Professional Documents
Culture Documents
net/publication/271836030
CITATIONS READS
12 1,491
2 authors:
Some of the authors of this publication are also working on these related projects:
MATHEMATICAL THINKING AMONG PRE-SERVICE TEACHERS: USING DYNAMIC MATHEMATICS SOFTWARE VERSUS PAPER AND PENCIL View project
All content following this page was uploaded by Shiau Wei Chan on 05 February 2015.
Student Misconceptions
Am. J. Phys. 11, 227 (1943); 10.1119/1.1990482
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
103.1.71.118 On: Thu, 05 Feb 2015 11:37:23
Malaysian Students’ Misconceptions about Measures of
Central Tendency: An Error Analysis
Zaleha Ismaila and Shiau Wei Chanb
a
Associate Professor Dr, Faculty of Education, Universiti Teknologi Malaysia, 81310 Skudai, Johor, Malaysia.
b
Dr, Faculty of Education, Universiti Teknologi Malaysia, 81310 Skudai, Johor, Malaysia.
Abstract. Earlier studies revealed that students encountered difficulties in understanding the concept of measures of central
tendency. However, there is negligible empirical support concerning this problem in Malaysian context. Thus, the purpose of
this study is to identify misconceptions held by tenth grade students from Malaysian secondary schools regarding measures
of central tendency. The instrument used was a statistical reasoning test. It was administered to 412 tenth grade students from
nine different schools. Overall, the results demonstrated that tenth grade students held considerable misconceptions about
measures of central tendency. Interestingly, some misconceptions were new which had not been identified in previous
research. This study provides implications for instructors and researchers planning learning goals as well as designing
instructional activities and assessments for future studies.
Keywords: Misconceptions, Measures of central Tendency, Secondary education
PACS: 01
INTRODUCTION
Statistics was considered as difficult domain to study because the statistical concepts are complex (Garfield and
Ben-Zvi, 2008). A number of students would feel nervous when learning statistics. Poor grasp of mathematical ideas
may cause difficulties troubles in statistics. Some of students perceived that statistics is similar to mathematics; that
is it has only one accurate answer (Ben-Zvi and Garfield, 2004). As a result, the difficulties that faced by the
students in comprehending statistical ideas may lead to misconceptions in statistical reasoning (Templaar, Gijselaers
and van der Loeff, 2006). Misconceptions, is described as a prototype for mistakes that reveals the misinterpretation
of statistical ideas (Cohen et. Al., 1996).
Groth and Bergner (2005) asserted that the measures of central tendency among students were incompatible.
From earlier studies, it was found that students had a lot of misconceptions in measures of central tendency, for
instance mode (Huck, 2009), median (Cooper and Shore, 2008), and mean (Olani et. al., 2011). When the data are
presented in the form of graphs, students confronted the problems as well (Lee and Meletiou-Mavrothesis, 2003).
According to Burrill & Camden (2005), most of the students cannot master the statistical concepts due to the
instruction that emphasize on the examination rather than conceptual knowledge. In addition, the conceptual
understanding of students is unable to be well measured owing to the traditional questions in the statistics (Garfield
and Chance, 2000). Students tended to learn by remembering the formulas with no comprehension of statistical
concepts which incapable to foster their conceptual understanding (Broers and Imbos, 2005).
In Malaysia topic of statistics are incorporated into mathematics and additional mathematics syllabus. Malaysian
students learn statistics from primary three to secondary five. However, this does not stop them from misconceptions
and making mistakes in solving basic statistics as indicated in the TIMSS data (Foo and Idris, 2010). Despite a
multitude of studies regarding misconceptions in statistics, little is known about the misconceptions regarding
measures of central tendency held by Malaysian students. This study was conducted to discover more about
Malaysian secondary students’ misconceptions in measures of central tendency. The findings of this study will
provide guidelines for preparing learning objectives, assessments and instructional activities to reduce
misconceptions.
LITERATURE REVIEW
A measure of central tendency is regarded as main aspects in making conjecture on data analysis and in
comprehending the ideas of distribution. It is a crucial component in inferential and descriptive statistics (Garfield
and Ben-Zvi, 2008). Average is judged as measures of central tendency which comprised of mode, median, and
mean by several statisticians. Nonetheless, average was interpreted in a different way based on problem context as
argued by Konold & Pollastek (2002, 2004), for instance fair share, data reduction, signal in noise, and typical
METHODOLOGY
Participants
The participants in this study were 412 tenth grade students, randomly selected from 14 classes of nine
secondary schools in Malaysia. Among them, 172 students (41.74%) were male and 240 students (58.25%) were
female. There were 229 Malay (55.58%), 159 Chinese (38.59%), 22 Indian (5.34%), 1 Kadazan Dusun (0.24%) and
Instrumentation
The instrument used in this study was a statistical reasoning test with five main questions as well as sub-
questions. This test was designed to assess the misconceptions held by students regarding measures of central
tendency (Question 1 to 4) and measures of variability (Question 5). Nevertheless, this paper will only discuss
misconceptions about measures of central tendency from Question 3. All the questions used were categorized
according to the table of specification according to the cognitive domain of Bloom’s Taxonomy (1956). Garfield and
Ben-Zvi (2008) contended that statistical reasoning corresponds to the Comprehension Level in Bloom’s Taxonomy
and may include some features from the Application and Analysis levels. Thus, in the statistical reasoning test used,
there are 7 items that test Comprehension, 7 items that assess Application abilities, and 2 items that examine the
student’s ability to Analysis. These items were developed based on the features of statistical reasoning where
students interpreted the data and connected one concept to another in first part of the test. In the second part of the
test, the students were asked to explain responses. The Question 3 is shown as below:
The following histogram in Figure 1 shows the scores ˄1-10˅obtained by 36 students in a mathematics test.
TABLE 1. Question 3a
% Correct: 37.62%
No Misconceptions Number of Percentage
students (%)
M14 Finding the mean of height of bars and dividing by the sum of values on the horizontal 2 0.49
axis
111 2 0 3 3 4 6 8 7 36
0.65
0 1 2 3 4 5 6 7 8 9 10 55
M15 Finding the mean of values on horizontal axis 2 0.49
0 1 2 3 4 5 6 7 8 9 10 55
1.53
111 2 0 3 3 4 6 8 7 36
M16 Finding the mean of height of bars and dividing by number of data on the horizontal axis 44 10.68
(omitting 4)
111 2 0 3 3 4 6 8 7 36
3.6
10 10
M17 Finding the mean of height of bars and dividing by number of data on the horizontal axis 21 5.10
(including 4)
111 2 0 3 3 4 6 8 7 36
3.27
11 11
M18 Finding the mean of values on horizontal axis (including the value of 4) and dividing by 1 0.24
number of data on the horizontal axis (omitting 4)
0 1 2 3 4 5 6 7 8 9 10 55
5.5
10 10
M19 Finding the mean of values on horizontal axis (omitting the value of 4) and dividing by 1 0.24
frequency of score
0 1 2 3 5 6 7 8 9 10 51
1.42
1 1 1 2 0 3 3 4 6 8 7 36
M20 Finding the mean of midpoint for ungrouped data in histogram 2 0.49
0.5(1) 1.5(1) 2.5(2) 5.5(3) 6.5(3) 7.5(4) 8.5(6) 9.5(8) 10.5(7) 273.5
7.597 7.6
111 2 0 3 3 4 6 8 7 36
M21 Finding the mean of midpoint for ungrouped data in histogram and dividing by midpoint 1 0.24
0.5(1) 1.5(1) 2.5(2) 5.5(3) 6.5(3) 7.5(4) 8.5(6) 9.5(8) 10.5(7) 273.5
4.52
0.5 1.5 2.5 3.5 4.5 5.5 6.5 7.5 8.5 9.5 10.5 60.5
M22 Dividing the product of frequency and value of data by number of data on the horizontal 11 2.67
axis (Omitting 4)
0(1) 1(1) 2(1) 3(2) 4(0) 5(3) 6(3) 7(4) 8(6) 9(8) 10(7) 260
26
10 10
M23 Dividing the product of frequency and value of data by number of data by number of 12 2.91
data on the horizontal axis (including 4)
Table 2 displays the two misconceptions M25 and M26 held by students and revealed by their answers to
Question 3b. There were 6.31% of students who believed that “mode is referred to as a peak, as the highest point on
the curve and as the top of the curve”(M25) as claimed by Huck (2009). In this question, the highest frequency in
histogram was 8, so students B035, F187, and J248 use this number as their answer. On the other hand, a new
misconception (M26) was observed in this question where 3.88% of students recognized “mode is the number
appears the most in the score,” as was the case for students A008, F160, and J278. The Number 1 appeared three
times in the histogram, so they selected 1 as the answer. 4.85% of students’ solutions were not appropriate and
11.17% of students did not put in the effort required to answer the problem.
TABLE 2. Question 3b
% Correct: 73.79%
No Misconceptions Number of students Percentage (%)
M25 Mode is referred to as a peak, as the highest point on the curve and as 26 6.31
the top of the curve (8)
M26 Mode is the number appears the most in the score (1) 16 3.88
Irrelevant 20 4.85
No attempt 46 11.17
From the Table 6, five misconceptions held by the students in Question 3c1 can be observed. There were 5.58%
of students having a tendency towards M27 and 2.18% of students held misconception M28. These two
misconceptions were consistent with the results of the study conducted by Cooper and Shore (2008). In this
question, three new misconceptions were discovered. The percentage of students who held misconceptions M29,
M30, and M31 were 3.64%, 16.26%, and 0.73% respectively. A high percentage of students, 39.56%, solved this
problem improperly and some of them did not show their work. A further 22.82% of students did not attempt this
question.
TABLE 3. Question 3c
% Correct: 9.22%
No Misconceptions Number of students Percentage
(%)
M27 Finding the median of frequencies of the data values 23 5.58
0, 1, 1, 1, 2, 3, 3, 4, 6, 8, 7
M28 Finding the median of the values on the horizontal axis 9 2.18
(1) 0 , 1 , 2 , 3 , 4 , 5 , 6 , 7 , 8 , 9 , 10 or
(2) 1, 2, 3, 4, 5, 6, 7, 8, 9, 10
M29 Dividing the number of data on the horizontal axis by two (1) or (2) = 15 3.64
5.5
M30 Adding the frequency of score and dividing it by two 67 16.26
M31 Finding median using formula = = 18.5 3 0.73
Irrelevant 163 39.56
No attempt 94 22.82
For Question 3d, 28 students (6.80%) knew that median was the most suitable measure of central tendency to
represent the data, but they did not know how to explain their answer as was the case for students F143, I232, J263,
K288, L309, L323, and M357. 152 students, including students A007, B033, D085, F128, F133, H212, I224, K28,
M334, and N411, believed that mean was the most appropriate measure of central tendency to characterize the data.
There were 73 students who answered “mode” including students C072, F130, F156, K281, and N403. 8 students
gave answers other than mean, mode, and median including G180, K285, and L330. The findings in this question
DISCUSSIONS
In this paper, the findings from Question 3 were used to answer the research question – [What are the
misconceptions held by tenth grade students in regards to measures of central tendency?] Misconceptions regarding
the mean of histograms were consistent with the result of the study conducted by Cooper and Shore (2008) including
M16, M17, and M18. The misconception concerning mode in histograms from this study verified the assertion of
Huck (2009), i.e. M25. Likewise, misconceptions regarding the median of histograms corroborated the findings of
Cooper and Shore’s (2008) study, which were M27 and M28. In addition to verifying the findings of other
researchers, this study also uncovered several new misconceptions. Eight new misconceptions regarding the mean
of histograms were discovered from the results of Question 3a, namely, M14, M15, M19, M20, M21, M22, M23 and
M24 as further defined in Table 1.
Also, a new misconception regarding the mode of histograms was found as demonstrated in Table 2, i.e. M26. In
addition, three new misconceptions concerning the median of histograms were discovered too, namely M29, M30
and M31 as discussed in Table 3. According to delMas, Garfield, and Ooms (2005) and Cooper and Shore (2008),
data represented graphically will cause students to have misconceptions. Such situations occur because of the
students’ limited understanding of graphs [30]. In short, the tenth grade students in this study held several faulty
conceptions in descriptive statistics, especially regarding measures of central tendency. Educators and teachers
should work towards remedying this situation. These misconceptions were happened because the students were not
taught on how to give explanation when solving the questions of measures of central tendency in schools. In
addition, they solved problems procedurally and memorize the formula without understanding the concepts.
From the findings, we noticed that a few students were able to solve the first part of the question, but could not
solve the second part of the question. Such situations occurred because the students seldom dealt with tasks that
required them to give explanations, thus they were not familiar with those tasks. Some students had a poor
understanding of statistical vocabulary needed to explain their thinking. On the other hand, some students were
unable to answer the first part of the question but managed to answer the second part of the question. This may have
stemmed from mathematical errors, failure to show any steps used to obtain their solution, or using inappropriate
methods to solve the problem.
In this study, many students did not attempt to answer some of the items and left them blank. This may have
occurred because they did not know how to solve the problem, did not want to answer the question or they did not
have enough time There were also some students who executed other methods to solve the problems incorrectly.
Their thinking and reasoning should be investigated. In future studies, these students should be interviewed in order
to understand their behavior.
CONCLUSION
By and large, Malaysian ten graders harbor numerous misconceptions regarding measures of central tendency.
Some misconceptions were similar to misconceptions discovered in earlier studies and a number of new
misconceptions were diagnosed in this study. The new findings of this study contribute to the statistics field. In
future studies, more structured empirical studies involving intervention are needed to overcome the misconceptions.
This study provides a direction for instructors and researchers to think about learning goals, use of appropriate
instruction and assessments tools in the statistics classroom to promote statistical reasoning.
ACKNOWLEDGMENTS
This study is funded by the Ministry of Higher Education, Malaysia and Universiti Teknologi Malaysia through
FRGS grant (Grant Vote No. R.J130000.7831.4F187).