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 statisti-
cally 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 Educa-
tion, 1989; Markovits & Sowder, 1994; McIntosh, Reys, Reys, Bana & Far-
rel, 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. Further-
more, 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 at-
tention. 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 edu-
cational revolution in Taiwan. The guidelines for a Nine-Year Joint Math-
ematics Curricula Plan (Ministry of Education in Taiwan, 2000) stresses
that the learning and teaching of mathematics should highlight meaning-
ful 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 num-
bers. This includes the ability to develop useful, flexible, and efficient
strategies (i.e., mental computation or estimation) for handling numeri-
cal 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 educa-
tion, 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 com-
ponents 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 com-

pare 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 rea-
sonableness 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 Tai-
wan, 2000)) was released and covered all subjects, including mathematics.
Number sense has been considered an important topic in school mathemat-
ics 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 cur-
ricula. If significant improvements in mathematics are to be made in Tai-
wan, the development of number sense should be considered as an impor-
tant 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 situa-
tions. 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, Tri-
gatti & 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 in-
struction 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 demon-
strated 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 ex-
tended 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 partic-
ipate 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 edu-
cational 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, in-
cluding 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 play-
ground. 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 consis-
tency 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-in-

terview 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 Review the related papers and design the number sense project
08/30/01–09/08/01 Pretest and pre-interview for both Exp- and Con-class
09/10/01–12/31/01 Number sense activities integrated into Exp-class
01/01/02–01/18/02 Posttest and post-interview for both Exp- and Con-class
01/19/02–02/17/02 Winter break
02/18/02–06/01/02 No number sense activities were integrated into classes
06/01/02–06/12/02 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 re-
sponses 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 dis-
cussion, 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 de-
velop 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 teach-

ing 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 devi-
ations, and t-test results for the experimental class and the control class.
The data indicate that there are statistically significant differences be-
tween 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 experi-
mental 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 in-
struction. 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 Mean score (Correct%) SD T-value

Pretest 12.35 (33.4%) 4.83 −10.949∗

Exp-classes Posttest 17.81 (48.1%) 5.29
(N = 37) Pretest 12.35 (33.4%) 4.83 −13.218∗
Retention 19.08 (51.2%) 5.46
Pretest 11.29 (30.5%) 4.09 −3.775∗
Con-classes Posttest 12.42 (33.6%) 3.94
(N = 38) Pretest 11.29 (30.5%) 4.09 −4.734∗
Retention 13.23 (35.9%) 4.35

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 TH2 TM1 TM2 TL1 TL2

(CH1) (CH2) (CM1) (CM2) (CL1) (CL2)

Pre-interview
Correct
NS-based 6 (9) 4 (5) 4 (4) 2 (3) 2 (3) 2 (3)
Rule-based 2 (4) 5 (3) 2 (0) 3 (1) 0 (0) 0 (1)
Couldn’t Ex 1 (0) 2 (1) 2 (2) 2 (2) 4 (2) 3 (2)
Incorrect
NS-based 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0)
Rule-based 0 (2) 0 (3) 0 (4) 2 (4) 0 (3) 0 (3)
Couldn’t Ex 6 (0) 4 (3) 7 (5) 6 (5) 9 (7) 10 (6)
Post-interview
Correct
NS-based 15 (10) 11 (6) 10 (5) 12 (5) 5 (3) 5 (3)
Rule-based 0 (4) 3 (3) 2 (3) 1 (2) 0 (3) 0 (4)
Couldn’t Ex 0 (0) 0 (4) 0 (2) 0 (2) 1 (1) 3 (1)
Incorrect
NS-based 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0)
Rule-based 0 (1) 0 (0) 1 (2) 1 (2) 1 (3) 0 (1)
Couldn’t Ex 0 (0) 1 (2) 2 (3) 1 (4) 8 (5) 7 (6)
Retention-interview
Correct
NS-based 15 (11) 12 (6) 11 (5) 12 (4) 5 (3) 5 (2)
Rule-based 0 (3) 1 (3) 0 (0) 2 (1) 0 (4) 1 (5)
Couldn’t Ex 0 (1) 1 (4) 1 (5) 0 (3) 4 (3) 2 (2)
Incorrect
NS-based 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0)
Rule-based 0 (0) 0 (0) 0 (1) 0 (2) 0 (0) 1 (0)
Couldn’t Ex 0 (0) 1 (2) 3 (4) 1 (5) 6 (5) 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. Two same sized watermelons were shared by five students equally. How many
watermelons can each student have? Why?
(1) Greater than a half of a watermelon
(2) Equal to 15 of a watermelon
(3) Equal to 25 of a watermelon
(4) Greater than 2.5 watermelons

Exp-A Con-B

Pre- Post- Retention Pre- Post- Retention

Correct
NS-based 1 (17%) 6 (100%) 6 (100%) 2 (33%) 3 (50%) 3 (50%)
Rule-based 2 (33%) 0 (0%) 0 (0%) 0 (0%) 1 (17%) 1 (17%)
Couldn’t explain 1 (17%) 0 (0%) 0 (0%) 1 (17%) 1 (17%) 1 (17%)
Incorrect
NS-based 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%)
Rule-based 0 (0%) 0 (0%) 0 (0%) 1 (17%) 0 (0%) 0 (0%)
Couldn’t explain 2 (33%) 0 (0%) 0 (0%) 2 (33%) 1 (17%) 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, (Pre-

interview 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 15 wa-
termelon. Since there are two watermelons, each person can get 25
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-inter-
view of TM2):

TM2: 2 ÷ 5 = 25 , so the answer is 2 ÷ 5 = 25 .

R: Can you tell me why the answer is 25 ?
TM2: Because 2 ÷ 5 = 25 , so the answer is 2 ÷ 5 = 25 .
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 = 25 ”, but could not give an explanation,
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 25 .
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 15 .
R: Can your tell me your reasons?
TL1: I don’t know.
During the post- and retention-interview, all six students in the exper-
imental 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 15 of one watermelon. Since there are
two watermelons, each one can get 25 watermelons.
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: 25 . I don’t know how to explain it.
One student gave an incorrect answer on pre-, post-, and retention-inter-
views. CL2 answered: “ 15 ” and “I don’t know how to explain it.”
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 stu-
dents 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 45 + 67 ? Why?
(1) 12 (2) 10 (3) 2 (4) 1 (5) Without calculating can’t find the answer

Exp-A and B Con-A and B

Pre- Post- Retention Pre- Post- Retention

Correct
NS-based 0 (0%) 3 (50%) 4 (66%) 0 (0%) 0 (0%) 0 (0%)
Rule-based 1 (17%) 0 (0%) 0 (0%) 1 (17%) 2 (33%) 2 (33%)
Couldn’t explain 0 (0%) 1 (17%) 0 (0%) 1 (17%) 1 (17%) 1 (17%)
Incorrect
NS-based 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%)
Rule-based 3 (50%) 1 (17%) 1 (17%) 2 (33%) 1 (17%) 1 (17%)
Couldn’t explain 2 (33%) 1 (17%) 1 (17%) 2 (33%) 2 (33%) 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: Can you tell me the answer for question 6 and your reasoning?
TH2: I think the answer is 2.
R: Can you tell me your reasons?
TH2: Because
4 6 28 30 58 23
+ = + = =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 de-
nominators) 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: The answer is 10.

R: Why? Please tell me your method.
TH1: Because 6 + 4 is equal to 10.
R: Can you do it another way?
TH1: (Hesitating for a while!) I don’t know.
Or (Pre-interview of CM1)
CM1: 1.
R: Why? Can you tell me your reasons?
CM1: 4
5
+ 67 = 10
12
, hence the answer is close 1.
R: Can you do it another way?
CM1: (Shaking her head!) I don’t know.
These methods were coded as incorrect and rule-based. Two students in
each class gave incorrect answers and no explanations. For example, (Pre-
interview 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 45 is less than 1 and near 1, 6
7
is also less than 1 and near 1.
Therefore, the sum is about 2.
Or (Post-interview of TH2)
TH2: The answer is 2.
R: Why? Can you tell me your reasons?
TH2: Since 45 is near 1 and 67 is also near 1, the sum is near 2.
R: Can you do it by different way?
TH2: I can find the common denominator and compute it.
R: Can you tell me what’s the difference between these methods.
TH2: 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.
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 or 7 is larger? Why?
8. Without calculating, which one 10 12
7 7
(1) 10 (2) 12 (3) same (4) Without calculating can’t find the answer

Exp-A Con-B

Pre- Post- Retention Pre- Post- Retention

Correct
NS-based 1 (17%) 4 (66%) 4 (66%) 0 (0%) 0 (0%) 0 (0%)
Rule-based 1 (17%) 0 (0%) 0 (0%) 2 (33%) 4 (66%) 4 (66%)
Couldn’t explain 1 (17%) 0 (0%) 0 (0%) 2 (33%) 1 (17%) 1 (17%)
Incorrect
NS-based 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%)
Rule-based 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%)
Couldn’t explain 3 (50%) 2 (33%) 2 (33%) 2 (33%) 1 (17%) 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, (Post-
interview 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 post-
and retention-interviews (50% vs. 0%) and (66% vs. 0%).
Table VI presents data from the fifth grader’ responses for the Exp-
7
and Con-class on comparing fractional size: without calculating, is 10 or
7
12
is larger? The fifth graders’ mathematics textbooks in Taiwan do not
teach children to find a least common denominator or cross-product proce-
dure for comparing fractions. The instruction for the Exp-class focused on
developing conceptual understanding through pictorial and verbal repre-
sentations. The textbooks used in the Con-class taught children to compare
fractions by using concrete materials. For example, when comparing 38 and
5
12
, the teacher gave students 24 pieces of chips as a unit and taught them
to find the pieces of 38 and 12
5
. 24 × 38 = 9; 24 × 12
5
= 10, therefore, 12
5
is
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
10
is larger than 12 .” This was coded as number sense-based. No student
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 pro-
vide an explanation. This was coded as correct and could not explain.
Three (50%) students in the Exp- and two (33%) students in the Con-
class gave incorrect answers and incorrect explanations. For example, (Pre-
interview 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 Exp-
class 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
7 7
TM2: You see the shaded area of 10
is larger than the shaded area of 12
,
7 7
then 10 > 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 denom-
inators. 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 Con-
class still could not compare fractions with the same numerators. For ex-
ample, (Post-interview of CL2):
CL2: I don’t know.
Or (Post-interview of TL1)
7 7
TL1: 12
> 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 stu-
dents 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 im-
proved greatly after they were provided with opportunities to explore num-
bers, 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 mean-
ingful and significant. Students’ responses in the Con-class in using num-
ber sense components reflected little change after instruction.

C ONCLUSION

Although this teaching experiment was implemented over only one se-
mester (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 lim-
ited 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 Exp-
class 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 Exp-
class 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 in-
struction (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 post-
and 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 bet-
ter 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 stu-
dents 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 13 and 38 , the teacher gave
the students 24 pieces of chips as a unit and taught them to find how many
pieces are equivalent to 13 and 38 . The textbooks show that 24 × 13 = 8;
24 × 38 = 9, therefore, 38 is larger due to having 9 pieces of chips. In the
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: 45 + 67 , students in
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 reason-
ing. 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. Fur-
thermore, 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 help-
ing 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 Tai-
wanese 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 construc-
tivist 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. Mathemat-
ics 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