Psychometric Properties of Adaptation of the Social Efficacy and Outcome Expectations Scale to Turkish

Abstract: The aim of this study was to examine the psychometric properties of the Social Efficacy and Outcome Expectations Scale (SEOES) on Turkish. The sample group included two groups of university students (ns= 440, 359). The validity of the scale was assessed using exploratory factor analysis, confirmatory factor analysis and concurrent validity, and the reliability was assessed using Cronbach Alpha’s internal consistency coefficient, test-retest method and item discrimination. The findings obtained from the exploratory factor analysis showed that 51.2 percent of total variance of the scale was explained and the scale consisted of two subdimensions like the original one. Confirmatory factor analysis indicated an acceptable fit to the data. The concurrent validity of the SEOES was respectively studied using (.65) Scale of Perceived Social Self-Efficacy; (.29) Satisfaction With Life Scale; and (-.36) UCLA Loneliness Scale. As for the scale’s reliability, the internal consistency was determined to be .91; .92 for Social Efficacy Scale; and .81 for Outcome Expectations Scale. The results of test-retest result (.90) and the findings of item analysis showed that the items in Turkish version of the scales were compatible with the original one. The present results provide evidence supporting the validity and reliability Turkish version of the SEOES.


Introduction
An important field in psychology is the study of self-efficacy (DeWitz and Walsh, 2002).In self-efficacy theory, it is of great importance for individuals to feel themselves efficient while engaging in social life and maintaining a healthy relationship with others (Bandura 1977(Bandura , 1997)).Self-efficacy means one's belief in his/her competence to cope with challenging situations in his/her life (Bandura 1977).In addition, "Perceived self-efficacy is defined as people's beliefs about their capabilities to produce designated levels of performance that exercise influence over events affecting their lives" (Bandura 1994, p.2).In other words, one's belief in his/her competence to overcome challenges he/she faces is seen as self-efficacy (Miller 2008).
Self-efficacy of the individual is influenced by four sources; performance accomplishments, vicarious experience, verbal persuasion and emotional arousal (Bandura 1977).It is known that high self-efficacy level is a positive factor for individuals to be successful in their school and working lives, whereas low self-efficacy discourages individuals in their new attempts if they fail in their previous attempts (Korkmaz 2005).
According to Social Cognitive Theory, both self-efficacy beliefs and outcome expectations are related to cognitive judgments that have effects on the behaviors of individuals, and they operate dependently (Bandura, 2001).In other words, social self-efficacy is the confidence the individuals feel when fulfilling tasks about the social relationships to initiate and maintain interpersonal relations (Smith and Betz 2000)."The influence of social self-efficacy on various structures is studied to examine social relational tasks" (Wright et al. 2013, p.218).Social efficacy is also known as social courage, which includes the skills to belong to a group, attend activities and assist the other in group (Conolly 1989).The expectations about the results caused by social behaviors are social outcome expectations (Wright et al. 2013) Outcome expectation is the intuition by the individual about the consequences he/she will face about some behaviors (Bandura 1977).
When the related literature is examined, although self-efficacy and social-efficacy are widely investigated, it is seen that there is not any research about the outcome expectations.The findings show a positive and statistically significant relation between social self-efficacy and self-esteem (Smith and Betz 2000), life satisfaction (Wright and Perrona 2010), self-efficacy, coping, empathy and extraversion (Di Giunta et al. 2010).However, a negative relation was found between social self-efficacy and social anxiety, and between shyness (Smith and Betz 2000;Palanci 2004) and loneliness (Hermann and Betz 2006).
As the social relationships established by an individual are considered as a reference for his/her later experiences, researchers have developed the Social Efficacy and Outcome Expectations Scale (SEOES) to measure social efficacy and outcome expectations together.The SEOES is consists of two sub-scales; Social Efficacy Scale (SES) and Outcome Expectations Scale (OES) (Wright et al. 2013).
The Social Efficacy Scale of consists 12 items and Outcome Expectations Scale consists of 6 items.SEOES questionnaires include a 5-point scale ranging from strongly disagree (1) to strongly agree (5).The scale does not include reversescored items.In the scale, total score is taken for each dimension.The highest and lowest scores to be respectively obtained from the scale are as follows: the SEOES (18-90), the SES (12-60), and the OES (6-30).Wright and his colleagues (2013) found that Cronbach alpha coefficient was found as the .96for SEOES, the .96for SES, and .90 for OES.Their study showed that there were positive relations between SEOES and Scale of Perceived Social Self-Efficacy (PSSE).In addition to this, 20-item in SEOES and two-factor confirmatory analysis indicated an acceptable fit to data: χ 2 =626.719,df =169; χ 2 /df =3.71, RMSEA =.066, SRMR =.05, CFI =.95, and NNFI =.95.However, at the end of the confirmatory analysis conducted for the SEOES, the sixth and ninth items were taken out from the scale as their error variances were found to be too high.Therefore, at the end of the confirmatory factor analysis consisting of 18 items, it was found that the error variances as χ 2 =448.454,df = 134; χ 2 /df =3.35, RMSEA =.059, SRMR =.04, CFI =.97, and NNFI =.96.In conclusion, the scale was completed with 18 items.In the final form of the scale, when the sixth and ninth items were taken out, the SES had 12 items and the OES had six items (Wright et al. 2013).
This adaptation study was carried out with 20 items including the sixth and ninth items mentioned above.We decided to include the sixth and ninth items based on two grounds; first, the confirmatory analysis results of the 20-item form of the SEOES proved acceptable and second, these items could be functional in Turkish culture.
In this study, construct validity and concurrent validity were investigated so that the scale could be accepted as valid.The exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) were applied for determining construct validity of the scale.The EFA was conducted to reveal the factor structure of the scale.The CFA was used to understand whether the earlier construct was confirmed or not (Buyukozturk et al. 2004).

The Sample Group
The Turkish version of the scale (SEOES) was applied to university students as the original study (Wright et al. 2013) was conducted with university students.This study was conducted with two separate groups of university students.The characteristics of these two sample groups are presented below.The sample groups in this study were constituted by undergraduate students from a middle class university in Turkey.

Sample Group 1
The first sample group of this study was consisted of 440 university students.284 (64.5%) of these students were female and 156 (35.5%) were male.The year-of-study distribution was: 84 (19.1%) first year students, 118 (26.8%) second year students, 110 (25.0%) third year students and 128 (29.1%) fourth year students.The ages of students ranged from 19 to 31 and the average age was 21.60.The participants of this sample group were studying in Faculty of Economics and Administrative Sciences, Faculty of Education and Faculty of Science and Letters of the university.An exploratory factor analysis work was conducted using the data obtained from the first sample group of this study.
The test-retest process of this study was completed with 81 volunteer university students from first sample group 1.In test-retest process; 49 (60.5%) of the participants were female and 32 (39.5%) were male.The year-of-study distribution was: 22 (27.2%)first year students, 28 (34.6%)second year students, 20 (24.7%) third year students and 11 (13.6%)fourth year students.Additionally, the ages of participants ranged between 19 and 25 and the average age was 21.

Sample Group 2
The second sample group of this study included 359 volunteer university students.191 (53.2%) of these students were female and 168 (46.8%) were male.The year-of-study distribution was: 58 (16.2%) first year students, 93 (25.9%) second year students, 92 (25.6%) third year students and 116 (32.3%) fourth year students.The ages of students ranged from 19 to 29 and the average age was 21.65.The participants were selected using convenience sample method.The participants of the second sample group were studying in Faculty of Economics and Administrative Sciences, Faculty of Education and Faculty of Science and Letters of the university.A confirmatory factor analysis and concurrent validity were conducted using the data obtained from the second sample group of this research.

Materials
The data were collected using Social Efficacy and Social Outcome Expectations Scale (SEOES), Scale of Perceived Social Self-Efficacy (PSSE), Satisfaction With Life Scale (SWLS), UCLA Loneliness Scale (ULS-8) and Personal Information Form.

Scale of Perceived Social Self-Efficacy (PSSE)
Scale of Perceived Social Self-Efficacy (PSSE) was developed by Smith and Betz (2000).Palanci (2004) adapted this scale to Turkish.The scale consists of 25 items and the total score is taken from it.A high score in the scale means that this social self-efficacy is high, whereas a low score means the social self-efficacy is low.PSSE questionnaires include a 5-point scale ranging from no confidence at all (1) to complete confidence (5).It was expressed that the PSSE original scale explained 40.7% of the total variance (Smith and Betz 2000) and the adapted version explained 44.2% of it (Palanci 2004).Cronbach Alpha reliability coefficient of the original scale was .94 and test-retest reliability coefficient was .82,while the former was .89 and latter was .68 in Turkish version of the scale.In this study, Cronbach Alpha reliability coefficient was found to be .93.

Satisfaction with Life Scale (SWLS)
Satisfaction with Life Scale was developed by Diener et al. (1985).It was adapted to Turkish by Yetim (1993).This scale does not include reverse-scored items.The scale consists of five items.SWLS questionnaires include a 7-point scale ranging from strongly disagree (1) to strongly agree (7).Cronbach Alpha reliability coefficient of the original scale was .87 and test-retest reliability coefficient was .82(Diener et al. 1985).Cronbach Alpha reliability coefficient of the Turkish version was found as .86 by Yetim (1993) and .74by Morsumbul (2004).Test-retest reliability of the Turkish version was .73(Yetim1993).In this study, Cronbach Alpha reliability coefficient was found to be .83.

UCLA Loneliness Scale (ULS-8)
UCLA Loneliness Scale (ULS-8) was developed by Hays and DiMatteo (1987).This scale was adapted to Turkish by Dogan et al. (2011).The scale consists of eight items, and the third and sixth items are reverse-scored.Total score is taken from the scale.A high score in this scale means loneliness level is high, whereas a low score means a low level of loneliness.ULS-8 questionnaires include a 4-point scale ranging from never (1) to always (4).The 8-item loneliness scale has one dimension.Cronbach Alpha reliability coefficient of the original scale was .84.Additionally, it was determined that there was a positive correlation value between social and emotional loneliness, whereas there was a negative value regarding the perceived social support.In this study, Cronbach Alpha reliability coefficient was found to be .79.

Personal Information Form
In order to collect demographic information of the individuals involved in study, an information collection form including information about age, sex and grades of the participant students was employed in this study.

Procedure
In data collection stage of this study, the assessment tools were prepared as a leaflet and distributed to volunteer students in a classroom environment.Before each assessment application, the researchers introduced themselves and told about the importance and purpose of the study.Additionally, the researchers told the participants that there will be no individual evaluation, their identity information are not required and the results will be used only for scientific purposes.
While collecting the data, counterbalancing method was adopted in order to prevent possible ordinal error and to boost the validity of the work.Four different leaflet were prepared by changing the order of the scales included.Thus, it was aimed to remove the effects of primacy and recency through counterbalancing.

Data Analysis
In this study, Exploratory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA) were employed for determining the construct validity of the Turkish versions of SEOES.Concurrent validity of the scale was conducted.The reliability of the SEOES was computed by using Cronbach Alpha internal consistency coefficient and test-retest methods.Additionally, item discrimination of the scale was assessed by computing item total correlations.SPSS 15.0 and LISREL 8.7 software were used for determining the validity and reliability values of SEOES.

Translation Process
Necessary permissions for the adaptation process were taken from the first author of the scale.The SEOES was translated into Turkish by four experts working in psychological counseling and guidance areas.The Turkish translations of scale were collected on one form based on the opinions of the two experts working in psychological counseling and guidance areas.This scale form was back-translated by two experts working in psychological counseling and guidance areas.Additionally, 15 experts with advanced English checked the original scale and Turkish translation version.The final version of the scale was prepared in accordance with the opinions of these experts.

Results
In this section, the findings obtained for validity and reliability of this study as a part of the adaptation of Social Efficacy and Social Outcome Expectations Scale to Turkish were presented.

Exploratory Factor Analysis (EFA) Results
In order to determine suitability of the SEOES for construct validity, Kaiser-Meyer-Olkin (KMO =.92) coefficient was calculated and Bartlett Test (χ 2 =3997.933,df =171 (p< .000)wasapplied.The EFA was also conducted to reveal factor structure of this study.
First, the EFA was conducted without any factorial limitations and a two-factor structure was obtained.Later, the factor structure of the scale was limited to two in order to obtain the same structure with the original one and Promax oblique rotation method was adopted.At the end of the analysis, the sixth item was taken out as its factor loading was less than .40 and the analysis was repeated.As seen in Table 1, the factor loadings of SES varied between .55 and .81,whereas those of the OES changed between .63 and .81.Eigenvalue of the SES was 7.54 and the total variance it explained was 39.7%, while the eigenvalue of the OES was 2.36 and the total variance it explained is 12.4%.The SEOES explained 52.1% of the total variance.Similar results were obtained at the end of exploratory factor analysis conducted using the data obtained the samples.

Confirmatory Factor Analysis (CFA)
A confirmatory factor analysis was carried out in order to confirm the construct validity obtained by exploratory factor analysis.In accordance with this aim, the analysis was carried out by applying the findings of this study to the confirmatory factor analysis and by considering suitability of the construct, fit statistics and modification suggestions.At the end of the analysis, the confirmatory factor analysis was repeated after making modifications between the items of 1-2 and 18-19.In confirmatory factor analysis, fit indices of a two-dimensional model were studied and Chi-square test value (χ 2 =411.81,df=149, χ 2 /df=2.76,p=0.00) was found to be statistically significant.Finding the Chi-square value statistically significant is not a desired situation and a high value has nothing to do with sample size.This value is found high in the studies where sample is large (Cokluk et al. 2012).If the Chi-square value is smaller than 5, it is assessed as a congruence in middle level.However, if it is smaller than 3, it is seen as an excellent congruence (Sumer 2000).In this study, it was obtained at excellent level.Other fit indices are respectively found as RMSEA=.070,GFI=.89,AGFI=.86,NFI=.96,NNFI=.97,CFI=.98,IFI=.98,RMR=.02 and SRMR=,05.If RMSEA value is .08 and smaller, it is assessed as good fit (Sumer 2000;Hooper et al.2008).The values of RMR and SRMR are .05and smaller, which is seen as an excellent congruence (Brown 2006).In fit indices of confirmatory factor analysis, the proximity of values such as GFI, AGFI, CFI and NFI to .90 is considered as an indicator of excellent fit (Hair et al. 1998).

Concurrent Validity of SEOES
Perceived Social Self-Efficacy Scale (PSSE), Satisfaction with Life Scale (SWLS) and UCLA Loneliness Scale (ULS-8) were used with regard to concurrent validity of the SEOES.In the analysis carried out as part of this validity, a positive statistically significant relation was found as .65 between the PSSE and the SEOES, .64 for the SES and .38 for the OES.A statistically significant relation was found as .29 between Satisfaction with Life Scale (SWLS) and SEOES total score, as .30for SES, .16 for OES, and -.37 for the PSSE.A statistically significant relation was found as -.36 between UCLA Loneliness Scale (ULS-8) and SEOES, -.37 for SES, -.17 for OES, -.43 for PSSE and -.28 for SWLS.The results obtained as part of the concurrent validity are shown in Table 2.

Reliability of SEOES
Reliability of the Social Efficacy and Social Outcome Expectations Scale was examined by computing Cronbach Alpha internal consistency coefficient and test-retest reliability coefficient.The Cronbach Alpha internal consistency coefficients, which are computed to ensure the reliability of the scale, were found .91 for the whole SEOES, .92 for the SES and .81for the OES.This process was applied again two weeks later for determining test-retest reliability and the reliability coefficient between the two applications was found as .90(p<.001) for SEOES.

Item Analysis
At the end of the item analysis of the SEOES, the mean changed between 3.99 and 4.38 and the standard deviation values varied between .57and .76.Additionally, corrected item-total correlations of the SEOES varied between .39 and .73.The obtained findings are shown in Table 3. Note.N = 440.SEOES = Social Efficacy and Outcome Expectation Scale; SD = standard deviation; ITC = item-total correlation.Scale range for items: 1 = strongly disagree, 2 = disagree, 3 = neither agree nor disagree, 4 = agree, and 5 = strongly agree.

Discussion and Conclusion
In this study, it was aimed to investigate the adaptation of the SEOES on Turkish university students, and to evaluate dimensionality, validity, and reliability of this translated version.
The KMO value of the SEOES Exploratory Factor Analysis (EFA) was found at an acceptable level (.92).The high value of KMO means that a variance in the scale is explained by other variances at a good level.If KMO value is .90or higher, it is interpreted as excellent (Sencan 2005).In the EFA, as the factor loading was under .40,the sixth item was taken out from the scale and the analysis was carried out with 19 items.However, the ninth item was included in the scale as its factor loading was .58.After conducting the EFA with 19 items, the factor loading values of the scale items varied between .55 and .81.In the original study of the scale (Wright et al. 2013), the factor loadings varied between .55 and .97 and the lowest factor loading obtained in studies was the same.Additionally, the SES explained 39.7% of total variance, the OES explained 12.4% of total variance and the SEOES explained 52.1% of the total variance of the scale.At the end of the analysis carried out with 20-item scale in the original study, the explained variances were as follows: 56.6% for SES, 11.70% for OES and 68.35%for SEOES.Close values were obtained although the total variance the scale explained in the adapted version was a bit lower than the values obtained in the original study (Wright et al. 2013).These findings showed that the SEOES measured the qualification to a great extent and in fact,this was desired in the Turkish version.As a result of the CFA conducted for testing the construct validity of the SEOES, it was observed that fit indices were found at a good level, the fit indices were close to the indices obtained in the original study (Wright et al. 2013), and 19-item form of the scale consisting of two sub-title dimensions was confirmed.At the end of analysis, it is determined that the scale (SEOSES) is valid.
As part of the concurrent validity of Turkish version of the SEOES, high and middle level correlation values were obtained using Perceived Social Self-Efficacy Scale (PSSE) (Smith and Betz 2000).In the original article, it was .74 between the PSSE and the SEOES, .77for the SES and .45for the OES (Wright et al. 2013).When the findings of this study are compared to the findings obtained in the original study, it is seen that the obtained results are similar.In addition to these, at the end of the concurrent validity conducted for adaptation of the SEOES to Turkish, middle and low level correlation values were obtained using Satisfaction With Life Scale and UCLA Loneliness Scale.The literature review shows that, a positive relation is seen between life satisfaction and social efficacy (Wright and Perrone 2010).
Based on these findings, it can be said that the higher the social efficacy level is, the more satisfaction with life is ensured.On the other hand, a negative relation is seen between loneliness and social efficacy (Hermann and Betz 2006).In the light of these findings, it can be said that the more the social efficacy level increases, the more the loneliness level decreases and thus lonely individuals have low social efficacy.
Cronbach Alpha internal consistency coefficient, which was calculated to determine the reliability of the Turkish version of the SEOES, was found .91 for SEOES, .92 for SES and .81for OES.Cronbach Alpha internal consistency coefficient was found .96for SEOES, .96for SES and .90 for OES in the original study (Wright et al. 2013).It can be said that the results of the reliability coefficients obtained in both works are close to each other.
The stability coefficient calculated for test-retest reliability of the scale was found as .90.It is seen that the findings obtained in the adapted version of the scale are similar to those obtained in the original study (Wright et al. 2013).At the end of the item analysis, it was determined that the item-total correlations varied between .39 and .73. in the original study (Wright et al. 2013), changed between .56 and .84.This asserts that the values obtained in the adapted scale were close to the values obtained in the original one.Additionally, the item-total correlation value was .25 and higher and it was seen as an adequate value (Ozdamar 1997).In this study, the obtained values met this requirement.
Although the SEOES gives good psychometric characteristics, the results of this study must be evaluated after taking some limitations into account.First, the participants in this study were mostly undergraduate students from a middle class university in Turkey.Thus, the findings may not be applied to a larger group of students from other educational environments.Second, the instrument was based on self-report questionnaires prepared for university students.
In conclusion, this adapted scale can be used by field experts and researchers in studies to be carried out about university students' social efficacy and outcome expectations.In addition to this, as a result of adaptation of the SEOES into different cultural structures, intercultural studies can also be conducted regarding this subject.

Figure 1 .
Figure 1.Path Diagram and Factor Loadings about the SEOES A path diagram showing standardized coefficients is shown in Figure-1.At the end of this analysis, the rates of tacit variances to explain observed variance are in the range of .49-.80 for the SES and .64-.78 for the OES.On the other hand, all items resulted in statistically significant "t" value in explaining tacit variances.

Table 3 .
Mean,Standard Deviation and Corrected Item-total Correlations Values The results of this study provide evidence that supports the validity and reliability of the scale.I am confident in my skills to share…