The Regulatory Emotional Self-Efficacy Scale : Issues of Reliability and Validity Within a Turkish Sample Group

The purpose of this study was to psychometrically evaluate the Turkish version of the Regulatory Emotional Self-efficacy Scale (RESE). The RESE, the Emotional Self-efficacy Scale, the Self-liking/Self-competence Scale, and the Oxford Happiness Questionnaire were applied to 303 university students in total, 180 were women (59.4%) and 123 were men (40.6%). According to results of confirmatory factor analysis applied in the study are founded enough conformity between the priori hypothesis model and the data. In addition, the metric invariance model shows that there were no gender differences on this confirmatory model. Internal consistency coefficients were all above the acceptable for the RESE’s sub-scale and total. Moreover, positive correlations were found between regulatory emotional self-efficacy dimensions and emotional self-efficacy, self-esteem, and happiness. According to these research findings, the RESE is a valid and reliable instrument for measuring regulatory self-efficacy in Turkish.


Introduction
Effective communication established with the environment affects the personality structure of the individual (Fossum & Barrett, 2000;Stemmler & Wacker, 2010).In this context, Bandura (1977) in his social learning theory states that self-efficacy beliefs are dynamic structures which the individual learns via experiences to the extent of his/her capacity.According to this, individual who can arrange and change the environmental situations is agent due to being able to shape his/her life.However, self-efficacy does not qualify the behaviors that the individual demonstrate but qualifies the perception of the individual on his/her belief in demonstrating these behaviors in a certain area depending on the experiences (Ajzen, 2002;Bandura, 1994;Bandura, 1997;Bandura, 2000;Zimmerman, 2000).So, self-efficacy beliefs of the individual affect his/her social and emotional functioning (Totan, 2014a).
The social learning theory is based on social-cognitive theory.Another important element of this theory is the concept of self-efficacy that is related to the beliefs that affect the results of the behaviors appearing in the individual (Bandura, 1997;Bandura & Locke, 2003;Özer, Totan, & Atik, 2011), outcome expectations, self-concept, Psychological Thought psyct.psychopen.eu| 2193-7281 and perceived control (Zimmerman, 2000).According to Bandura (1997), self-efficacy which is described as the beliefs of individuals in being successful in certain fields, and it significantly determines their job performance.
The understanding of a certain field in question has caused the occurrence of several measurement scales.Bandura (2006) argued that rather than the holistic and unique measurement scales oriented to evaluate self-efficacy, it is necessary to produce measurement scales that separately evaluate each self-efficacy field.The main reason of this is the difficulty of analyzing self-efficacy with all its dimensions and also the inconsistencies of these measurement scales in gaining the sufficient information in situational evaluations.For this reason, the measurement scales related to self-efficacy should be arranged according to the scope of the related field.
Factorial studies on self-efficacy revealed that specific self-efficacy areas existed.Emotional self-efficacy is nothing but a hierarchical process in which the individuals perceive, understand, describe emotions and use them as the regulator of their thoughts (Kirk, Schutte, & Hine, 2008).It is described in another study as the perceived capacity oriented to deal with the negative affect (Muris, 2001).Petrides, Furnham, and Mavroveli (2007) describe trait emotional self-efficacy as self-perceptions in which the individual reports himself/herself oriented to his/her emotional wishes in the context of emotional intelligence.In another point of view, Choi, Kluemper, and Sauley (2013) discuss the emotional self-efficacy as a dynamic structure rather than in the context of the stability of the emotional intelligence.They describe the concept of emotional self-efficacy as the individual's perception of his/her success of controlling or transforming emotional life and they also structured the measurement questions under the roof of "can do".Managing the negative affect and expressing the positive ones forms the perception on the regulatory factors of emotional self-efficacy within the affective field (Caprara & Gerbino, 2001as cited in Caprara et al., 2008).The reason why there are so many explanations that evaluate the emotional functioning in the theory of self-efficacy could be stemming from the researchers' different points of view on the transformation of emotional intelligence to self-efficacy.Caprara et al. (2008) state that individuals managing their emotional experiences in their daily lives, notably differ from each other because a) individuals' skills on managing the daily lives and b) the perception of sufficiency oriented to regulate the emotions differ from each other.In fact, the individual can do anything that s/he doesn't believe to be successful when s/he faces compelling situations.Within this concept, the researchers describe the dimensions of self-efficacy with the view of emotional regulation in managing negative affect oriented to the effects of troubled and tragic experiences and self-efficacy in managing positive affect oriented to experiences that include pride, enthusiasm and cheer.
In addition, regulatory self-efficacy which is established by Caprara et al. (2008) takes its foundations from previous studies (Bandura, Caprara, Barbaranelli, Gerbino, & Pastorelli, 2003;Caprara, Caprara, & Steca, 2003).However, Caprara et al. (2008) reach the ultimate result of the study via the studies they conducted on Italian, American and Bolivian participants.According to the result, they determined that the explained variance was above 30% in Italian group, above 29% in American group and above 30% in Bolivian group.It was concluded that the Cronbach alpha coefficients were above .64 in all these groups.The researchers concluded the existence of metric invariance models (both for three countries and gender) in the results of multi group confirmatory factor analysis.When the researchers examined the relationship between the similar and opposite scales, they also stated that there were statistically significant relationships between the scales and the variables oriented to adjustment as self-esteem, positive affect and pro-social behavior in a positive way.There were also statistically significant relationships between the scales and the variables oriented to maladjustment as negative affect, shyness, irritability, aggression and anxiety/depression in a negative way.
Caprara and his colleagues (2008) describe self-efficacy in managing negative affect as regulating negative emotion due to difficulties and challenges in life, and the belief in overcoming such emotions as anger, rage, deformity and discouraging feelings.The researchers (Bandura, 1993;Bandura, 1997;Totan, 2014a;Vignoles, Regalia, Manzi, Golledge, & Scabini, 2006) describe self-efficacy in managing positive affect as the competence belief in showing or experiencing such affects as amusement, enthusiasm and pride as a result of successful and satisfying events.The examination of the structure of regulatory self-efficacy suggested its validity and reliability in a number of cultures like USA, Bolivia, and Italy (Caprara et al., 2008;Zhao et al., 2013).Therefore, the aim of this study is to translate the RESE into Turkish and to conduct validity and reliability studies on Turkish youth in the context of classical test theory.

Method Participants
The participants of this study included university students who were receiving education in Adnan Menderes University in Faculty of Education in the academic period of spring semester 2013.Participants were selected via convenience sampling.Total number of participants was 303 (N = 303), of which 180 (59.4%) were women and 123 (40.6%) were men.Age range was between 18 to 25 (M = 20.88,SD = 1.63).In this group of participants, n = 93 (30.9%) were freshmen, n = 87 (28.7%) were sophomores, n = 66 (21.8%) were juniors, and n = 57 (18.8%) were seniors.

Measures
Regulatory Emotional Self-Efficacy Scale (RESE) -RESE devised by Caprara et al. (2008) examines the individual's competence belief oriented to show or experience the affects occurring in the affective field due to positive or negative events.It is composed of 12 items and in the form of five point scale.This scale is composed of two sub-dimensions: 1) perceived self-efficacy in expressing positive affect and 2) perceived self-efficacy in expressing negative affect.Also, negative affect has two sub-factors: 1) perceived self-efficacy in managing anger/irritation and 2) perceived self-efficacy in managing despondency.Caprara and colleagues (2008) reported that the RESE had strong validity and reliability evidence in USA, Italy and Bolivia cultures.Kirk, Schutte, and Hine (2008) and it has four dimensions including regulating emotions, using thoughts as supporter, understanding and perceiving the thoughts.This scale contains 32 items and the answers are given in the form of five point scale [not sure, hardly sure, a little sure, mostly sure, exactly sure].The researchers determined that four dimensions explained 44% of the total explained variance during the studies to develop the scale.They concluded that the internal consistency of the scale was at the level of .96.Totan, İkiz, and Karaca (2010) who adapted the scale to Turkish culture, stated that the scale maintained its four-factored structure in Turkish language and reported that the Cronbach alpha coefficient was above .90for the sub-dimensions and the total scale.Doğan, Totan, and Sapmaz (2013) reported ESES' internal consistency as .91.In addition, Cronbach alpha was calculated as .92for the total score on ESES in this study.

Emotional Self-Efficacy Scale (ESES) -ESES was developed by
Self-Liking/Self-Competence Scale -This scale was developed by Tafarodi and Swann (2001) and adapted to Turkish language by Doğan (2011).This scale has two dimensions, 16 items, and the answers are given in the form of five point scale.The total score on the scale (the sum of both dimensions) is evaluated as self-esteem.Tafarodi and Swann (2001) concluded that the scale confirmed its two-factored structure as a result of the confirmatory factor analysis (χ 2 = 6.56, df = 103, CFI = .92,RMSEA = .006)and they also determined that the internal consistency coefficients for both the sub-dimensions and the total was between .82 to .90.Doğan (2011) determined that the scale is reliable and valid in Turkish language.The researcher concluded as a result of the confirmatory factor analysis that the two-dimension structure of the scale was confirmed (χ 2 = 258.40,df = 98, CFI = .97,RMSEA = .049)and he determined the internal consistency coefficients as between .72 to .83 for the sub-dimensions and the total score on the scale.Cronbach alpha coefficient for total score was found to be at .89.Also, Totan (2014b) reported .89as Cronbach alpha coefficient and .91 as McDonald omega coefficient for total score.
Oxford Happiness Questionnaire (OXQ) -The OXQ was developed by Hills and Argyle (2002) and adapted to Turkish by Doğan and Akıncı Çötok (2011).This scale is composed of 7 items and the answeres are given in the form of five point scale.Hills and Argyle (2002) reported that the internal consistency value of the scale was as much as .91 to .92, and the explained variance in the factor analysis of the structure validity was above 60%.

Procedure
In this study, first of all permission was taken from Dr. Gian Vittorio Caprara to adapt the RESE to Turkish.After seeking the necessary permissions and reaching the original form of the scale, this was translated into Turkish language.Translations were conducted by three people who had Ph.D. degree in the Department of Psychological counseling and guidance at Adnan Menderes University.The researcher turned these different translations' forms to final form.A Turkish language expert examined this final form in terms of Turkish language.Then the author applied the resulting form to 30 third grade psychological counseling and guidance students, and examined the comprehensibility of all items.Turkish form was applied to university students who received education in Adnan Menderes University in Aydın city of Turkey in the spring semester of 2013 academic year.Also, the validity and reliability studies of the RESE were conducted.Concerning the structure validity, explanatory and confirmatory factor analyses were used, while in the internal consistency calculations Cronbach alpha coefficient was determined.
Moreover, these studies have examined the relationship between regulatory emotional self-efficacy dimensions and emotional self-efficacy, self-esteem, and happiness.Statistical analyses were conducted using AMOS and SPSS packages.The significance level was accepted as .05for item discriminant analysis.Bonferroni correction method was applied for correlational analysis' p values (Manly, 2005).

Results
The aim of this study was to examine the psychometric characteristics of the RESE with a sample of Turkish university students.With this aim, firstly explanatory factor analysis (EFA) with promax rotation for increasing the factors' relation and then confirmatory factor analysis (CFA) were used.The RESE has specific expectations when applied to different cultures from Turkish population.Therefore, EFA were used before CFA.During explanatory factor analysis, the aim was to reach the existence of positive affects of the three dimensions determined in the original form of the RESE as follow: a) perceived self-efficacy in managing anger/irritation, b) perceived self-efficacy in managing despondency and c) perceived self-efficacy in expressing affect.
During confirmatory factor analysis, it was concluded that the factors of perceived self-efficacy in managing anger/irritation, and perceived self-efficacy in managing despondency formed the dimension of perceived self-efficacy in managing negative affect in the second level and the structural model of Caprara and his colleagues (2008) on the issue of this dimension covariated with the factor of perceived self-efficacy in expressing positive affect (see Table 1).Note.F 1 = Perceived self-efficacy in expressing positive affect (POS); F 2 = Perceived self-efficacy in managing despondency (DES); F 3 = Perceived self-efficacy in managing anger/irritation (ANG); h 2 = Communalities.
As a result of exploratory factor analysis, 3 factors whose promax rotation and eigenvalue were over 1, were obtained.There is no secondary factor loading for all items.The first factor of them whose eigenvalue was 2.27 explained the 18.93% of the total variance, the second factor whose eigenvalue was 2.12 explained 17.69% of the total variance and the third factor whose eigenvalue was 1.29 explained 10.77% of the total variance; and these three factors in total explained 47.39% of the total variance (KMO of sampling adequacy = .72,Bartlett's test of sphericity χ 2 (66) = 483.56,p =< .05).It is therefore suggested that the RESE maintained its three-factored structure of the original in the Turkish form and that following confirmatory analysis, as in the original model, a negative second-order factor formed from two first-order factors and a first-order positive factor model were used (see below,  As a result of the confirmatory factor analysis, it was determined that the dimension of perceived self-efficacy in managing negative affect (NEG) was a second-order factor with two first-order factors as perceived self-efficacy in managing anger/irritation (ANG) and perceived self-efficacy in managing despondency (DES) and it was covariated with perceived self-efficacy in expressing positive affect (POS) which is a first-order factor (see Figure 1).
All of the standardized parameter estimations were statistically significant in a positive way, and they were determined as .39-.68 for perceived self-efficacy in managing anger/irritation, as .38-.56 for perceived self-efficacy in managing despondency and .49-.69 for perceived self-efficacy in expressing positive affect.This result, because of determining the confirmation of two dimensions as positive and negative, was examined over two genders as multi-group factor analysis evaluated as gender invariance.According to the partial metric invariance model, the difference in men and women in terms of regulatory emotional self-efficacy was insignificant in metric invariance model (Δχ 2 = 12.93, df = 10, p = .23).Internal consistency studies were conducted via Cronbach alpha.According to the results, it is suggested that the internal consistency was determined as .52 for perceived self-efficacy in managing anger/irritation, as .58for perceived self-efficacy in managing despondency, as .61for perceived selfefficacy in managing negative affect, as .69for perceived self-efficacy in expressing positive affect and as .60 for the total score.Descriptive statistic results as shown in Table 2 indicated that item 1, 2, 3, 4, 8, 9, and 12 had higher average middle point than the item scaling's middle point 3, which is the mid-point of the scale items and other items were a little lower than the mid-point.When the distinctiveness powers of sub and normal groups of 27% were examined, it was concluded that the results of independence of t-tests showed sufficiency in separating the individuals who took high and low scores from all the items for self-efficacy in managing negative affect and perceived self-efficacy in expressing positive affect.
When the relationships between the two dimensions which are within the regulatory self-efficacy fields and selfefficacy, self-esteem, and happiness were examined, it was observed that the relationship between perceived self-efficacy in managing negative affect and perceived self-efficacy in expressing positive affect was quite low and insignificant (see Table 3).In addition, it was concluded that perceived self-efficacy in managing negative affect had an effect size with emotional self-efficacy at the level of r 2 = .04,with self-esteem at the level of r 2 = .02 and with happiness at the level of r 2 = .02;perceived self-efficacy in expressing positive affect had a correlation with emotional self-efficacy at the level of r 2 = .21 in small coefficients.measuring the concept of emotional self-efficacy while especially for different factor structures obtained in China (Zhao et al., 2013), regulatory emotional self-efficacy had different qualifications in Chinese culture.Some studies examined the links between ill-being and emotional regulation (Garnefski & Kraaij, 2006;Garnefski, Kraaij, & Spinhoven, 2001;Silk, Steinberg, & Morris, 2003).The importance of measurement scales of regulatory emotional self-efficacy increases.So, the RESE can be used both in psychotherapy, psychological counseling and other mental health services and also in the studies oriented to the effects of resourcing to the variables in the field of positive psychology.

Limitations and Suggestions
The current study suggested that the RESE's total structure is a valid and reliable scale when adopted for a Turkish sample group of students and it also maintained its original form which had been previously confirmed in Bolivia, Italy, and USA (Caprara et al., 2008).But, the study has got some limitations.Firstly, it is the scale's psychometric characteristics, which were examined with a specific sample group of students from the faculty of education.This scale should be examined with other population groups from different professions and different age groups in Turkey.Secondly, multiple comparisons can be made, when such groups experience mood disorder in which the effect of negative experiences is clear like post-traumatic stress disorder (PTSD), and clinical depression.In addition, multiple group comparisons can be made with neighbors cultures such as Bulgarian, Greek, Arabian, Persian, Georgian, and Armenian.Finally, similar scale validity, incremental and decremented validity studies of the RESE should be developed.

Figure 1 .
Figure 1.Confirmatory factor analysis for the RESE.

Table 1
Exploratory Factor Analysis With Promax Rotation for RESE

Table 2
Descriptive Statistics for the RESE's Items, and Item Discriminant Between Lower and Upper 27% GroupsNote.a 27% lower-upper groups' n = 82, degree of freedom = 162.