The Effect of Manipulatives on Mathematics Achievement and Attitudes of Secondary School Students

The purpose of this study is to investigate the effect of manipulatives (concrete learning materials) both on the academic achievement of secondary school students in mathematics and on their attitudes towards mathematics. Pretest-posttest control group experimental model, which is one of the quasi-experimental research designs, was used in the study. The study group consisted of 48 seventh grade students (24 in experiment group and 24 in control group) studying in a state school in the Southeastern Region of Turkey in 2014-2015 school year. The ages of students range between 13 and 14. Mathematics achievement test and mathematics attitude scale were applied in order to collect the research data. As a result of the research, posttest mathematics academic achievement scores of experiment and control groups were found to differ significantly in favor of posttests in both groups. The scores of attitude towards mathematics for experiment and control groups were significantly different in posttests in favor of the experiment group.


Introduction
"Teacher!What is this of use for us in daily life?" is one of the most frequent questions that secondary school mathematics teachers are confronted with.One reason why students are right to ask this question is the fact that mathematics is a discipline mainly related to abstract concepts (Hartshorn & Boren, 1990;Laski, Jor'dan, Daoust, & Murray, 2015;Tuncer, 2008).Piaget, one of the founders of constructivism, emphasizes that abstract aspect of mathematics as well.Constructivism, which underlies the mathematics curriculum for elementary education revised in 2005-2006 school year, is a student-focused approach based on construction of complex structures of knowledge by students through active interrelation with concrete materials and on interpreting abstract mathematical concepts by means of concrete experiences (Bruner, 1977;Dienes, 1971;Piaget, 1965;White, 2012), as opposed to traditional mathematics instruction.According to Piaget (1965), students cannot understand and learn the abstract mathematical concepts presented in form of words and symbols solely by way of direct explanation; they do not have the necessary cognitive maturity for it.That is why, he argues that it is possible for students to learn abstract mathematical concepts as a result of their experiences with concrete objects and concrete materials.
of "manipulative" as a mathematical concept, several definitions which present the features of the manipulatives are found.
Manipulatives enable students to integrate their knowledge and associate them with their thoughts in order to understand mathematical concepts thoroughly (Boggan, Harper, & Whitmire, 2010;Kelly, 2006); they contribute to students' communication with their own mathematical thinking and to bringing their mathematical ideas to a higher cognitive level (Balka, 1993;Ojose & Sexton, 2009;O'Shea, 1993).They also evoke amusement in the teaching process by providing active participation of both students and teachers (Boggan, Harper, & Whitmire, 2010;Ojose & Sexton, 2009) and in this way, lead to permanent learning through creating equality of opportunity among students (Şahin, 2013).As stated by Tunç, Durmuş and Akkaya in their research in 2011, manipulatives are instruction materials which facilitate teaching and learning (Tunç, Durmuş, & Akkaya, 2011), and make positive contributions to conceptualization and interpretation processes of students (Ross & Kurtz, 1993).However, manipulatives do not only contribute to the cognitive aspect of the learner, they also enhance the development of psychomotor skills (Cope, 2015;Kocaman, 2015) by addressing several senses of the learner such as sight, touch or hearing (Hynes, 1986;Kennedy, 1986;McClung, 1998;Perry & Howard, 1997), inside and outside the class.As stated by Kocaman (2015), manipulatives should not be thought of as a remedy providing an advantage against every difficulty that students face in mathematics.Instead, as underlined by Gökmen (2012), teachers and students need to make a common sense out of the manipulatives they will use.Otherwise, manipulatives would not go beyond being a means of entertainment alone and cannot help the learner to comprehend any concepts.That is why, it was emphasized in many studies that manipulatives should be understood and applied properly especially during the teaching-learning process (Ball, 1992;Clements & McMillen, 1996;Moyer, 2001;Thompson, 1992).At this point, Kelly (2006) underlines the need to provide the teachers with a sound basis of information about when, why and how to use manipulatives.It was also stated that teachers need to explain the objectives to the students explicitly and introduce the manipulatives, giving the necessary information, during the process of teaching.Thus, the manipulatives will go beyond being instruments of play and entertainment from the perspective of the students and will be transformed into materials contributing to their learning.Nevertheless, Kelly (2006) emphasizes that manipulatives should be used not occasionally but all along, as they enhance the conceptualization process of the students positively.As Ojose and Sexton (2009) put it, long-term application of manipulatives leads to assimilation of abstract mathematical concepts through observation of models and thus, to an increase in the achievement of students.Therefore, in order for teachers to develop and use relevant, effective and productive manipulatives, they firstly need to be well informed about the features of manipulatives in educational environments, principles of preparing manipulatives, their advantages and limitations and to have the capability to explain those materials clearly to the students (Akkan & Çakıroğlu, 2011;Tutak, Kılıçarslan, Akgül, Güder, & İç, 2012).It was revealed that the application of manipulatives effectively in the course of teaching process, as stated above, contributes to increasing the mathematics exam grades of students, creating an outstanding performance of students in classes, decreasing anxiety against mathematics, rising cognitive flexibility and positively affecting their attitudes towards mathematics (Bullmaster, 2013;Clements, 1999;Driscoll, 1983;Ojose & Sexton, 2009;Raphael & Wahlstrom, 1989;Sowell, 1989;Suydam, 1986).
On the other hand, there are also some other researches putting forth that there is not a positive correlation between the manipulatives and mathematics achievement due to the prejudices of teachers against manipulatives (Drickey, 2006;McClung, 1998), and that manipulatives do not facilitate learning in mathematics instruction (Fuson & Briars, 1990;Sowel, 1989;Wearne & Hibert, 1988), which have given way to studies that do not support the use of manipulatives in mathematics instruction (Ball, 1992;Thompson P. & Thompson A., 1990;Winograd & Flores, 1986).Thus, studies have emerged in literature, which drove inconsistent conclusions as to the effects of using manipulatives upon the achievement of students, especially with the students at higher levels of education, who have gained abstract thinking skills.This situation indicates to us that the effects of manipulatives on the achievement and attitude of students have yet not been explicitly clarified.

Purpose of the Study
The purpose of this study is to investigate the effect of manipulatives used for the subject of circles and spheres in the seventh grade mathematics course upon the achievement and attitudes of students.To this end, answers to following questions were sought.

Research Design
In this study investigating the effect of manipulatives on academic achievement and attitude towards mathematics course, pre-test/post-test control group model which is one of the quasi-experimental designs, was used.In pre-test/post-test control group model, one of the groups, formed previously for some other purpose, is randomly selected as the experiment group, whereas the other group is taken for the control group.The groups are evaluated once before the experiment and once after it.The model allows controlling errors originating from sources such as date, maturity, testing and instrument considerably, as the effects of those variables on experiment and control groups are expected to be the same (Kaptan, 1998, p. 85).In experimental method, the researcher is able to control certain influences, paths or environmental conditions and thus, to observe and understand how objects or the behaviors of the individuals (the subjects) are affected and change.Experimental method is also the ideal method for finding out the cause and effect relations in the research (Kaptan, 1998, pp. 74-75).

Study Group
The study group of this research consists of 48 seventh-grade students (24 in experiment and 24 in control groups) studying at classes 7-B (experiment group) and 7-G (control group) at a state school in the Southeastern Region in Turkey in the spring semester of 2014-2015 school year.The students in experiment group (11 female, 13 male) and control group (10 female, 14 male) are between 12-13 years of age.Since the study group was selected from among groups that had already been formed, convenience sampling was used (Creswell, 2013).As the mathematics lesson academic achievement scores and attitude towards mathematics lesson scores were equivalent between the two groups, one of the groups was randomly assigned as the experiment group, and the other one as the control group.

Data Collection Tools
Mathematics achievement test and attitude towards mathematics lesson scale were used for collecting the research data.

Mathematics Achievement Test
A multiple choice test consisting of 25 items and oriented to "Circle and Sphere" subject in the "Geometry" learning domain of seventh grade mathematics lesson in secondary school was developed.At least 2 questions assessing each of the acquisitions appeared in the test.The test was checked by a mathematics teacher, an expert in mathematics instruction, an expert in assessment and evaluation and a Turkish teacher.It was given its final form after the pilot scheme.KR-20 reliability coefficient of the achievement test was calculated as .82.This value demonstrates the reliability of the test as an assessment tool.

Attitude towards Mathematics Lesson Scale
Attitude Towards Mathematics Lesson Scale (ATMLS): "Mathematics Attitude Scale" developed by Baykul (1990) in order to assess the attitudes of students towards mathematics lesson was used in the study.The scale was devised for the study named "Changes in Attitudes towards Mathematics and Science Lessons from the Fifth Grade of Primary School until the Last Grade of High-Schools and Equivalent Schools and Some Factors Believed to be related to Success in University Entrance Exams".There are 30 items in the scale, 15 of which are positive and 15 are negative.According to the results of factor analysis carried out on the scale, the variance ratio was found .56.Cronbach alpha reliability coefficient of attitude towards mathematics lesson scale was calculated as .96.The items of the scale are in five-point Likert type, with choices as "strongly agree, agree, neither agree nor disagree, disagree, strongly disagree".The calculated reliability value of the attitude scale in this study was .94,which indicates the scale is reliable.

Analysis of the Data
In order to test whether the distribution of the scores obtained from the attitude scale are normal, Shapiro-Wilk test is used if the sample size of the group is less than 50 and Kolmogorov-Smirnov (K-S) test is used if it is larger than 50 (Büyüköztürk, 2014, p. 42).Shapiro-Wilk test was applied in this study, as the sample size was less than 50 and it was found that the observation values did not display normal distribution (Table 1).Since the distribution was not normal, Wilcoxon Signed-Rank Test and Mann Whitney U-Test, which are nonparametric tests, were applied for the evaluation of scores of the mathematics achievement test and attitude towards mathematics lesson scale.

Experimental Procedure
The research was carried out in two separate seventh-grade classes of a state school.Attention was paid that the classes would feature similar qualities in terms of their potentialities, during the selection of the classes.To this end, the former mathematics achievement mean scores of the classes were compared and two classes which had equivalent mathematics achievement were included in the study.One of the classes was assigned as the experiment group and the other as the control group randomly.Both of the groups were applied the mathematics achievement test and attitude towards mathematics lesson scale.During the study, the experimental procedure of which took 4 weeks (5 classes per week, 20 classes in total), manipulatives designed in accordance with the topics of "length in circle and sphere" and "angles" were used with the experimental group by the researchers, whereas the lesson was conducted by means of traditional methods in the control group.Designed teaching materials were tangible and allowed the students to practice and revise sufficiently.
Figure1.The manipulatives used in the study

Findings
Is there a significant difference between pretest scores of experiment and control groups before the application?As shown in Table 2, there was no significant difference between the mathematics achievement mean scores of the experiment and control groups (p>.05).According to this finding, the mathematics achievement levels of the experiment and control groups before the application can be said to be equivalent.
Is there a significant difference between posttest scores of experiment and control groups after the application?According to the findings in Table 3, there was significant difference between the achievement posttest mean scores of experiment and control groups (p<.001).This finding indicates that manipulatives are effective in increasing the achievement scores of the experiment group significantly.
Is there a significant difference between the achievement pretest-posttest scores of the experiment and control groups?Total 24 When Table 4 is analyzed, a statistically significant increase (Z=-4.048,p=000<.001) is observed in the achievement posttest scores ( =77.00) of the experiment group in comparison to its pretest scores ( =47.79).
The increase in achievement posttest scores ( =53.75) of the control group is also found to be statistically significant (Z=-3.250,p=.001<.05)compared to its pretest scores ( =44.13).According to the results of Mann-Whitney U test conducted in order to assess whether there is a significant difference between the attitudes towards mathematics lesson of experiment and control groups before the application, as shown in Table 5, there was no statistically significant difference between mean scores of the experiment and control groups (U=216.500,p>.05).This finding demonstrates that the experiment and control groups had similar attitudes towards mathematics lesson before the application.According to the results of Mann-Whitney U test conducted in order to assess whether there is a significant difference between the attitudes towards mathematics lesson of experiment and control groups after the application, as shown in Table 6, there was a statistically significant difference between the attitude mean scores of the experiment and control groups in favor of the experiment group (U=81.000,p<.001).This finding puts forth that manipulative learning materials are substantially effective in increasing the scores of attitude towards mathematics lesson with the experiment group students.According to the results of Mann-Whitney U test conducted in order to assess whether there is a significant difference between the attitude towards mathematics lesson pretest-posttest scores of the experiment group, as shown in Table 7, there was no statistically significant difference, though the pretest mean scores of the experiment group were higher than the posttest scores (U=212.000,p>.05).This finding indicates that manipulative learning materials increase the scores of attitude towards mathematics lesson in students of the experiment group, but the effect is not significant.
According to the results of Mann-Whitney U test conducted in order to assess whether there is a significant difference between the attitude towards mathematics lesson pretest-posttest scores of the control group, posttest mean scores were observed to decrease significantly compared to the pretest mean scores (U=178.500,p<.05).This finding shows that there was a significant decrease in the attitude of control group students towards mathematics lesson after the application.

Discussion, Results and Suggestions
Various researchers have been carried out recently, investigating the effects of manipulatives in mathematics instruction.According to the findings of those studies, it was observed that manipulatives increase the mathematics achievement (Clements, 1999;Kayma, 2010;Sowell, 1989;Tuncer, 2008).Besides, there are also some other researches suggesting that, students are more active, their motivation for learning is higher and they adopt a positive attitude towards mathematics lesson, when manipulatives are employed in mathematics classes (Enki, 2014;Gürbüz, 2007;Sowell, 1989;Yağcı, 2010).
In reference to the comparison of achievement posttest scores of the experiment and control groups, the high scores in the posttest of the experiment group indicates that the manipulatives prepared for the "circle" and "sphere" topics contribute considerably to effective learning.However, there exist some studies in literature which suggest that manipulatives are not effective on the mathematics achievement of the students (Boakes, 2009;Boyraz, 2008;Enki, 2014;Kaplan, Baran, Işık, Kal, & Hazer, 2013;McClung, 1998;White, 2012).
Based on our knowledge that the levels of attitudes of the experiment and control groups towards mathematics lesson before the application are similar, it was concluded that manipulatives are effective in increasing the scores of attitude towards mathematics lesson for the experiment group.There are other studies demonstrating that manipulatives are effective upon the attitudes of students towards mathematics lesson (Enki, 2014;Sowell, 1989).In the study by Enki (2014), the students pointed out that, as opposed to former traditional instruction methods, learning through activities including the use of manipulatives gave them pleasure and increased their motivation and allowed them to learn while having fun.Positive feedbacks of this kind are indicators of a positive attitude towards the mathematics lesson and the usage of manipulatives in mathematics instruction.
Comparing the pretest-posttest scores of attitude towards mathematics lesson of the control group, it can be seen that posttest mean scores were less than pretest mean scores at a statistically significant level, which shows that there was a significant decrease in the attitude of control group students towards mathematics lesson after the application.The underlying reason may be the diminishing of students' attention towards the mathematics lesson delivered by means of traditional methods, or the change in students' attitudes due to abstract topics.It can also be inferred from the data that the scores of attitude might have decreased because the topics selected for the study are the last ones in the mathematics curriculum before the end of the semester.
In conclusion, with reference to the idea that usage of manipulatives is beneficial for concretizing abstract topics, an increased usage of manipulatives in teaching-learning processes can be suggested.The retention test could not be applied, as the topic selected for the study was the last one before the holiday.The effects of manipulatives upon retention may be investigated by longer-term studies.
1) Is there a significant difference between a. Pretest scores of experiment and control groups b.Posttest scores of experiment and control groups c.Pretest-posttest scores of either group in terms of academic achievement in mathematics lesson?2) Is there a significant difference between d.Pretest scores of experiment and control groups e. Posttest scores of experiment and control groups f.Pretest-posttest scores of either group in terms of their attitudes towards mathematics course?

Table 1 .
Normality test results for the attitude scale

Table 2 .
Comparison of achievement posttest mean scores of experiment and control groups

Table 3 .
Comparison of achievement posttest mean scores of experiment and control groups

Table 4 .
Results of Wilcoxon test for the academic achievement pretest-posttest scores of experiment and control groups

Table 5 .
Comparison of attitude scale pretest scores of experiment and control groups

Table 6 .
Comparison of attitude scale posttest scores of experiment and control groups

Table 7 .
Comparison of attitude scale pretest-posttest scores of experiment and control groups