Education and Science

Education and Science
Vol 39 (2014) No 176 145-157
The Adaptation of the Sources of Mathematics Self-Efficacy Scale for
Turkish Context *
Eyüp Yurt 1, Ali Murat Sünbül 2
Abstract
Keywords
The aim of the study is to adapt the Sources of Middle School
Mathematics Self-Efficacy Scale developed by, Usher and Pajares
(2009) for Turkish context. After the Turkish version was formed
through required procedures, it was administered with 750
middle school students of 6th, 7th and 8th grades. The ages of the
students ranged between 12 and 15. The 48% of the students were
female (n=408) and 52%of the students were male (n=342). 32% of
the students were attending (n=242) sixth grade, 34’% of them
(n=257) seventh and 34% (n=251) eighth grade. Construct validity
of the scale was investigated via Exploratory and Confirmatory
Factor Analysis. Criterion validity of the scale was also done to
see if it serves its purpose well. Reliability of the scale was tested
by computing Cronbach Alpha, corrected total item correlation
coefficient, and t-tests comparing the total item scores of top 27%
and bottom 27% participants. The results indicated that the
Turkish version of the SMSMSEC consists of four factors.
Cronbach alpha values of the factors ranged between 0.80 and
0.94. Corrected total item correlation coefficients ranged between
0.77 and -0.25. Criterion validity results indicated the scale serves
its purpose well. The t-test results indicated that there were
significant difference between the total scores of top 27% and
bottom 27% of the participants for all items.
Sources of Self-efficacy
Social Cognitive Theory
Self- efficacy
Mathematics
Middle School
Article Info
Received: 06.09.2014
Accepted: 11.17.2014
Online Published: 12.16.2014
DOI: 10.15390/EB.2014.3442
Introduction
Self-efficacy is used in recent studies on learning and motivation more often than other
concepts such as self-identity and self-respect (Şahin, 2013). The reason for this can be that self-efficacy
can explain the students’ performances more compared to other concepts available in the literature
(Bong and Clark, 1999; Bong and Skaalvik, 2003; Ferla, Valcke and Cai, 2009). Self-efficacy refers to a
person’s belief related to his or her capacity to perform a task well (Bandura, 1997). Self-efficacy is
about being aware of what a person can accomplish rather than being aware of what they want to
accomplish (Senemoğlu, 2007). Research on self-efficacy has indicated that self efficcay has an impact
on indivual’s choice of activity, motives, persistence, learning and achivements (Bandura, 1997;
Schunk and Pajares, 2005; Schunk and Zimmerman, 1998). People tend to choose the activities they
This study was developed from the first author’s doctoral dissertation and was presented as a oral presentation at the 2 th
WCEIS conference.
1 Necmettin Erbakan University, A. K. Education Faculty, Education Science Department, Turkey, eyupyurt@gmail.com
2 Necmettin Erbakan University, A. K. Education Faculty, Education Science Department, Turkey, asunbul@konya.edu.tr
*
145
Education and Science 2014, Vol 39, No 176, 145-157
E. Yurt, A. M. Sünbül
think they will succeed and tend to avoid the ones they think they will fail (Bandura, 1997). On the
other hand, they set big goals for themselves in areas where they have high self-efficacy (Bandura,
1997). People with high self-efficay also exhibit high level of effort while they are working on task and
they exhibit high level of persistence when they meet obstacles (Bandura, 1997). There are studies
(Schunk and Pajares, 2005; Zimmerman and Kitsantas, 1999; Collins, 1982, as cited in Schunk and
Pajares, 2009, s. 39) which indicate that students who believe they can achieve are more likely to
complete a task than the ones who do not believe they can achieve even if both group of students have
similar or same level of ability. In other words, in situations where people have same abilities in the
outset of a task, people with high self-efficacy tend to learn more and be more successful. Studies have
shown that self-efficacy plays a major role both in academic success at any level and in all types of
successful behaviors (Schunk, 2011).
Bandura (1997) states that a person’s self-efficacy comes from four sources. These sources are
personal experiences, vicarious experiences, social persuasions, emotional and physiological states.
Personal experiences have a permanent impact on self-efficacy of a person. As a result of this, personal
experiences are the most prominent source of a person’s self-efficacy (Bandura, 1997). After the
completion of a task, a person evaluates his performance. If his evaluation is positive, his confidence
boosts regarding his capacity and beliefs of completing similar tasks in the future. If the evaluation is
negative, his confidence wanes regarding similar future tasks. For example, if a student who always
gets AA on tests got BB on a particular test despite studying hard, he would feel disappointed. This
particular student may start to doubt himself. On the other hand, when a student who always gets CC
on a particular course’s tests got BB, this student tends to develop confidence in his success on that
particular course. In short, people’s personal experiences affect their self-efficacy either positively or
negatively (Bandura, 1997).
Another source of self-efficacy is vicarious experiences, which people gain by observing others
performing particular tasks. Students constantly observe significant people such as their parents,
teachers, siblings and peers and these observations affect their self-efficacy. For example, if a student’s
classmate who gets AA on a test, his own self-efficacy increases because he believes that he could also
get a similar mark on the exam. Vicarious experiences may also have negative impact on a person’s
self-efficacy. If the student observes that his peers have failed in a task, he tends to believe that he
would fail in that particular task too. This situation applies especially when the person has limited
relevant experience or lacks any judgment regarding his capacity to perform the task (Bandura, 1997).
The third source of self-efficacy is social persuasions. People’s self-efficacy may increase or
decrease as a result of encouragement and discouragement of their parents, teachers and friends. For
example, teachers’ encouragement may increase students’ self-efficacy for academic goals and
achievement. Students need encouragement of their teachers and families especially when they
believe their capacity to complete a task is not good enough. However, over- encouragement for a
person in tasks beyond his capacity may lead to future mistakes and lessening of self-efficacy
(Bandura, 1997).
The final source of self-efficacy is emotional and physiological states. People tend to reach a
judgment about their capacity based on their physical responses in stressful situations. People who
possess similar capacities and skills may exhibit different physical responses in stressful situations.
These physical responses in stressful situations alter a person’s self-efficacy. People who can control
these responses get stressed less and their self- efficacy remains unchanged (Bandura, 1997).
Studies on self-efficacy in the literature mostly focus on high school and university students
(Usher, 2009). Studies on self-efficacy in the Turkish context have investigated self-efficacy of teachers
and students at faculties of education or prospective teachers (Akbaş and Çelikkaleli, 2006; Azar, 2010;
Çalışkan, Selçuk and Özcan, 2010; Coşkun, 2010; Çapri and Çelikkaleli, 2008; Çetin, 2008; Durdukoca,
2010; Ekici, 2006; İpek and Acuner, 2011; Morgil, Seçken and Yücel, 2004; Maden, 2010; Terzi and
Mirasyedioğlu, 2009; Yaman, Koray and Altunçekiç, 2004; Yılmaz and Çimen, 2008; Yılmaz, Yılmaz
146
Education and Science 2014, Vol 39, No 176, 145-157
E. Yurt, A. M. Sünbül
and Türk, 2010). There are very few studies (Arslan, 2012; Arslan, 2013; Çetin, 2009) investigating selfefficacy of students at middle school. One of the reasons is due to a lack of suitable assessment
instrument to evaluate middle school students’ self-efficacy. For example, the literature review
revealed that there are two instruments for assessing middle school students’ self-efficacy in
mathematics (Işıksal and Aşkar, 2003; Karadeniz, Büyüköztürk, Akgün, Çakmak and Demirel, 2008).
One of these instruments (Işıksal and Aşkar, 2003), aims to assess self-efficacy of students in specific
areas of mathematics (e.g. symmetry and equations). The other one aims to assess students’ selfefficacy for self-regulated mathematic skills (Karadeniz et al., 2008). The literature review also
revealed that there is no Turkish self-efficacy scale which aims to measure mathematic self-efficacy
sources based on the sources of self-efficacy pointed out by Bandura (1997). Therefore, this study aims
to adapt the Sources of Middle School Mathematics Self-Efficacy Scale developed by Usher and
Pajares (2009) for the Turkish context.
Many studies have indicated that self-efficacy plays an important role in students’ academic
success (Schunk and Zimmerman, 1998; Usher and Pajares, 2006; Schunk, 2011). Additionally, selfefficacy is an inherent component of every type of successful behavior (Schunk, 2011). Adopted
measurement instrument will help to identify the sources of middle school mathematics self-efficacy
scale. Thus, the relations of selfefficacy sources with the cognitive, affective and motivational variables
related to mathematics can be examined in a variety of research. In the light of this research both
theoretical and practical knowledge will be offered. It is hoped that this study may close a gap in the
literature regarding a lack of Turkish version of Sources of Middle School Mathematics Self-Efficacy
Scale.
Method
Participants
The participants of the study are 750 middle school students. 48% of the participant are female
(n=342) , 52% of the participants are male ( n=408). 32% of the participants are 6th grade (n=242); 34%of
them are 7th grade (n=257) and 34% of them are 8th grade students (n=251). The participants’ ages
range between 12-15 years. Data from 266 students was used for Exploratory Factor Analysis; data
from 254 was used for Confirmatory Factor Analysis and data from 230 students were used for
criterion validitym.
The instrument
The English version of the Sources of Middle School Mathematics Self-Efficacy Scale
(SMSMSEC thereafter) was taken from the article by Usher and Pajares (2009). The SMSMSEC consists
of 24 items; 6 items for personal experiences, 6 items for vicarious experiences, 6 items for social
persuasions and finally 6 items for emotional and physiological states. Items 3, 19, 20, 21, 22, 23, and
24 on the scale are reversed scored and the rest of them are positive scored items. Answering the scale
takes about 15 to 25 minutes. The respondents of the scale are required to indicate on a Likert scale of
definitely disagree (1) to definitely agree (100).
The originals scale’s construct validity was investigated via Exploratory Factor Analysis,
Confirmatory Factor Analysis. For its construct validity, A. Bandura, B. J. Zimmerman and D. H.
Schunk were consulted. It was found that structure of the scale with four factors had the best fit for
the model. To investigate the criterion validity of the scale, scores of areas such as self-concept, math
skills self-efficacy, math courses self-efficacy, self-regulatory self-efficacy, task-goals and Semester
GPA were taken into consideration. The correlational values of those selected scores and four factors
in the SMSMSEC ranged between 0.88 and -0.65. The internal consistency of the original scale was
tested via Cronbach alpha. The alpha values for four factors in the scale ranged between, 0.84 and 0.88
(Usher and Pajares, 2009).
147
Education and Science 2014, Vol 39, No 176, 145-157
E. Yurt, A. M. Sünbül
Analyzis of the Data and Procedures
As in the original study, Exploratory Factor Analysis and Confirmatory Factor Analysis were
done using the scores gathered from Turkish students to investigate the factorial structure of the scale.
By this way, construct validity of the scale was investigated via two types of factorial analyses.
Exploratory factor analysis (EFA) aims to uncover the underlying structure of a relatively
large set of variables which are in interaction with each other (Stevens, 2009; Çokluk, Şekercioğlu and
Büyüköztürk, 2010). EFA reveals which factor covers a particular item and the strength of its relation
to the main factor via factor loading values. Items factor loading value is required to be 0.45 and
above. However, items with a 0.30 factor loading value can remain in the scale (Kline, 2011). There are
many techniques to reveal factors in factor analysis. These techniques can be listed as principle
component analysis, image factor analysis, maximum likelihood estimation factor analysis, basic
factor analysis. The most widely used and the easiest one to interpret is principle component analysis
(Büyüköztürk, 2002). Since it is intended to gather the variables under a smaller number of
components reducing the number of them, principal component analysis was used (Tabachnick and
Fidell, 2007). SPSS 18.0 was used to compute the exploratory factor analysis.
Confirmatory Factor Analysis is used to test whether data fits a measurement model of factors
and their indicators (Çokluk et al., 2010). In other words, it is used to test whether measures of a
construct are consistent with a researcher's understanding of the nature of that construct or factor.
AMOS 19.0 was used to compute Confirmatory Factor Analysis.
In Confirmatory Factor Analysis, many fit indices are used to test the adequacy of the model.
Fit indices are used to test the fit between actual data and the hypothesized model. To test the fit of the
model, using several fit indices at the same time is advised because of their relative weaknesses and
strengths (Kline, 2011). Additionally, there are four fit indices which are advised to be reported in
SEM analysis (Kline, 2011). These are Chi-Square Goodness, (2); Root Mean Square Error of
Approximation (RMSEA); Standardized Root Mean Square Residual (SRMR); and Comparative Fit
Index (CFI). There are additional indices such as Goodness of Fit Index (GFI); Adjusted Goodness of
Fit Index (AGFI); Incremental Fit Index (IFI) and Normed Fit Index (NFI) (Bollen, 1989; Bentler, 1992;
Anderson ve Gerbing, 1984; Jöreskog ve Sörbom, 1993; Cole, 1987; Sümer, 2000). In the present study
all of the fit indices mentioned above were computed.
As a first step, an un-rotated principle component analysis was computed to investigate the
factor structure of the scale. Second, equimax rotation was used to interpret factors more easily.
Equimax rotation, which is a hybrid of varimax and quartimax rotations and rotates simultaneously,
simplifies factors and variables (Çokluk et al., 2010). After the completion of exploratory factor
analysis, Confirmatory Factor Analysis was done to see if the original scale is valid in the Turkish
context. In Confirmatory Factor Analysis, maximum likelihood estimation factor analysis was
computed.
Additionally, to determine the scales suitability in the Turkish context, criterion validity was
carried out. As a result of this aim, scores from four factors of the scale, the mathematic anxiety scale
and self-efficacy scales were used to compute Pearson's product-moment coefficient.
Corrected item-total correlation was computed to test the items’ facilityto distinguish people
in terms of their characteristics assessed in the scale and to investigate the test's internal consistency, ttest was investigated to see if there is a significant difference between top 27% and bottom 27 % of the
participants based on the total scores (Büyüköztürk, 2011). Cronbach Alpha was computed to test the
scale’s internal reliability. Additionally, factors’ means and standard deviation values of the scale
were calculated. Finally, correlations between factors were computed using Pearson's productmoment coefficient.
148
Education and Science 2014, Vol 39, No 176, 145-157
E. Yurt, A. M. Sünbül
Ethical issues and conducts of scale development process
Required permission for adaptation of the scale was obtained by e-mail from Ellen Usher who
was the correspondent author of the original scale development study. In order to perform validity
and reliability tests, required application permissions were obtained from the Ministry of National
Education office in Konya. Besides, willingness of the participants was a priority in filling out the scale
form.
Results
Translation of the SMSMSEC into Turkish
The original English scale was translated in to Turkish separately by three professional
interpreters. The translated versions of the scale were examined by 3 experts in the field of
psychological assessment end evaluation. A tentative version of the scale was formed based on their
recommendations. As a next step, this tentative version was examined in terms of its suitability in
cultural context, linguistics, scientific methods and evaluation by experts using Expert Evaluation
Form (EEF). EEF consisted of two sections. The first section gave general information about the study
and a brief literature review on the theoretical underpinning of the scale. The second section consisted
of Likert scale items corresponding to the each evaluation criteria. The Likert scale ranged between 1
“this item is definitely unsuitable” to 5 “this item is definitely suitable”. There was also space for
suggested corrections. Based on the EEF, items with a mean score of 4.0 above and standard deviation
of 0.7 and below were selected. This version of the scale was back translated into English by a linguist
and an education specialist. Both Turkish and English translations were checked against the original
form by two English instructors at a Turkish University. They indicated that both forms express the
same ideas in the original Scale.
Results of the Exploratory Factor Analysis
One assumption of Exploratory Factor Analysis is sampling adequacy (Çokluk et al., 2010).
To test the sampling adequacy, Kaiser-Meyer-Olkin Coefficient was calculated and it was 0,932. Based
on this result, assumption of sampling adequacy was met (Tavşancıl, 2010). Another assumption of
the Exploratory Factor Analysis is normal distribution of the data (Çokluk et al., 2010). The
distribution of the data was tested via Bartlett test to compute its chi-square value (2=4927.7, p<0.001).
The results indicated the data gathered met the assumptions of exploratory factor analysis. The
computation of exploratory factor analysis showed that there were four factors whose eigenvalue is
above 1. Additionally 5 of the items got high loading in more than one factor. To boost the loading
value of the items, factor analysis was computed again using Equamax rotation technique. The results
of this procedure showed that the present data showed a model similar to the one in the original
study. The results of the factor analysis are presented in table 1.
149
Education and Science 2014, Vol 39, No 176, 145-157
E. Yurt, A. M. Sünbül
The percentage of variance that was explained by the scale with four factors was 69%. Factors’
eigenvalues and their percentage of variances were 10.51 and 43.78%; 3.11 and 12.96%; 1.78 and 7.40%;
1.20 and 4.99% respectively. These results indicated that items were similar to the ones in the original
scale with four factors.
Table 1. Results of the Exploratory Factor Analysis of the Sources of Middle School
Mathematics Self-Efficacy Scale
Factor Loading Values *
Total Factor
Items
Variance
Faktor 1
Faktor 2
Faktor 3
Faktor 4
i4
,69
,32
,75
i2
,80
,49
-,22
,70
i1
,74
,46
-,21
,69
i3
,62
-,40
,65
i6
,66
,44
,65
i5
,60
,45
,60
i10
,65
,23
,77
i8
,60
,74
i9
,60
,72
i7
,57
,23
,72
i11
,60
,28
,38
,61
i12
,31
,30
,45
i15
,83
,21
-,24
,83
i14
,80
,24
,83
i16
,80
,28
,31
,77
i17
,77
,39
,27
-,20
,71
i13
,72
,33
,28
-,26
,68
i18
,60
,34
,31
,62
i23
,83
-,21
-,23
,85
i21
,78
,85
i22
,82
-,30
,84
i24
,80
-,21
-,24
,83
i19
,69
,81
i20
,70
,80
*indicates values below ±0, 20
Results of the Confirmatory Factor Analysis
The SMSMSEC, which was developed based on Bandura’s Social Cognitive Theory (1997) and
was considered acceptable by Turkish experts, was investigated via Confirmatory Factor Analysis. As
in the original study, its fit for the four- factor model was computed. The fit indices (2=825.67,
p<0.001, CFI=0.88, GFI=0.78, AGFI=0.74, RMSE=0.09, SRMR=0.07, NFI=0.83) showed that the model
does not fit the data well. When calculated modifications indices are examined, error covariance of
some items (i4-i5, i1-i2, i9-i11, i7-i8, i10-i12, i10-i8, i13-i14, i13-i16, i14-i16, i15-i18, i16-i17, i19-i22, i22i23, and i23-i24) were found to be related at a significant level. These pairs of items were also found to
be under the latent factor in the original study. After consultation with an expert, the model was
tested again by feeding the error correlations between these items into the model. The results of the
final confirmatory factor analysis are presented in figure 1. The items’ factor loading values ranged
between 0.42 and 0.94 and all items factor loading values are statistically significant (p<0.001).
150
Education and Science 2014, Vol 39, No 176, 145-157
E. Yurt, A. M. Sünbül
Figure 1. CFA Result for the Four Factor Model, n= 254,2=488.15, N=254, sd=233, p<0.00.1
When the fit indices of the model is examined, chi-square value is significant (2=488.15,
N=254, sd=233, p<0.001). However, as the sampling size increases so does the likelihood of having a
significant chi-square value and to counter this situation, looking at ratio of 2/sd is suggested in the
literature (Büyüköztürk, Akgün, Özkahveci and Demirel, 2004). The calculated ratio of 2/sd was 2.20.
Fit indices values were RMSEA=0.07, SRMR=0.07, CFI=0.95, GFI=0.87, AGFI=0.85, NFI=0.90 and
IFI=0.95 (see table 2).
Table 2. Fit indices of the Sources of Middle School Mathematics Self-Efficacy Scale and acceptable fit
indices values
Original
Turkish
Acceptable
Fit Indices
Sources
Scale
Scale
Indices Values
2/sd
2.44
2.10
≤5
Bollen (1989), Sümer (2000)
IFI
0.95
≥0.90
RMSEA
0.04
0.07
≤0.08
Browne and Cudeck (1993), Hu and
Bentler (1999), Sümer, (2000), Byrne
SRMR
0.04
0.07
≤ 0.08
(1998)
GFI
0.87
≥0.85
Anderson and Gerbing (1984), Jöreskog
and Sörbom (1993), Cole (1987), Marsh,
AGFI
0.85
≥0.80
Balla and McDonald (1988)
NFI
0.90
≥0.90
Bentler (1992), Sümer (2000)
CFI
0.96
0.95
≥0.95
151
Education and Science 2014, Vol 39, No 176, 145-157
E. Yurt, A. M. Sünbül
Results of the Reliability Analysis
The reliability of the SMSMSEC’s each component was tested via Cronbach alpha.
Additionally, items’ discriminatory power was tested by (i) calculating corrected item-total
correlation; (ii) running a t-test comparing the total scores of top 27% and bottom 27 % of the
participants for each item. The results are displayed in table 3.
Vicarious
Experiences
Physiological
States
Social
Persuasions
Personal
Experiences
Tablo 3. Matematik Öz-yeterlik Kaynakları Ölçeği Faktörlerinin Düzeltilmiş Madde Toplam
Korelasyonları ve Üst %27, Alt %27 Puanları Arasındaki t testi Sonuçları
t
t
Factor
Item Corrected itemFactor
Item Corrected item(Top 27%(Top 27%Name
No total correlation
Name
No total correlation
bottom 27%)
bottom 27%)
i1
,72
15,92**
i13
0.72
21.61**
i2
,72
17,45**
i14
0.74
19.43**
i3
,35
4,79**
i15
0.73
19.51**
i4
,62
11,9**
i16
0.77
26.89**
i5
,53
13,7**
i17
0.77
28.56**
i6
,63
18,43**
i18
0.70
28.28**
i7
,62
15,34**
i19
-0.28
-4.29**
i8
,52
15,05**
i20
-0.25
-2.33*
i9
,56
13,45**
i21
-0.31
-3.1**
i10
,56
18,31**
i22
-0.27
-3.08**
i11
,65
16,25**
i23
-0.33
-4.03**
i12
,51
9,83**
i24
-0.35
-3.72**
*p<0.05, **p<0.001
The results indicate that the SMSMSEC’s items’ corrected item-total correlation ranged
between 0.77 and -0.25. T-tests comparing the total scores of top 27% and bottom 27 % of the
participants for each item indicate that there is a significant difference in scores from all items.
Cronbach alpha values for the Turkish and the original versions are presented in table 4. The
Cronbach alpha values of the Turkish version ranged between 0.80 and 0.94.
Table 4. Cronbach Alpha values of the Sources of Middle School Mathematics Self-Efficacy Scale
Alpha Value
Factor Name
Original
Turkish
Personal Experiences
0.88
0.87
Vicarious Experiences
0.84
0.80
Social Persuasions
0.88
0.93
Physiological States
0.87
0.94
152
Education and Science 2014, Vol 39, No 176, 145-157
E. Yurt, A. M. Sünbül
In the next step, mean values and standard deviation of the SMSMSEC’s four components
were calculated (see table 5). Correlation values between its components were also calculated. The
correlation values between the SMSMSEC’s four components ranged between -0.30 and 0.71. The
mean value for the Personal Experiences was 439.47 (Sd=127.34), for Vicarious Experiences, 419.87
(Sd=144.21); for the Social Persuasions, 361.33 (Sd=178.15); for the Emotional and Physiological
Situations; 201.30 (Sd=174.92).
Table 5. The Sources of Middle School Mathematics Self-Efficacy Scale‘s Correlation values between
its components and mean values and standard deviation of its four components
̅
𝐗
Değişkenler
Sd
V1
V2
V3
V4
V1 Personal Experiences
439.47
127.34
V2 Vicarious Experiences
419.87
144.21
0.48**
V3 Social Persuasions
361.33
178.15
0.71**
0.61**
V4 Physiological States
201.30
174.92
-0.54**
-0.30**
-0.49**
**p<0.01
Results of the Criterion Validity
To investigate whether the scale serves well, criterion validity was carried out. The scores
from SMSMSEC’s four components were correlated with scores from the mathematic anxiety scale
(Bindak, 2005) and self-efficacy scale (Karadeniz et al., 2008). The results are displayed in table 6. The
results indicated that SMSMSEC’s four components and the mathematic anxiety scale and self-efficacy
scale produced medium and high correlation values.
Table 6. Correlation between the mathematic anxiety scale and self-efficacy scale and The Sources of
Middle School Mathematics Self-Efficacy Scale‘s four components
Personal
Vicarious
Social
Physiological
Variables
Experiences
Experiences
Persuasions
States
Self-Efficacy
0.69**
0.51**
0.67**
-0.56**
Mathematic Anxiety
-0.74**
-0.49**
-0.61**
0.72**
**p<0.01
Discussionsn and Conclusion
This study investigated the validity and reliability of Usher and Pajares‘s Sources of Middle
School Mathematics Self-Efficacy Scale (2009) in Turkish context. The construct validity of the scale
was carried out via Exploratory factor analysis, and then via confirmatory factor analysis.
The results of the exploratory factor analysis produced results similar to the ones from the
original study. The factors revealed via exploratory factor analysis in this study were given the
Turkish translations of the original scale; Mastery Experiences; Vicarious Experiences; Social
Persuasions; Physiological States. The SMSMSEC’s four-component structure was investigated via
confirmatory factor analysis as it was done in the original study. When the fit indices (2/sd= 2.10;
RMSEA=0.07, SRMR=0.07, CFI=0.95, GFI=0.87, AGFI=0.85, NFI=0.90, IFI=0.95) were examined, the ratio
of 2/sd (2.10) and IFI (0.95) showed that the model fits the data very well (Bollen, 1989; Sümer, 2000).
The other fit indices were examined, the values of RMSEA (0.07) and SRMR (0.07) fit indices were very
close. RMSEA and SRME fit indices ideally need to be as close to zero as possible. However, a value of
0.05 and below for RMSEA and SRME fit indices is considered acceptable in the literature (Browne
and Cudeck, 1993). Additionally, other researchers (Browne and Cudeck, 1993; Hu and Bentler, 1999;
Sümer, 2000; Byrne, 1998) suggested a value of 0.10 and below for RMSEA and SRME fit indices are
acceptable too (Anderson and Gerbing, 1984; Marsh, Balla and McDonald, 1988; Cole, 1987). In short,
the model tested in this study fit the data at an acceptable level since it had many and complicated
factors; and the values for RMSEA, SRMR ranged between 0.05 and 0.08.
153
Education and Science 2014, Vol 39, No 176, 145-157
E. Yurt, A. M. Sünbül
The other fit indices, GFI and AGFI, are absolute fit indices. Some researchers (Hooper,
Coughlan and Mullen, 2008) suggested that a value 0.95 and below for GFI and AGFI indices show
perfect fit while some others (Baumgartner and Hombur, 1996; Hooper, Coughlan and Mullen, 2008)
suggested a value ranging between 0.90-0.95 show nearly perfect fit. On the other hand some
researchers (Anderson and Gerbing, 1984; Jöreskog and Sörbom, 1993; Cole, 1987; Marsh, Balla and
McDonald, 1988) stated that a value of 0.85 for GFI and a value of 0.80 and above for AGFI are
acceptable. The values of GFI (0.87) and AGFI (0.85) computed in this study are close to 0.90.
Therefore, the fit of the modal is acceptable.
The other fit indices are NFI and CFI fit indices. A value of 0.95 and above for them are
suggested to show perfect fit (Hu and Bentler, 1999), a value ranging between 0.90 and 0.95 shows
acceptable level of fit (Bentler, 1992; Sümer, 2000). The fit indices calculated in this study for NFI (0.90)
and CFI (0.95) are acceptable.
When the fit indices of this study and the original study are compared, the ratio of 2/sd and
CFI are very close to each other, and they both show acceptable fit. The values for RMSEA (0.04) and
SRMR (0.04) calculated in the original study show god fit while the values for RMSEA (0.07) and
SRMR (0.07) calculated in this study show acceptable fit.
The t-test results from the total scores of top 27% and bottom 27% of the participants showed
that there is a significant difference for mean values of all items. Therefore the items ‘discrimination
power was good. Cronbach alpha values calculated for the SMSMSEC’s four components ranged
between 0.80 and 0.94. These values are very close to the ones calculated in the original study.
Additionally, criterion validity was investigated by correlating scores from the SMSMSEC’s four
components, the mathematic anxiety scale and self-efficacy scale. The correlations ranged between
0.72 and -0.74. Therefore, usefulness of the scale is quite high.
In conclusion, Turkish version of the SMSMSEC consists of four factors as in the original
model; four -factor model fit the data from students participated in the research; internal consistency
of the factors is at an acceptable level and it serves its purpose well. When these results are taken into
consideration, the Turkish version of the SMSMSEC can be used to assess the Turkish middle school
students’ sources of mathematics self-efficacy.
154
Education and Science 2014, Vol 39, No 176, 145-157
E. Yurt, A. M. Sünbül
References
Akbaş, A., & Çelikkaleli, Ö. (2006). Sınıf öğretmeni adaylarının fen öğretimi öz yeterlik inançlarının
cinsiyet, öğrenim türü ve üniversitelerine göre incelenmesi. Mersin Üniversitesi Eğitim Fakültesi
Dergisi, 2(1), 98-110.
Anderson, J. C., & Gerbing, D. W. (1984). The effect of sampling error on convergence, improper
solutions, and goodness-of-fit indices for maximum likelihood confirmatory factor analysis.
Psychometrika, 49(2), 155-173.
Arslan, A. (2012). İlköğretim Öğrencilerinin Öz Yeterlik İnancı Kaynaklarının Öğrenme ve
Performansla İlgili Öz Yeterlik İnancını Yordama Gücü. Kuram ve Uygulamada Eğitim Bilimleri,
12(3), 1907-1920.
Arslan, A. (2013). Investigation of relationship between sources of selfefficacy beliefs of secondary
school students and some variables. Kuram ve Uygulamada Eğitim Bilimleri, 13(4), 1983-1993.
Azar, A. (2010). Ortaöğretim fen bilimleri ve matematik öğretmeni adaylarının öz yeterlilik inançları.
Zonguldak Karaelmas Üniversitesi Sosyal Bilimler Dergisi, 6(12), 235-252.
Bandura, A. (1997). Self-efficacy: The exercise of control. New York: Freeman.
Baumgartner, H., & Hombur, C. (1996). Applications of structural equation modeling in marketing
and consumer research: A review. International Journal of Research in Marketing, 13, 139-161.
Bentler, P. M. (1992). On the fit of models to covariances and methodology to the Bulletin. Psychological
Bulletin, 112, 400–404.
Bindak, R. (2005). İlköğretim Öğrencileri İçin Matematik Kaygı Ölçeği. Fırat Üniversitesi Fen ve
Mühendislik Bilimleri Dergisi, 17(2), 442-448.
Bollen, K. A. (1989). Structural equations with latent variables. New York: Wiley.
Bong, M., & Clark, R. E. (1999). Comparison between self-concept and self-efficacy in academic
motivation research. Educational psychologist, 34(3), 139-153.
Bong, M., & Skaalvik, E. M. (2003). Academic self-concept and self-efficacy: How different are they
really?. Educational psychology review, 15(1), 1-40.
Browne, M. W., & Cudeck, R. (1993). Alternative ways of assessing model fit. Sage Focus Editions, 154,
136-136.
Büyüköztürk, Ş. (2002). Faktör analizi: Temel kavramlar ve ölçek geliştirmede kullanımı. Kuram ve
Uygulamada Eğitim Yönetimi Dergisi, 8(4), 470-483.
Büyüköztürk, Ş. (2011). Veri analizi el kitabı (15. Baskı). Ankara: Pagem A Yayıncılık.
Büyüköztürk, Ş., Akgün, Ö. E., Özkahveci, Ö., & Demirel, F. (2004). Güdülenme ve öğrenme
stratejileri ölçeğinin Türkçe formunun geçerlik ve güvenirlik çalışması. Kuram ve Uygulamada
Eğitim Bilimleri, 4(2), 207-239.
Byrne, B. M. (1998). Structural Equation Modeling with LISREL, PRELIS and SIMPLIS: Basic Concepts,
Applications and Programming. Mahwah, New Jersey: Lawrence Erlbaum Associates.
Cole, D. A. (1987). Utility of confirmatory factor analysis in test validation research. Journal of
Consulting and Clinical Psychology, 55, 1019-1031.
Coşkun, M. K. (2010). Din kültürü ve ahlak bilgisi öğretmenlerinin öz yeterlik algılarının çeşitli
değişkenler açısından incelenmesi. Sosyal Bilimler Araştırmaları Dergisi, 1, 95-109.
Çalışkan, S., Selçuk, G. S., & Özcan, Ö. (2010). Fizik öğretmeni adaylarının öz yeterlik inançları:
cinsiyet, sınıf düzeyi ve akademik başarının etkileri. Kastamonu Eğitim Dergisi, 18(2), 449-466.
Çapri, B., & Çelikkaleli, Ö. (2008). Öğretmen adaylarının öğretmenliğe ilişkin tutum, ve meslek
yeterlik inançlarının cinsiyet, program ve fakültelerine göre incelenmesi. İnönü Üniversitesi Eğitim
Fakültesi Dergisi, 9(15), 33-53.
155
Education and Science 2014, Vol 39, No 176, 145-157
E. Yurt, A. M. Sünbül
Çetin, B. (2008). Marmara üniversitesi sınıf öğretmeni adaylarının bilgisayarla ilgili öz yeterlik
algılarının incelenmesi. Dicle Üniversitesi Ziya Gökalp Eğitim Fakültesi Dergisi, 11, 101-114.
Çetin, B. (2009). Yeni ilköğretim programı (2005) uygulamalarının ilköğretim 4. ve 5. sınıf
öğrencilerinin öz yeterliliklerine etkisi. Pamukkale Üniversitesi Eğitim Fakültesi Dergisi, 25(1), 130141.
Çokluk, Ö., Şekercioğlu, G., & Büyüköztürk, Ş. (2010). Sosyal bilimler için çok değişkenli istatistik: SPSS ve
LISREL uygulamaları. Ankara: Pegem Akademi.
Durdukoca, Ş. F. (2010). Sınıf öğretmeni adaylarının akademik öz yeterlik algılarının çeşitli
değişkenler açısından incelenmesi. Abant İzzet Baysal Üniversitesi Dergisi, 10(1), 69-77.
Ekici, G. (2006). Meslek lisesi öğretmenlerinin öğretmen öz yeterlik inançları üzerine bir araştırma.
Eurasian Journal of Educational Research, 24, 87-96.
Ferla, J., Valcke, M., & Cai, Y. (2009). Academic self-efficacy and academic self-concept: Reconsidering
structural relationships. Learning and Individual Differences, 19(4), 499-505.
Hooper, D., Coughlan, J., & Mullen, M.R. (2008). Structural equation modelling: Guidelines for
determining model fit. Journal of Business Research Methods, 6, 53–60.
Hu, L. T., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis:
Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary
Journal, 6(1), 1-55.
Işıksal, M., & Aşkar, P. (2003). İlköğretim öğrencileri için matematik ve bilgisayar öz-yeterlik algısı
ölçekleri. Hacettepe Üniversitesi Eğitim Fakültesi Dergisi, 25, 109-118.
İpek, C., & Acuner, H. Y. (2011). Sınıf öğretmeni adaylarının bilgisayar öz yeterlik inançları ve eğitim
teknolojilerine yönelik tutumları. Ahi Evran Üniversitesi Eğitim Fakültesi Dergisi, 12(2), 23-40.
Jöreskog, K. G., & Sörbom, D. (1993). LISREL 8: Structural equation modeling with the simplis command
language. Lincolnwood: Scientific Software International, Inc.
Karadeniz, Ş., Büyüköztürk, Ş., Akgün, Ö. E., Çakmak, E. K., & Demirel, F. (2008). The Turkish
adaptation study of motivated strategies for learning questionnaire (MSLQ) for 12–18 year old
children: Results of confirmatory factor analysis. Turkish Online Journal of Educational Technology,
7(4), 108-117.
Kline, R. B. (2011), Principles and Practice of Structural Equation Modeling (3nd Edition ed.). New York:
The GuilfordPress.
Maden, S. (2010). Türkçe öğretmenlerinin drama yöntemini kullanmaya yönelik öz yeterlikleri.
Mustafa Kemal Üniversitesi Sosyal Bilimler Enstitüsü Dergisi, 7(14), 259-274.
Marsh, H.W., Balla, J. R., & McDonald, R. P. (1988). Goodness-of-fit indexes in confirmatory factor
analysis: The effect of sample size. Psychological Bulletin, 103, 391-410.
Morgil, İ., Seçken, N., & Yücel, A. S. (2004). Kimya öğretmen adaylarının öz yeterlik inançlarının bazı
değişkenler açısından incelenmesi. Balıkesir Üniversitesi Fen Bilimleri Enstitüsü Dergisi, 6(1), 62-72.
Schunk, D. H. (2011). Learning theories: An educational perspective. Pearson Education, Inc.
Schunk, D. H., & Pajares, F. (2005). Competence beliefs in academic functioning. In A. J. Elliot & C.
Dweck (Eds.), Handbook of competence and motivation (pp. 85–104). New York: Guilford Press.
Schunk, D. H., & Pajares, F. (2009). Self-efficacy theory. In K. R. Wentzel & A. Wigfi eld (Eds.),
Handbook of motivation at school (pp. 35–53). New York: Routledge.
Schunk, D. H., & Zimmerman, B. J. (Eds.). (1998). Self-regulated learning: From teaching to self-reflective
practice. Guilford Press.
Senemoglu, N. (2007). Gelişim ögrenme ve ögretim. Ankara: Gönul Yayincilik.
Sümer, N. (2000). Yapısal eşitlik modelleri: Temel kavramlar ve örnek uygulamalar. Türk Psikoloji
Yazıları, 3(6), 49-74.
Stevens, J. P. (2009). Applied Multivariate Statistics for the Social Sciences. New York: Taylor & Francis.
156
Education and Science 2014, Vol 39, No 176, 145-157
E. Yurt, A. M. Sünbül
Şahin, R. (2013). Öğrenme Psikolojisi. M. Baloğlu (Ed.), Sosyal bilişsel kuram içinde (s.111-140). Ankara:
Nobel Akademik Yayıncılık.
Tabachnick, B. G., & Fidell, L. S. (2007). Using multivariate statistics (5th ed.) Boston: Allyn and Bacon.
Tavşancıl, E. (2010). Tutumların ölçülmesi ve SPSS ile veri analizi (4. Basım). Ankara: Nobel Yayın
Dağıtım.
Terzi, M., & Mirasyedioğlu, Ş. (2009). İlköğretim matematik öğretmen adaylarının matematiğe yönelik
öz yeterlik algılarının bazı değişkenler açısından incelenmesi. TUBAV Bilim Dergisi, 2(2), 257-265.
Usher, E. L. (2009). Sources of middle school student's self-efficacy in mathematics a qualitative
investigation. American Educational Research Journal, 46(1), 275–314.
Usher, E. L., & Pajares, F. (2006). Inviting Confidence in School: Invitations as a Critical Source of the
Academic Self-Efficacy Beliefs of Entering Middle School Students. Journal of Invitational Theory &
Practice, 12, 7-16.
Usher, E. L., & Pajares, F. (2009). Sources of self-efficacy in mathematics: A validation study.
Contemporary educational psychology, 34(1), 89-101.
Yaman, S., Koray, C. Ö., & Altunçekiç, A. (2004). Fen bilgisi öğretmen adaylarının öz yeterlik inanç
düzeylerinin incelenmesi üzerine bir araştırma. Türk Eğitim Bilimleri Dergisi, 2(3), 355-366.
Yılmaz, G., Yılmaz, B., & Türk, N. (2010). Beden eğitimi ve spor öğretmenlerinin mesleklerine ilişkin
öz yeterlik düzeylerinin incelenmesi. Selçuk Üniversitesi Beden Eğitimi ve Spor Bilim Dergisi, 12(2),
85-90.
Yılmaz, M., & Çimen, O. (2008). Biyoloji eğitimi tezsiz yüksek lisans öğrencilerinin biyoloji öğretimi öz
yeterlik inanç düzeyleri. Yüzüncü Yıl Üniversitesi Eğitim Fakültesi Dergisi, 5(1), 20-29.
Zimmerman, B. J., & Kitsantas, A. (1999). Acquiring writing revision skill: Shifting from process to
outcome self-regulatory goals. Journal of educational Psychology, 91(2), 241.
157