared to the control class, the experimental class made much more progress on number sense tests. This indicates that the teaching of number sense is highly effective. The retention test results further indicated that the teaching experiment was meaningful. Tables III and IV report the results of students’ responses in the experimental class and the control class on pre-, post-, and retention-interviews. The results indicate that students in the teaching class (except for the students in the low level) progressed much more in their use of number sense strategies when responding to interview questions. The results of the retention-interview further show that students’ learning was meaningful and therefore long-lasting. However, students’ responses in the control class did not advance in the use of number sense strategies after the instruction. In order to highlight the differences in the development and use of number sense strategies between the experimental class and the control class, several interview questions were presented before instruction, after instruction, and four month after the study.
TABLE II The mean scores (correct percent), standard deviations, and t-tests results of number sense pretest, posttest, and retention tests for both experimental and control classes Class Test Pretest Posttest Pretest Retention Pretest Posttest Pretest Retention Mean score 12.35 17.81 12.35 19.08 11.29 12.42 11.29 13.23 (Correct%) (33.4%) (48.1%) (33.4%) (51.2%) (30.5%) (33.6%) (30.5%) (35.9%) SD 4.83 5.29 4.83 5.46 4.09 3.94 4.09 4.35 T-value −10.949∗ −13.218∗ −3.775∗ −4.734∗

Exp-classes (N = 37)

Con-classes (N = 38)

Note. ∗ P < 0.01.

TEACHING AND LEARNING NUMBER SENSE

123

TABLE III The results of students’ responses in the experimental class (and control class) on pre-, post-, and retention-interviews TH1 (CH1) Pre-interview Correct NS-based Rule-based Couldn’t Ex Incorrect NS-based Rule-based Couldn’t Ex Post-interview Correct NS-based Rule-based Couldn’t Ex Incorrect NS-based Rule-based Couldn’t Ex Retention-interview Correct NS-based Rule-based Couldn’t Ex Incorrect NS-based Rule-based Couldn’t Ex TH2 (CH2) TM1 (CM1) TM2 (CM2) TL1 (CL1) TL2 (CL2)

6 (9) 2 (4) 1 (0) 0 (0) 0 (2) 6 (0)

4 (5) 5 (3) 2 (1) 0 (0) 0 (3) 4 (3)

4 (4) 2 (0) 2 (2) 0 (0) 0 (4) 7 (5)

2 (3) 3 (1) 2 (2) 0 (0) 2 (4) 6 (5)

2 (3) 0 (0) 4 (2) 0 (0) 0 (3) 9 (7)

2 (3) 0 (1) 3 (2) 0 (0) 0 (3) 10 (6)

15 (10) 0 (4) 0 (0) 0 (0) 0 (1) 0 (0)

11 (6) 3 (3) 0 (4) 0 (0) 0 (0) 1 (2)

10 (5) 2 (3) 0 (2) 0 (0) 1 (2) 2 (3)

12 (5) 1 (2) 0 (2) 0 (0) 1 (2) 1 (4)

5 (3) 0 (3) 1 (1) 0 (0) 1 (3) 8 (5)

5 (3) 0 (4) 3 (1) 0 (0) 0 (1) 7 (6)

15 (11) 0 (3) 0 (1) 0 (0) 0 (0) 0 (0)

12 (6) 1 (3) 1 (4) 0 (0) 0 (0) 1 (2)

11 (5) 0 (0) 1 (5) 0 (0) 0 (1) 3 (4)

12 (4) 2 (1) 0 (3) 0 (0) 0 (2) 1 (5)

5 (3) 0 (4) 4 (3) 0 (0) 0 (0) 6 (5)

5 (2) 1 (5) 2 (2) 0 (0) 1 (0) 6 (6)

Notes. 1. Each interview included 15 questions. 2. NS- (rule)-based represents number sense- (rule)-based method; Couldn’t Ex means couldn’t give clear explanation.

Table IV presents a summary of results from the fifth grader’s responses in the Exp- and Con-class on the pre-, post-, and retention-interview about question 1. This question focused on investigating the students’ conceptual understanding of fractions. Data indicated that only one student could use

124

D.-C. YANG

TABLE IV Results from fifth grade students’ responses to the following problem: 1. (1) (2) (3) (4) Two same sized watermelons were shared by five students equally. How many watermelons can each student have? Why? Greater than a half of a watermelon Equal to 1 of a watermelon 5 Equal to 2 of a watermelon 5 Greater than 2.5 watermelons Exp-A PreCorrect NS-based Rule-based Couldn’t explain Incorrect NS-based Rule-based Couldn’t explain PostRetention PreCon-B PostRetention

1 (17%) 2 (33%) 1 (17%) 0 (0%) 0 (0%) 2 (33%)

6 (100%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%)

6 (100%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%)

2 (33%) 0 (0%) 1 (17%) 0 (0%) 1 (17%) 2 (33%)

3 (50%) 1 (17%) 1 (17%) 0 (0%) 0 (0%) 1 (17%)

3 (50%) 1 (17%) 1 (17%) 0 (0%) 0 (0%) 1 (17%)

Notes. 1. Percentages may not add to 100 because of rounding. 2. Each class had six students who were interviewed.

the number sense-based method to answer question 1. For example, (Preinterview of TH2): R: Can you tell me the answer and your reasons about question 1? TH2: One watermelon was shared by five persons, so each one got 1 wa5 termelon. Since there are two watermelons, each person can get 2 5 watermelon. TH2 gave the correct answer and reasonable explanations, so this response was coded as Number sense-based method. Two (33%) students gave the correct answer, by using the rule-based method. For example, (Pre-interview of TM2): TM2: 2 ÷ 5 = 2 , so the answer is 2 ÷ 5 = 2 . 5 5 R: Can you tell me why the answer is 2 ? 5 TM2: Because 2 ÷ 5 = 2 , so the answer is 2 ÷ 5 = 2 . 5 5 R: Can you explain it in a different way? TM2: I don’t know.

TEACHING AND LEARNING NUMBER SENSE

125

Since TM2 used the rule “2 ÷ 5 = 2 ”, but could not give an explanation, 5 this response was coded as “ruled-based”. Another student produced the correct answer but also couldn’t explain it (Pre-interview of TH1): TH1: The answer is 2 . 5 R: Can your tell me your reasons? TH1: I don’t know how to explain it. This was coded as “correct and could not explain.” Two students gave incorrect answers and also could not explain them (Pre-interview of TL1): TL1: The answer is 1 . 5 R: Can your tell me your reasons? TL1: I don’t know. During the post- and retention-interview, all six students in the experimental class could use the number sense-based method to explain their reasoning. In addition, four of the six students could support their answers with pictorial representations. For example, (Post-interview of TM2): R: Can you tell me your answer and reasons about question 1? TM2: As you can see the picture, each watermelon can be divided into 5 parts. Each person can have 1 of one watermelon. Since there are 5 two watermelons, each one can get 2 watermelons. 5 In Con-class, the students’ responses showed not much difference on the pre-, post-, and retention-interviews. Two (33%) students used the number sense-based method to answer question 1 on the pre-interview, however, only three (50%) students could apply the number sense strategy to explain their reasons on the post- and retention-interviews. One student was able to give the correct answer, yet could not explain why on the pre-, post-, and retention-interviews. For example, (pre-, post-, and retention-interviews of CM2): R: Can you tell me your answer and reasons about question 1? CM2: 2 . I don’t know how to explain it. 5 One student gave an incorrect answer on pre-, post-, and retention-interviews. CL2 answered: “ 1 ” and “I don’t know how to explain it.” 5 The data indicate that students in the Exp-class more frequently used number sense strategies during the post and retention-interviews (100% vs. 17%), and were more successful than students in the control class (100% vs. 50%). Table V presents a summary of the results from fifth grader’s responses for question 6. This question focused on investigating whether or not students could use benchmarks. It is a more challenging question than the

126

D.-C. YANG

TABLE V Results from fifth grade students’ responses to the following problem: 6. Without calculating, select the best estimate for 4 + 6 ? Why? 7 5 (1) 12 (2) 10 (3) 2 (4) 1 (5) Without calculating can’t find the answer Exp-A and B PreCorrect NS-based Rule-based Couldn’t explain Incorrect NS-based Rule-based Couldn’t explain PostRetention PreCon-A and B PostRetention

0 (0%) 1 (17%) 0 (0%) 0 (0%) 3 (50%) 2 (33%)

3 (50%) 0 (0%) 1 (17%) 0 (0%) 1 (17%) 1 (17%)

4 (66%) 0 (0%) 0 (0%) 0 (0%) 1 (17%) 1 (17%)

0 (0%) 1 (17%) 1 (17%) 0 (0%) 2 (33%) 2 (33%)

0 (0%) 2 (33%) 1 (17%) 0 (0%) 1 (17%) 2 (33%)

0 (0%) 2 (33%) 1 (17%) 0 (0%) 1 (17%) 2 (33%)

above question. The data indicate that no student in either class could use the benchmark on the pre-interview. One student in each class applied the rule-based method. For example, (pre-interview of TH2): R: TH2: R: TH2: Can you tell me the answer for question 6 and your reasoning? I think the answer is 2. Can you tell me your reasons? Because 28 30 58 23 4 6 + = + = =1 , 5 7 35 35 35 35 the answer is close to 2. R: Can you do it by another way? TH2: (Hesitate for a while!) I don’t know. R: How did you know to find the sum of two fractions with different denominators? TH2: I learned it at private school. This kind of question (find the sum of two fractions with different denominators) is not taught until the seventh grade in the Taiwanese new mathematics curricula. However, TH2 had learned this way at a private school. Three (50%) students in the Exp-class and two (33%) students in the Con-class gave incorrect answers based on rule-methods. For example, (Pre-interview of TH1):

TEACHING AND LEARNING NUMBER SENSE

127

TH1: R: TH1: R: TH1: CM1: R: CM1: R: CM1:

The answer is 10. Why? Please tell me your method. Because 6 + 4 is equal to 10. Can you do it another way? (Hesitating for a while!) I don’t know. 1. Why? Can you tell me your reasons? 4 + 6 = 10 , hence the answer is close 1. 5 7 12 Can you do it another way? (Shaking her head!) I don’t know.

Or (Pre-interview of CM1)

These methods were coded as incorrect and rule-based. Two students in each class gave incorrect answers and no explanations. For example, (Preinterview of TL1): TL1: The answer is about 1. R: Can you tell me your reasons? TL1: I guess. Or (Pre-interview of CL2) CL2: I don’t know. Three (50%) students in the post-interview and four (66%) students in the retention-interview used benchmarks to solve this question. For example, (Post-interview of TH1): TH1: 2. R: Can you tell me your reasons? TH1: Because 4 is less than 1 and near 1, 5 Therefore, the sum is about 2. Or (Post-interview of TH2) TH2: R: TH2: R: TH2: R: TH2: The answer is 2. Why? Can you tell me your reasons? Since 4 is near 1 and 6 is also near 1, the sum is near 2. 5 7 Can you do it by different way? I can find the common denominator and compute it. Can you tell me what’s the difference between these methods. I can save time, so I don’t need to spend a lot of time to do complex computation. It is also more meaningful and useful to me.

6 7

is also less than 1 and near 1.

However, no student in the control class could use 1 as a benchmark. In the Con-class, two (33%) students in the post- and retention-interview applied

128

D.-C. YANG

TABLE VI Results from fifth grade students’ responses to the following problem
7 7 8. Without calculating, which one 10 or 12 is larger? Why? 7 7 (1) 10 (2) 12 (3) same (4) Without calculating can’t find the answer

Exp-A PreCorrect NS-based Rule-based Couldn’t explain Incorrect NS-based Rule-based Couldn’t explain PostRetention Pre-

Con-B PostRetention

1 (17%) 1 (17%) 1 (17%) 0 (0%) 0 (0%) 3 (50%)

4 (66%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 2 (33%)

4 (66%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 2 (33%)

0 (0%) 2 (33%) 2 (33%) 0 (0%) 0 (0%) 2 (33%)

0 (0%) 4 (66%) 1 (17%) 0 (0%) 0 (0%) 1 (17%)

0 (0%) 4 (66%) 1 (17%) 0 (0%) 0 (0%) 1 (17%)

written algorithms. They found the common denominator first and got the exact answer, then decided on the answer. Two (33%) students continued to use the same method before and after instruction. For example, (Postinterview and retention-interview of CL2): CL2: I don’t know. As shown in Table VI, students in the Exp-class were not only more successful on the post- and retention-interviews than the pre-interview in applying benchmarks (50% vs. 0%) and (66% vs. 0%), but they also used benchmarks more frequently than students in the Con-class on the postand retention-interviews (50% vs. 0%) and (66% vs. 0%). Table VI presents data from the fifth grader’ responses for the Exp7 and Con-class on comparing fractional size: without calculating, is 10 or 7 is larger? The fifth graders’ mathematics textbooks in Taiwan do not 12 teach children to find a least common denominator or cross-product procedure for comparing fractions. The instruction for the Exp-class focused on developing conceptual understanding through pictorial and verbal representations. The textbooks used in the Con-class taught children to compare fractions by using concrete materials. For example, when comparing 3 and 8 5 , the teacher gave students 24 pieces of chips as a unit and taught them 12 5 5 5 to find the pieces of 3 and 12 . 24 × 3 = 9; 24 × 12 = 10, therefore, 12 is 8 8 larger because it equals 10 pieces of chips. On the pre-interview, one student (17%) in the Exp-class responded: “A pizza was cut into 10 pieces and the other pizza was cut into 12 pieces.

TEACHING AND LEARNING NUMBER SENSE

129

Each piece in the 10 pieces group is larger than in the 12 piece group, so 7 7 is larger than 12 .” This was coded as number sense-based. No student 10 in the Con-class could apply a number sense-based method. One student in the Con-class memorized the rule but was unable to explain his reason. For example, (Pre-interview of CH1): R: Can you tell me your answer and reasoning for question 8? CH1: Since the numerators are the same, and one of the denominators is smaller, then that fraction is larger. R: Why is the fraction larger, if the denominator is smaller? CH1: I don’t know. I memorized the rule I learned in mathematics class. One student in each class gave the correct answer but they could not provide an explanation. This was coded as correct and could not explain. Three (50%) students in the Exp- and two (33%) students in the Conclass gave incorrect answers and incorrect explanations. For example, (Preinterview of TL2):
7 7 TL1: 12 > 10 . R: Why? Can you tell me your reason? TL1: Because the numerators are same, I compared the denominators. 7 7 The 12 is larger than 10, hence 12 > 10 .

Or (Pre-interview of CL2) CL2: I don’t know. On the post- and retention-interviews, four (66%) students in the Expclass applied the number sense methods. One used a meaningful way and three used pictorial representations to help solve the problem. For example, (Post-interview of TH1 and TM2): R: Can you tell me your answer and reasoning for question 8? TH1: A pizza was cut into 10 pieces and another was cut into 12 pieces. Each piece in the 10 piece group is larger than in the 12 piece group, 7 7 so 10 is larger than 12 . Or TM2: You see the shaded area of 7 7 then 10 > 12 .
7 10

is larger than the shaded area of

7 , 12

7 7 However, no student in the Con-class could compare 10 and 12 in a meaningful way. Four (66%) students in the Con-class applied rule-based methods. For example, (Post-interview of CH1 and CH2):

CH1: Since the numerators are the same, I only need to compare denominators. The fraction that has the smaller denominator is the larger fraction.

130

D.-C. YANG

R: Why? Can you tell me your reasons? CH1: I don’t know. I memorized the rule I learned in mathematics class. Two (33%) students in the Exp-class and one (17%) student in the Conclass still could not compare fractions with the same numerators. For example, (Post-interview of CL2): CL2: I don’t know. Or (Post-interview of TL1) TL1:
7 12 7 , 10

>

because 12 is larger than 10.

These responses were coded as incorrect and could not explain. In examining the summary of students’ responses, we found that students in the Exp-class advanced in their use of number sense strategies on the post- and retention-interviews as compared to the pre-interview (66% vs. 0%). The data also indicated that students in the Exp-class were more successful in using number sense-based methods than students in the Con-class on the post- and retention-interviews (66% vs. 0%). In summary, the data indicates that there was an apparent change in the Exp-class after instruction and that students’ number sense ability improved greatly after they were provided with opportunities to explore numbers, operations, and their relationships, and to communicate their ideas in a conducive learning environment. The post-interviews revealed that students’ responses reflected number sense strategies. Furthermore, the results of retention-interviews indicated that students’ learning was meaningful and significant. Students’ responses in the Con-class in using number sense components reflected little change after instruction.

C ONCLUSION Although this teaching experiment was implemented over only one semester (about 4 and a half months) and limited to two classes from one school, the improvement made by the fifth graders in the experimental class is readily apparent. Though the generalizations of this study are limited due to the small sample size, the results do provide some important and interesting findings, as follows: 1. Students’ performance on number sense group tests in both the Expclass and the Con-class indicated that there were statistically significant differences for the scores on the post-tests and retention-tests as compared with the pretests at the α = 0.01 level. However, the scores for the Expclass increased 44% after instruction (the mean score went from 12.35 to

TEACHING AND LEARNING NUMBER SENSE

131

17.81), while the scores for the control class increased only 10% after instruction (the mean score went from 11.29 to 12.42). This indicated that the Exp-class made much more progress on number sense tests as compared with the Con-class. These data demonstrate that the teaching of number sense activities is effective and helpful in the development of students’ number sense. 2. In examining the individual interviews of students in the Exp-class and Con-class on pre-, post-, and retention-interviews, the changes and progress made by students in the Exp-class is apparent. Students in the Exp-class were not only able to effectively use number sense strategies after instruction, but they were also more successful than students in the Con-class in using number sense strategies to solve problems on the postand retention-interviews. The interview results showed that these number sense activities were effective in helping children develop their number sense abilities. Furthermore, the results of the retention-interviews further demonstrated that students’ learning in the Exp-class was meaningful and therefore long-lasting. The data from the interviews provide opportunities not only to better understand students’ improvement in using number sense strategies, but also to explore students’ thinking and comprehension of numbers and operations. 3. There is a major difference between number sense activities designed for this study and the mathematics textbooks used in Taiwanese classes. The mathematics textbooks used for the fifth grade level do not teach students to find a least common denominator or the cross-product procedure for comparing fractions. The textbooks teach students to use a unit quantity to order fractions. For example, when comparing 1 and 3 , the teacher gave 3 8 the students 24 pieces of chips as a unit and taught them to find how many pieces are equivalent to 1 and 3 . The textbooks show that 24 × 1 = 8; 3 8 3 24 × 3 = 9, therefore, 3 is larger due to having 9 pieces of chips. In the 8 8 context of this situation, students knew how to order fractions when the unit was given. However, they were unable to solve the problem if the unit quantity was not given. On the contrary, the instruction for the Exp-class focused on the development of conceptual understanding through pictorial representations. The teacher in the Exp-class encouraged children to draw diagrams to help them understand fractions and also encourage students to represent fractions by verbal language. This result is consistent with the study of Cramer et al. (2002). 4. The use of benchmarks for students in the Exp-class was much more prevalent after instruction as compared with students in the control class. The use of benchmarks is not introduced in the Taiwanese textbooks. The-

132

D.-C. YANG

refore, the progress made by students in the experimental class is apparent. For example, when students were asked to estimate: 4 + 6 , students in 5 7 the Exp-class knew to select the 1 as a benchmark to decide the answer, but students in the Con-class were unable to use benchmarks to find the answer. In summary, this study demonstrates that students’ number sense can be effectively developed through establishing a classroom environment that encourages communication, exploration, discussion, thinking, and reasoning. The results of this research study confirm the earlier studies (Markovits & Sowder, 1994; Yang, 2002a, 2002b; Yang & Reys, 2001a, 2001b) that children’s number sense can be fostered through appropriate teaching. Furthermore, the process-oriented teaching model practically applied in this study not only supports the NCTM process standards and the guidelines for a Nine-Year Joint Mathematics Curricula Plan, but also can be used as a guide to help teachers implement the mathematics curriculum. In fact, this teaching model provides an artifact that could go a long way toward helping teachers actually embed these processes in their practice. The teaching of number sense focuses on conceptual understanding. It directs children to pursue meaningful learning. This further confirms the statement in the Guidelines that students’ learning of mathematics should be meaningful. From this research study, one can gather that well-designed instructional lessons not only help children develop number sense, but actually promote the development of critical thinking and reasoning about numbers and operations.

ACKNOWLEDGEMENTS This paper is a part of a research project supported by the National Science Council in Taiwan with grant no. NSC 90-2521-S-415-001. Any opinions expressed in here are those of the author and do not necessarily reflect the views of the National Science Council in Taiwan. The author gratefully acknowledges the help of teacher Miss Lin, a Taiwanese 5th grade teacher, without whose cooperation and teaching skills this paper would not have resulted.

R EFERENCES
Anghileri, J. (2000). Teaching number sense. Trowbridge, Wiltshire: Cromwell Press. Australian Education Council. (1991). A national statement on mathematics for Australian school. Melbourne: Curriculum Corporation.

TEACHING AND LEARNING NUMBER SENSE

133

Cobb, P., Wood, T., Yackel, E., Nicholls, J., Wheatly, G., Trigatti, B. & Perlwitz, M. (1991). Assessment of a problem-centered second-grade mathematics project. Journal for Research in Mathematics Education, 22, 3–29. Cockcroft, W.H. (1982). Mathematics Counts. London: HMSO. Cramer, K.A., Post, T.R. & delMas, R.C. (2002). Initial fraction learning by fourth- and fifth-grade students: A comparison of the effects of using commercial curricula with the effects of using the rational number project curriculum. Journal for Research in Mathematics Education, 33, 111–144. Fraivillig, J. (2001). Strategies for advancing children’s mathematical thinking. Teaching Children Mathematics, 454–459. Greeno, J.G. (1991). Number sense as situated knowing in a conceptual domain. Journal for Research in Mathematics Education, 22, 170–218. Howden, H. (1989). Teaching number sense. Arithmetic Teacher, 36, 6–11. Hsu, C.Y., Yang, D.C. & Li, F.M. (2001). The design of “The fifth and sixth grade number sense rating scale”. Chinese Journal of Science Education, 9, 351–374. Japanese Ministry of Education. (1989). Curriculum of Mathematics for the Elementary School. Tokyo: Printing Bureau. Lin, F.L. (1989). Strategies used by adders in proportional reasoning. In Proceedings of the 13th International Conference for the Psychology of Mathematics Education, Vol. 2, Paris, France, pp. 234–241. Markovits, Z. & Sowder, J.T. (1994). Developing number sense: An intervention study in grade 7. Journal for Research in Mathematics Education, 25, 4–29. McIntosh, A., Reys, B.J. & Reys, R.E. (1992). A proposed framework for examining basic number sense. For the Learning of Mathematics, 12, 2–8. McIntosh, A., Reys, B.J., Reys, R.E., Bana, J. & Farrel, B. (1997). Number sense in school mathematics: Student performance in four countries. Mathematics, Science, and Technology Education Centre, Edith Cowan University. Ministry of Education in Taiwan. (2000). Nine-year-integrated mathematics curriculum standards for national schools from grade 1 to 9 in Taiwan (pp. 19–86). Taiwan: Author (in Chinese). National Council of Teachers of Mathematics (NCTM). (1989). Curriculum and evaluation standards for school mathematics. Reston, VA: NCTM. National Council of Teachers of Mathematics (NCTM). (2000). The principles and standards for school mathematics. Reston, VA: NCTM. National Research Council. (1989). Everybody counts. A report to the nation on the future of mathematics education. Washington, DC: National Academy Press. Ning, T.C. (1992). Child’s meanings of fractional number words, Unpublished Doctoral Dissertation, The University of Georgia, Athens, GA. Reys, B.J. (1994). Promoting number sense in middle grades. Teaching Mathematics in the Middle School, 1(2), 114–120. Reys, R.E. & Yang, D.C. (1998). Relationship between computational performance and number sense among sixth- and eighth-grade students in Taiwan. Journal for Research in Mathematics Education, 29, 225–237. Sowder, J. (1992a). Estimation and number sense. In D.A. Grouws (Ed.), Handbook of research on mathematics teaching and learning (pp. 371–389). New York: Macmillan. Sowder, J. (1992b). Making sense of numbers in school mathematics. In G. Leinhardt & R. Hattrup (Eds.), Analysis of arithmetic for mathematics teaching (pp. 1–51). Hillsdale, NJ: Erlbaum.

134

D.-C. YANG

Treffers, A. (1991). Meeting innumeracy at primary school. Educational Studies in Mathematics, 26, 333–352. Warrington, M.A. & Kamii, C. (1998). Multiplication with fractions: A Piagetian constructivist approach. Mathematics Teaching in the Middle School, 3, 339–343. Yang, D.C. (2002a). Teaching and learning number sense: One successful process-oriented activity with six grade students in Taiwan. School Science and Mathematics, 102, 152– 157. Yang, D.C. (2002b). The study of sixth grade students’ development of number sense through well-designed number sense activities. Chinese Journal of Science Education, 10(3), 233–259. Yang, D.C. & Reys, R.E. (2001a). Developing number sense. Mathematics Teaching, 176, 39–41. Yang, D.C. & Reys, R.E. (2001b). One fraction problem: Many solution paths. Mathematics Teaching in the Middle School, 7, 164–166.

Graduate Institute of Mathematics Education, National Chiayi University, 85, Wen Lung, Ming-Hsiung, 621, Chiayi, Taiwan E-mail: dcyang@ms21.hinet.net