An Investigation of Mathematics Anxiety of Primary School Teachers

The purpose of this study is to determine the level of mathematics anxiety in primary school teachers in Turkey and to investigate the influence of a number of variables over the anxiety level. 559 teachers were participated in the study. In this study, mathematics Anxiety Scale Intended for Teachers and Candidate Teachers (MKÖ-Ö) was used as the data collection tool. For data analysis, "t" test, analysis of variance (ANOVA), and Tukey test were applied. According to the research results, we have found that total anxiety scores for teachers were at the “I am anxious a little” level. Also, there was a significant difference between teachers’ anxiety level and gender variable, and female teachers’ anxiety levels were found to be higher than male teachers. Additionally, teachers’ anxiety levels decreased as their age and tenure increased. According to the variable of the program of study, the anxiety levels of teachers who graduated from primary school teaching program were lower than the graduated teachers from other programs. Moreover, it is found that the anxiety levels of mathematics teachers who like their job were lower than those who do not like their job.


Introduction
Mathematics, which is taught at all levels of education life and is one of the basic disciplines, is a subject that students are afraid of most and fail the most.These negative responses cause development of mathematics anxiety in time.Richardson and Suinn (1972) stated of mathematics anxiety that it involved "feelings of tension and anxiety that interfere with the manipulation of numbers and the solving of mathematical problems in a wide variety of ordinary life and academic situations".Similarly, Levine (1993) defined the mathematics anxiety as the anxiety felt at the solution of mathematical problems and at the manipulation of numbers during the daily and academic life.
The factors such as the nature of mathematics, quality of teaching, motivation, peer group, the method used by the teacher, not being able to associate with the daily life, unsuitability of topics with the cognitive level of students, students' negative attitude against mathematics, inadequate level of their mathematical pre-knowledge, teacherstudent relationship may be counted among the reasons for development of mathematics anxiety.Several researches have stated that there is a reciprocal relationship between these factors (Baloğlu, 2001;Birgin, Baloğlu, Çatlıoğlu & Gürbüz, 2010;He, 2007;Ma & Xu, 2004;Shields, 2006).
According to some studies that investigated the mathematics anxiety of teachers, mathematics anxiety is also common among teachers and teacher candidates as much as the students and teachers even transfer this anxiety to the students in conscious or unconscious ways (Baloğlu, 2001;Lazarus, 1974;Norwood, 1994;Pantic & Wubbels, 2010).Considering these reasons, analyzing in-service teachers and pre-service teachers' thoughts and beliefs on the role of teacher and mathematics education is important in terms of showing teachers how they should and should not behave in their own classes (Baki & Gökçek, 2007).
When literature was analyzed, it was seen that the mathematics anxiety studies are focused on factors that affects this anxiety.It is seen that the most researched variable is gender.Nevertheless, there is not an exact consensus in research results.While some researchers found that women have higher mathematics anxiety compared to men (Meehan, 2007;Tapia & Marsh, 2004), some researchers could not determine a meaningful difference (Dreger & Aiken, 1957;Fee-Fulkerson, 1983).Considering the studies conducted about teachers, age and tenure are seen as the variables that affect on mathematics anxiety is searched.These studies showed parallel results (Garrosa, Moreno-Jimenez, Liang & González, 2008;Martinussen, Richardsen & Burke, 2007;Suran & Sheridan, 1984).Although some researchers determined that older teachers have higher mathematics anxiety levels than younger teachers (Betz, 1978;Dew, Galassi & Galassi, 1984), some researchers stated that there was a reverse relationship between mathematics anxiety and age and tenure variables (Byrne, 1994;Gold, 1985).In some studies, it is determined that age and tenure variables are not effective in mathematics anxiety (Wiggins, 1984).Not many studies are conducted in the literature in terms of considering the program of study variable.In some studies, it was determined that primary school teachers who were graduated from education faculties were less anxious while comparing to primary school teachers who were graduated from other programs (Gündüz, 2005).Some studies did not find a meaningful difference between graduated program and anxiety (Kilmen & Demirtaşlı, 2009).It was also determined in the literature that teachers who like their profession have less anxiety than the ones who do not like it (Lovett & Cameron, 2011;Scheopner, 2010;Wigfield & Eccles, 2000).Almost all of the studies on this variable have come up with same results.
It is also obvious that the teachers who have mathematics anxiety will create negative opinions and behaviors among students rather than educating students who are innovative, investigative, knowledge producer and user.Considering the potential effect of primary school teachers on their students' attitudes toward math, who have mathematics anxiety, primary school teachers can make a significant positive attitudinal impact on their students.

Aim of the study
When related resources are researched, it can be seen that very few studies have analyzed to the teachers' mathematics anxiety.This shows that current research is an original study.Starting from this point, the study aimed to determine the mathematics anxiety level of teachers and whether their mathematics anxiety differs depending on some variables.For this purpose, the answers to following questions were sought: 1-What is the level of teachers' mathematics anxiety?2-Does teachers' mathematics anxiety differ depending on gender, age, tenure and the program of study?
3-Does teachers' mathematics anxiety vary by the level of their job satisfaction?

Research model
In this study, survey model, which is among the quantitative data collection methods, was used.Survey models are research approaches which aim to depict situations occurred in the past or existing now as they are.The event, individual, or the object of the research is tried to be defined in its own circumstances and as it is.No effort to change and affect them is shown (Karasar, 2002).

Research group
The study group of this research was composed of 559 randomly selected primary school teachers who working in the primary schools in Turkey.Criterion sampling method, which is a part of the purpose sampling methods, was used for determining the sample.When selecting primary school teachers' different criteria such as age, tenure and gender etc. have been taken into consideration.The observation units in criterion sampling can be composed persons, events, objects and situations from particular qualities (Yıldırım & Şimşek, 2005).The participants in the study were 289 female, 270 male teachers.

Data collection tool
Mathematics Anxiety Scale 'Intended for Teachers and Candidate Teachers' (MKÖ-Ö) was used as the data collection tool in this study.The first part of the questionnaire consists of demographics questions while the second part includes mathematics anxiety scale about the anxiety towards mathematics lesson developed by Üldaş (2005).
MKÖ-Ö is a scale to measure the mathematics anxiety which is composed of 39 items and 7 subscales.Scale and subscales were found to be a valid and reliable scale (Üldaş, 2005).These subscales expresses the anxiety emerging from mathematics, regarding mathematics teaching, learning, solving problems (problem solving, arithmetic operation), doing mathematical interpretations, making mathematical mistakes, and mathematical selfadequacy.The mathematics understanding anxiety subscale is concerned with understanding and learning mathematical concepts and problems.The mathematics description anxiety subscale is concerned with the anxiety that arises when teaching or expressing mathematics.The problem solving anxiety subscale reflects the anxiety that arises when dealing with a mathematical problem.The arithmetical operation anxiety subscale indicates anxiety arising in mathematics-related situations where four processes are involved or where numbers are involved.The self-adequacy anxiety subscale is anxiety about mathematical competence.The interpretation anxiety subscale is concerned with analyzing and synthesizing mathematical data.The making mistake anxiety subscale contains anxiety about making mistakes while doing mathematical operations.MKÖ-Ö is a four-point likert-scale questionnaire graded from 1 to 4. The Cronbach's alpha for the reliability of MKÖ-Ö was .90.The Cronbach alpha values of the subscales were found as follows: mathematics understanding .87,mathematics description .81,problem solving .80,arithmetical operation .83,self-adequacy .78,interpretation .85,making mistake .81.

Data analysis
During data analysis, frequency (f) and percentage (%) techniques were used for descriptive statistics analysis about personal information.The questionnaire items were scaled and rated as "I am not anxious (1)", "I am anxious a little (2)", "I am anxious (3)", "I am very anxious (4)".The limits of these values are as follows: Table 1.Questionnaire choices gravity and limits The data collected from the questionnaire that was administered to teachers are analyzed by the help of a statistical software using "t" test, analysis of variance (ANOVA) and Tukey test.

Findings and comments related to anxiety levels of teachers of mathematics
Arithmetic mean and standard deviation scores that were computed from the whole scale and subscales are presented in Table 2.When the teachers' mathematics anxiety scores were analyzed, it was seen that the total of their anxiety scores is 2.13, which means their overall anxiety level was at the level of "I am anxious a little".When anxiety scores at the subscales were analyzed, it was seen that understanding mathematics anxiety (1.42) was at "I am not anxious" level, description mathematics anxiety (2.41), arithmetic operation anxiety (2.19), mathematical interpretation anxiety (1.91) and making mistakes anxiety (2.13) was at "I am anxious a little" level.Teachers' solving problems anxiety (2.63) and teachers' self-adequacy anxiety (2.50) was at "I am anxious" level.

Findings and comments related to the analysis of teachers' mathematics anxiety levels based on various variables
This part includes the findings and comments on whether the teachers' mathematics anxiety levels is statistically significant based on gender, age, tenure variables, program of study and job satisfaction.On Table 3, data regarding teachers' total scores from the total scale and subscales based on their gender with independent group t-test are presented.In the independent group t-test carried out for mathematics anxiety averages based on gender variable, a statistically significant difference at the level of p<0.05 was found at the subscales of Description Mathematics, Self-Adequacy, Making Mistakes.This difference was in favor of male teachers in all three subscales.Thus, mathematics anxiety levels of female teachers were higher than those of male teachers in these subscales.Again in the total scale, a significant difference at the level of p<0.05 was found and this difference also was in favor of male teachers.Further, no significant difference was observed between male and female teachers at the subscales of Understanding Mathematics, Problem Solving, Arithmetic Operations and Mathematical Interpretation.But, the anxiety scores of male teachers were lower than the female teachers at the subscales except for the Arithmetic Operations.
On Table 4, data regarding teachers' total scores from the total scale and subscales according to their age gathered with one-way analysis of variance and Tukey test are presented.Note.The letters a, b and c show group differences for the same subscale (p<0.05).
In the ANOVA test carried out for the mathematics anxiety averages based on age variable, a statistically significant difference at the level of p<0.05 was found for the subscales of Description Mathematics, Self-Adequacy, Interpretation and Making Mistakes.Also in the total scale, a significant difference at the level of p<0.05 was found.There were no significant differences in the other subscales.
In the total mathematics anxiety scores on age variable, teachers aged between 20-29 had the highest anxiety scores (M=2.26),whereas, teachers aged 50 and more had the lowest anxiety scores (M=2.02).The total mathematics anxiety of teachers declines as they get older.
According to the data obtained from the Tukey test, the scores of the total mathematics anxiety do not have any statistically significant difference between teachers who aged 20-29 and 30-39.Also, the scores of the total mathematics anxiety do not have any statistically significant difference between teachers who aged 40-49 and 50 and over.But, there was a significant difference between teachers aged 20-39 and 40 and over.
On Table 5, the data are presented regarding to teachers' total scores that from the total scale and subscales based on their tenure gathered with one-way analysis of variance and Tukey test.Note.The letters a, b and c show group differences for the same subscale (p<0.05).
In the ANOVA test carried out for mathematics anxiety averages based on tenure variable, a statistically significant difference at the level of p<0.05 was found for the subscales of Description Mathematics, Self-Adequacy and Making Mistakes.Also in the total scale, a significant difference at the level of p<0.05 was found.There were no significant differences in the other subscales.
While the teachers with a tenure of 1-5 years had the highest (M=2.24)anxiety scores among the total mathematics anxiety scores regarding their tenure, the teachers with a tenure of 21 years and more had the lowest (M=2.02)anxiety scores.When teachers get older, their total mathematics anxiety levels decline.Thus, it can be said that when experience increases, mathematics anxiety reduces.
According to the data obtained from the Tukey test, the total teachers' mathematics anxiety scores with tenure of 1-5 years and 6-10 years do not have any statistically significant difference.Also, teachers' total mathematics anxiety scores with tenure of 11-15 years, 16-20 years and 21 and more years do not have any statistically significant difference.However, there is a significant difference between teachers with tenure of 1-10 years and teachers with tenure of 11 and more years.Based on these findings and teachers' total anxiety scores, it can be said that experienced teachers have less mathematics anxiety.
Data gathered by the independent group t-test applied to average scores teachers had from the whole scale and the subscales regarding the program of study are presented on Table 6.In the independent group t-test carried out for mathematics anxiety scores based on program of study variable, a statistically significant difference at the level of p<0.05 was found for the subscales of Understanding Mathematics, Description Mathematics, Self-Adequacy, Mathematical Interpretation and Making Mistakes.This difference was in favor of teachers who were graduates of primary school teaching program for all five subscales.Thus, the anxiety levels of primary school teaching graduates in these subscales were less than those of teachers who were graduates of other programs.Again, a significant difference at the level of p<0.05 was found in the whole scale and this difference was also in favor of primary school teaching graduates.In the other subscales, no significant difference between the primary school teaching graduates and teachers who graduated from other programs was found.
Besides, when anxiety scores in all the subscales were analyzed, it was seen that the primary school teaching graduates have lower mathematics anxiety scores than the graduates of other programs.Only in the Arithmetic Operations subscale, the anxiety scores of primary school teaching graduates are higher than other programs.However, this difference was statistically not different.Data gathered by the one-way variant analysis and Tukey test applied to average scores teachers got from the whole scale and the subscales regarding the job satisfaction are presented on Table 7 Note.The letters a, b and c show group differences for the same subscale (p<0.05).
In the ANOVA test carried out for mathematics anxiety averages based on job satisfaction variable, a statistically significant difference at the level of p<0.05 was found in all subscales except Arithmetical Operation.Again in the total scale, a significant difference at the level of p<0.05 was found.
While teachers who were not satisfied with their job had the highest anxiety scores (M=2.41) in total mathematics anxiety scores regarding job satisfaction, the teachers who were satisfied with their job had the lowest anxiety scores (M=1.86).When the teachers' job satisfaction reduces, their total mathematics anxiety increases.According to the data obtained from the Tukey test, there was a statistically significant difference between the total mathematics anxiety scores of teachers who were satisfied, were partially satisfied, and were not satisfied with their job.

Discussion and Conclusion
When the mathematics anxiety scores of teachers were analyzed, it was seen that teachers' total anxiety score is 2.13, meaning of that is the level of "I am anxious a little".In similar studies, the mathematics anxiety scores of teachers were found to be low (Fennema, Peterson, Carpenter & Lubinski, 1990;McAnallen, 2010;Perkkila, 2003;Uusimaki & Nason, 2004;Üldaş, 2005).In a study conducted by Norwood (1994), the researcher reached to the result that the majority of teachers who are responsible for mathematics education have mathematics anxiety and they could transfer their anxiety to their students.In this context, as a positive result, it can be stated that the total anxiety results of teachers are low in our research.
When the anxiety scores in the subscales were analyzed, it was found out that understanding mathematics anxiety (1.42) was at "I am not anxious" level, while description mathematics anxiety (2.41), arithmetic operation anxiety (2.19), mathematical interpretation anxiety (1.91) and making mistakes anxiety (2.13) were at "I am anxious a little" level.Teachers' Problem solving anxiety level (2.63) and self-adequacy anxiety level (2.50) were found to be at "I am anxious" level.
A significant difference at the level of p<0.05 was found between the mathematics anxiety of teachers and the gender variable, and this difference was in favor of male teachers.Thus, the anxiety levels of female teachers are higher than male teachers.When related literature was reviewed, it was seen that gender was the most researched variable.Nevertheless, there is not still an exact reconciliation in the research results.For example, many researchers (Alexander & Martray, 1989;Bander & Betz, 1981;Benson, 1989;Brush, 1980;Frost, Hyde & Fennema, 1994;Haynes, Mullins & Stein, 2004;Meece, 1981;Meehan, 2007;Şahin, 2008;Tapia & Marsh, 2004;Tobias, 1978Tobias, , 1986Tobias, , 1990;;Zettle & Raines, 2000) identified that women have higher mathematics anxiety than men.These results were consistent with the results of our research.
However, some research results identified that there was not a significant difference between mathematics anxiety and gender variable (Akın, 2002;Bindak, 2005;Brush, 1978;Cooper & Robinson, 1991;Dreger & Aiken, 1957;Fee-Fulkerson, 1983;Ling, 1982;Oropesa, 1993).Also, in a study conducted by Linn and Kessel (1996), a significant difference favoring the girls was found between the gender variable and mathematics anxiety.When the mathematics anxiety scores in the subscales regarding the gender were analyzed, a statistically significant difference at the level of p<0.05 was found in the subscales of Description Mathematics, Self-Adequacy, Making Mistakes.This difference was in favor of male teachers in all three subscales.Thus, the anxiety levels of female teachers in these subscales were higher compared to male teachers.Again in the whole scale, a significant difference at the level of p<0.05 was found and this difference was also in favor of male teachers.
No significant difference was found between male and female teachers in the subscales of Understanding Mathematics, Solving Problems, Arithmetic Operations and Mathematical Interpretation.Still, in all subscales except for the Arithmetic Operations, anxiety scores of male teachers were lower than anxiety scores of female teachers.The reason for this may be the self-confidence which exists in the nature of men more than women.Or, it can be thought that men take pleasure from the mathematical activities more than women.
It was determined that there was a significant difference at the level of p<0.05 between the mathematics anxiety of teachers and their age and tenure.When teachers get older and their tenure increases, their total mathematics anxiety reduces.Among the reasons why young teachers have more anxiety are that they cannot cope with problems, may not engage in their job, and have lack of experience.However, it can be said that the increase in professional experience and maturity of teachers, the improvement in their ability to decide, the increase in their personal success and being more resistant to anxiety are effective in the reduction of anxiety while getting older.In parallel, Üldaş (2005) found out in his research that there is a statistically significant and inversely proportional relation between the mathematics anxiety scores of teachers and age variable.Again in parallel, Girgin (1995) reached the same result on primary teachers that anxiety decreases when age increases.Similar to these two studies, it is possible to encounter many studies which determine the result that anxiety decreases when age and tenure increases (Aslan et al., 1997;Byrne, 1994;Garrosa et al., 2008;Gold, 1985;Martinussen et al., 2007;Suran & Sheridan, 1984).The reason for this may be that people are more realistic when they are young but when they get older, they are more likely to speak as they are supposed to speak and answer as they are supposed to answer.In the contrary, some studies reached to the result that anxiety will increase with the increase of age and tenure (Betz, 1978;Dew, Galassi & Galassi, 1984;Richardson & Suinn, 1972).Also, some studies cannot find a meaningful relation between these variables and anxiety (Wiggins, 1984).Also, based on the tenure variable, when the mathematics anxiety score averages in the subscales were analyzed, no statistically significant difference at the level of p<0.05 was found in Understanding Mathematics, Solving Problems, Arithmetic Operations and Mathematical Interpretation subscales.Whereas in Description Mathematics, Self-Adequacy and Making Mistakes subscales, a significant difference at the level of p<0.05 was found in parallel with the age variable results.Thus, when the teachers get older, their mathematics anxiety scores in these subscales decrease.
Based on program of study variable, a significant difference at the level of p<0.05 was found.Thus, a significant difference in favor of primary school teaching graduates was found between the primary school teaching graduates and the graduates of other programs.It can be argued that primary school teachers' beliefs on the education policies and qualifications may differ based on what program they graduated.It can also be said that the mathematics anxiety of primary school teachers have decreased with the lesson of mathematics teaching.Not many studies were carried out in the literature on program of study variable.In his study, Gündüz (2005) determined that the primary school teachers graduated from education faculties are less anxious compared to primary school teachers graduated from other programs.This study supports the results of that research.Some studies did not find a significant difference between program variable and anxiety (Kilmen & Demirtaşlı, 2009;Kırılmaz, Çelen & Sarp, 2003).Some researchers reached the result that primary school teaching graduates were more anxious than the primary school teachers who graduated from other faculties (Ertürk & Keçecioğlu, 2012).Furthermore, when the mathematics anxiety score averages regarding the program of study variable were analyzed, a statistically significant difference at the level of p<0.05 was found in the subscales of Understanding Mathematics, Description Mathematics, Self-Adequacy, Mathematical Interpretation and Making Mistake.This difference was in favor of primary school teaching graduate teachers in all five subscales.Thus, the anxiety levels of primary school teaching graduates in these subscales were less than teachers who graduated from other programs.No significant difference was found between primary school teaching graduate teachers and teachers who graduated from other programs in the subscales of Problem Solving and Arithmetic Operations.
Based on the job satisfaction variable, a statistically significant difference at the level of p<0.05 was found in mathematics anxiety.It was seen that the anxiety levels of the teachers who like their job are lower than those who do not like it.The results of our study are consistent with the findings of previous research.The negative attitudes developed against the profession emerge as anxiety.It was observed that the anxiety levels of the teachers who like their job were much lower than the teachers who like profession less or do not like their profession (Bruinsma & Jansen, 2010;Eren, 2012;Gündüz, 2005;Lovett & Cameron, 2011;Scheopner, 2010;Tuğrul & Çelik, 2002;Wigfield & Eccles, 2000).Also regarding the job satisfaction variable, when the mathematics anxiety score averages in the subscales were analyzed, a statistically significant difference at the level of p<0.05 was found in the subscales of Understanding Mathematics, Solving Problems, Description Mathematics, Self-Adequacy, Mathematical Interpretation and Making Mistakes.This difference was in favor of teachers who were satisfied with their job in all six subscales as expected.Namely, when the job satisfaction of teachers increases, their anxiety decreases.On the other hand, when the arithmetic problems anxiety scores regarding the job satisfaction were analyzed, no statistically significant difference at the level of p<0.05 was observed.Based on all these results, starting from the point that the young teachers with a less tenure of office have more mathematics anxiety, these teachers may be supported to decrease their anxiety.It was seen that classroom teachers who graduated from classroom teaching programs have less anxiety than classroom teachers who graduated from other programs.Because of this, it may be afforded that classroom teacher assignments are not made from the programs out of the field.It was determined that the professional satisfaction of teachers directly affect the educational studies.Because of this, necessary amendments about economic, social, physical and employee rights issues to increase the professional satisfaction of teachers may be made.A relation was found between the program a classroom teacher graduates from and their anxiety level.The effects of teachers being graduated from different programs on their anxiety can be researched by using qualitative research methods.It was seen that most of the mathematics anxiety studies were made on students and candidate teachers.Because of this, mathematics anxiety studies especially on teachers may be made.

Table 2 .
Mathematics anxiety scores of the teachers (n=559)

Table 3 .
T-Test Based on Gender

Table 6 .
T-Test based on the program of study .