Social Constraints Among Bereaving Turkish Adults: Testing The Construct And Convergent Validity And Relıability of the Social Constraints Scale (Scs) 1

AJESI - Anadolu Journal of Educational Sciences International, 2018; 8(2): 204-231 DOI: 10.18039/ajesi.454571 Social Constraints Among Bereaving Turkish Adults: Testing The Construct And Convergent Validity And Relıability of the Social Constraints Scale (Scs)1 Idil AKSOZ-EFE 2, Ozgur ERDUR-BAKER3 Submitted by 11.03.2018 Accepted by 27.07.2018 Research Paper Abstract Recent studies revealed the significance of social support on human functioning and wellbeing, and also the adverse effect of social constraints and lack of support. Especially after negative life events such as loss of a loved one, individuals search for support and when not found, adaptation becomes problematic. Thus, measuring and monitoring the perceptions of social constraints is important as well as the social support. This study aimed to conduct the initial reliability and validity analyses of the Social Constraints Scale (SCS) in a Turkish population. SCS and Two-Track Bereavement Questionnaire (TTBQ) were administered to 323 bereaved individuals (222 females and 101 males) with a mean age of 35.31 (SD = 14.63). Relational active grieving (RAG) and biopsychosocial functioning subscales of TTBQ were used to test the criterion-related validity of the SCS. Results suggested that SCS had good reliability and validity results, and had significant positive correlations with both RAG and biopsychosocial functioning. Moreover, according to the results, when the bereaved perceive more social constraints, they also have more problems with functioning and maladaptive grieving. Results of the initial construct validity, criterion-related validity and internal consistency tests showed that, the Turkish SCS is a valid and reliable scale. Keywords: Social constraints scale, bereavement, validity, reliability 1 This study is part of the PhD thesis of the first author under supervision of the second author. 2 Atatürk Üniversitesi, e-mail: aksoz.efe@atauni.edu.tr 3 Orta Doğu Teknik Üniversitesi, e-mail: erdur@metu.edu.tr 204 AJESI - Anadolu Journal of Educational Sciences International, 2018; 8(2): 204-231 DOI: 10.18039/ajesi.454571 Introduction In dealing with difficult life events, the importance of social support is well established as related literature reports substantial empirical evidence. Social sharing and social support help to normalize negative life experiences (Vachon & Stylianos, 1988); reduce the intensity of negative emotions and boost the positive mood (McGloshen & O’Bryant, 1988; Sheridan, Sherman, Pierce, & Compas, 2010); help to reduce grief symptoms while adjusting to the loss of a loved one (Riley, LaMontage, Hepworth, & Murphy, 2007); help individuals feel loved, understood, cared and provide them courage to cope with negative events and maintain well-being (Lepore, 2001); and improve physical and psychological states (Pennebaker & Harber, 1993). But what if the existing social environment does not facilitate sharing and is not willing to provide support that much needed by the stressed individuals? What if the social network constrains individuals by introducing emotional, informational, or instrumental barriers? Studies point out that, social constraints described as the perceived communication barriers especially after stressful situations, may have adverse effects on functioning (Lepore, 1992, 2001; Lepore, Silver, Wortman, & Wayment, 1996; Schuster, Kessler, & Aseltine, 1990). Even though extensive research is available on the protective role of social support, much less is known about the nature of social constraints in terms of how they are perceived and how they impact psychological responses of individuals experiencing adverse life events. Moreover, existing knowledge about culture-sensitivity of the definition of social constraint is less than sufficient. Therefore, the current study focuses on the exploration of social constraints among bereaving adults in a Turkish sample and examination of the psychometric properties of the Turkish version of the Social Constraints Scale. Three types of social constraints were defined by Lepore et al. (1996), (a) lack of supportive and empathic individuals in the social network, (b) inappropriate suggestions of 205 AJESI - Anadolu Journal of Educational Sciences International, 2018; 8(2): 204-231 DOI: 10.18039/ajesi.454571 the social network, and (c) lack of the social network in times of disclosures of negative life event survivors. Regardless of their type, when constraints occur, they cause to be felt isolated, stigmatized, misunderstood, and unsupported (Lepore, 2003). Furthermore, when individuals in stressful situations (i.e., cancer patients) perceive social constraints, they tend to suppress their emotions, and have higher levels of stress. On the contrary, when stressed individuals perceive support, they maintain their well-being (Lepore, 2001). After a major life event revealing emotions, such as death of a loved one, individuals seek for social support to disclose their emotions and thoughts (Rime, Mesquita, Philippot, & Boca, 1991). When the basic security beliefs of individuals’ are threatened, they search for support from their social network to disclose, thus cope with the difficulties they face with, and recover from the overwhelming emotions. Sharing the emotions in the social network plays a critical recovering role in the aftermaths of negative events, and starts systematically soon after the event (Rime et al., 1991). After the loss of a loved one, bereaved individuals need to experience their grief process in a social network, not in isolation (Pressman & Bonanno, 2007; Stroebe & Schut, 1999). Sharing emotions and thoughts about the deceased may strengthen the bonds among the bereaved and their network (Pressman & Bonanno, 2007) as well as helping them adjust to the loss and function better (Pennebaker & Harber, 1993; Riley et al., 2007). On the other hand, lack of social support was defined as a risk factor for bereaved (Sanders, 1988) and was associated with lower positive emotions (McGloshen & O’Bryant, 1988), higher stress levels and intrusive thoughts and feelings (Lepore, 2001), and higher depressive symptoms (Lepore et al., 1996). In order to measure and gain a deeper understanding of the social constraints, Lepore et al. (1996) developed Social Constraints Scale (SCS). The 10-item scale, 5 items repeated twice for (a) “the most significant one” for the bereaved and (b) “others” in the social network, aimed to explore the social constraints experience. Lepore et al. (1996) presented a 206 AJESI - Anadolu Journal of Educational Sciences International, 2018; 8(2): 204-231 DOI: 10.18039/ajesi.454571 positive association between intrusive thoughts and social constraints among mothers bereaved by the loss of their infants. This relationship appeared to be positively moderated by social constraints. That is, when bereaved mothers experienced higher levels of social constraints, their intrusive thoughts had stronger associations with their depressive symptoms. On the other hand, when perceived less social constraints, the association between their intrusive thoughts and depressive symptoms was weaker (Lepore et al., 1996). Later on, Lepore and Ituarte (1999) adapted the SCS for the trauma survivors, specifically for cancer patients. The 15-item version of the scale (SCS-15) was administered to 97 female cancer patients and results showed that the patients were optimistic and experienced lower levels of negative emotions when they were supported by friends and spouses and experienced higher levels of negative emotions when they were constrained (Lepore & Ituarte, 1999). Aside from bereavement, several studies explored the relationships between social constraints and stress (Lepore, Ragan, & Jones, 2000; Lewis, Derlega, Clarke, & Kuang, 2006); quality of life (Eton, Lepore, & Helgeson, 2001); PTSD symptoms (Groff, 2014); mood and emotional experience (Agustsdottir et al., 2010; Lewis et al., 2006; Schmidt, 2002); depression (Agustsdottir et al., 2010; Braitman et al., 2008; Groff, 2014; Ullrich, Lutgendorf, & Stapleton, 2002); and rumination (Pennebaker & Harber, 1993), showing the importance of monitoring and measuring the social constraints. These studies presented that perceptions of social constraints are positively associated with stress, avoidance, intrusive thoughts, anxiety, rumination and depression, and are negatively associated with quality of life, positive mood and emotional experience, optimism, and adjustment to bereavement. Lepore (2001) also found that while experiencing stressful life events, perceptions of social constraints may result in additional traumas. Furthermore, gender, age and cultural background are significant determinants of social interactions and disclosures of bereaved individuals. Gender and age influence the 207 AJESI - Anadolu Journal of Educational Sciences International, 2018; 8(2): 204-231 DOI: 10.18039/ajesi.454571 social interactions (Farberow, Gallagher-Thompson, Gilewski, & Thompson, 1992; Lubben, 1988; Oliver, Pearson, Coe, & Gunnell, 2005; Pennebaker & Harber, 1993; Vaswani, 2011), emotional abilities and expressivity (Lindinger-Sternart, 2015; Vaswani, 2011), and helpseeking behaviors (Mackenzie, Scott, Mather, & Sareen, 2008; Oliver et al., 2005). Females and older people seek more help than males and younger people, and females are more likely to express their emotions. Studies also revealed that, social networks of older widowers generally consist of small groups of men (Farberow et al., 1992; Lubben, 1988) and widowers perceive less support than widows (Fry, 2001; Stelle & Uchida, 2004; Stroebe, Stroebe, & Abakoumkin, 1999), showing that usually older widowers are more isolated from their social network than widows (Stelle & Uchida, 2004); and elderly bereaved think and share about their emotions less than the younger bereaved (Pennebaker & Harber, 1993). Considering together, these results suggest that age and gender are significant indicators, in other words, males and seniors are more prone to experience social constraints. In a broader perspective, culture determine the differences in emotional experiences and socially acceptable disclosures of emotions after loss of a loved one (Bolak-Boratav, Sunar, & Ataca, 2011; Hunter, 2007; Nagel, 1988; Nordanger, 2007); searching for meaning, thinking about the deceased and patterns of sharing emotions and thoughts (Pressman & Bonanno, 2007); help-seeking behavior (Mojaverian, Hashimoto, & Kim, 2013; Ono, Araki, Mitani, & Yoshimura, 1999); and supportive or unsupportive perceptions of the social networks (You & Lu, 2014). In their study investigating the Turkish cultural display rules of emotions, Bolak-Boratav et al. (2011) found that happiness was the most acceptable emotion to display, and contempt and disgust were the least acceptable. They also found that it is more acceptable to display the emotions in a private setting then a public setting, and displaying sadness is more acceptable towards women than men, and towards people at the same level of social status than high or low status (Bolak-Boratav et al., 2011). In other words, cultural 208 AJESI - Anadolu Journal of Educational Sciences International, 2018; 8(2): 204-231 DOI: 10.18039/ajesi.454571 context define the ones to disclose (i.e., who), the ones to share with (i.e., to whom), the topics to disclose (i.e., what), and the conditions to disclose (i.e., how and where). While the cultural perceptions on social support was revealed by the related literature, little is known about the cross-cultural perceptions and equivalency of social constraints. Thus, in order to understand the prevalence and structure of social constraints, it is important to study the subject in different cultures. Since the influence of social constraints in stressful life events is manifested in several studies, it is important to have tools to measure it. Recently, studies emerged from nonEnglish-speaking samples (i.e., Chinese by You and Lu, 2014 and Greek by Koutrouli, Anagnostopoulos, Tsikkinis, Papastylianou, and Lepore, 2015), and showed the associations of social constraints with quality of life, emotional experience, and stress, using SCS-15 for trauma survivors. You and Lu (2014) expressed the need for research on social constraints among minority groups and in different cultural settings, and explored its relationship with quality of life in a Chinese breast cancer survivor group. They found that social constraints were negatively associated with quality of life and positive affect, and positively associated with intrusive thoughts and negative affect among breast cancer survivors. They also compared the scores of the Chinese sample to scores of Caucasian samples from previous studies, and found cultural differences revealing that Chinese sample reporting higher social constraints than Caucasion sample (You & Lu, 2014). Also, Koutrouli et al. (2015) explored the psychometric properties of the Greek SCS-15. Greek version of the scale had a threefactor structure; unsupportive behaviors, avoidant behaviors, and suggestions for distraction and pretense, all significantly correlated with intrusions and distress. Unsupportive behaviors showed the highest correlation with distress while suggestions for distraction had the highest correlation with intrusive thoughts (Koutrouli et al., 2015). 209 AJESI - Anadolu Journal of Educational Sciences International, 2018; 8(2): 204-231 DOI: 10.18039/ajesi.454571 These studies show the cross-cultural validation of the SCS and explore the social interactions in different settings with the evidence of cultural differences. However, literature needs more cross-cultural evidence from different samples to gain a universal understanding of the topic and provide valid and reliable tools to measure it. In Turkish literature, there are no studies exploring the social constraints among bereaved individuals. Social interactions in Turkish literature is limited to studies investigating the negative perceptions of social support. Thus, this study aims to adapt the SCS into Turkish and test its initial psychometric properties to respond to the gap in the literature. Methodology In this section, the characteristics and selection method of the participants are presented. Then, data collection procedure is described and finally, psychometric properties of the measures used in the current study are reported. Participants and Procedure Target population of this study was Turkish adults who experienced loss of a loved one in the last 5 years. This limit was set to follow Rubin et al.’s (2009) suggestion that grieving process of the bereaving individuals change significantly after five years. The inclusion criteria were (a) being an adult, and (b) experiencing the loss of a loved one in last five years. After gathering the approval from Middle East Technical University Human Subjects Ethics Committee, the data were collected via pen and pencil forms from volunteered bereaving individuals by researcher in approximately two weeks. Students and personnel of Middle East Technical University were reached by the first author, and they were also asked to refer other bereaved individuals as possible participants. In addition to these 210 AJESI - Anadolu Journal of Educational Sciences International, 2018; 8(2): 204-231 DOI: 10.18039/ajesi.454571 participants, bereaved individuals in the first author’s network (i.e., distant relatives and acquaintances) were reached. The final study group included participants from different regions of Turkey. A total of 323 bereaving (222 females and 101 males) with an age range of 18 to 81 (M = 35.31, SD = 14.63) participated in the study. Since the purpose of the current study was to adapt and test the psychometric analyses of the Social Constraints Scale, following Sousa and Rojjanasrirat’s (2011) suggestions, at least five participants per item were aimed to be reached for the study group. Age of the participants were categorized depending on Bischof’s (1976) and Santrock’s (2011) suggestions (early adulthood = 18-25, middle adulthood = 26-40, late adulthood = 41-60, old age = 61-beyond). Of the participants, 109 (33.7%) of them were in early adulthood group, 115 (35.6%) were in middle adulthood group, 79 (24.5%) were in late adulthood group, and 20 (6.2%) were in old age group. According to the participants’ reports, 139 (43%) of the deceased were females and 184 (57%) were males, and the mean age of the deceased was 62.82 (SD = 21.30) with a range of 9 to 101. The mean for the time since death was 2.63 years (SD = 1.59). Of the reported losses, 154 (47.7%) of them were natural anticipated (e.g., cancer, old age), while 125 (38.7%) of them were natural sudden (e.g., heart attack) and 42 (13%) of them were violent (e.g., suicide). Of the deceased ones, 91 (28.2%) of them were grandparents of the participants, while 59 (18.3%) were parents, 56 (17.3%) were aunts and uncles, 37 (11.5%) were in-laws, 21 (6.5%) were other acquaintances, 20 (6.2%) were friends, 17 (5.3%) were cousins, 11 (3.4%) were siblings, 10 (3.1%) were spouses, and 1 (0.3%) was child of the participants. Only 18 (5.6%) of the participants reported to get professional psychological help after their loss while 297 (92%) had no professional help. Measures 211 AJESI - Anadolu Journal of Educational Sciences International, 2018; 8(2): 204-231 DOI: 10.18039/ajesi.454571 The Social Constraints Scale (SCS). To measure the social constraints of the bereaved in disclosing their thoughts and feelings, Lepore et al. (1996) developed a 10-item 5point rating scale (1 = almost never to 5 = almost always). They explored the relationships among depressive symptoms, intrusive thoughts and social constraints of 98 bereaved mothers who lost their infant babies because of SIDS. More specifically, the purpose was to assess the social barriers occurring when bereaved mothers wanted to disclose their trauma related thoughts and emotions. The scale asks the perceptions of social constraints with five questions posed twice for “the most important person” for the bereaved and “other people” in the social network separately. First, a two-factor structure was tested by Lepore et al. (1996), however, exploratory factor analysis results revealed a one-factor structure, accounting for the 27% of the variances. The total score is used to define the level of the constraints; higher scores indicate higher levels of constraints. The internal consistency value was α = .77 (Lepore et al., 1996). The Two-Track Bereavement Questionnaire (TTBQ). Rubin et al. (2009) developed TTBQ to explore the responses to loss of a loved one. The 56-item questionnaire consists of five factors; relational active grieving (RAG), close and positive relationship with the deceased, conflictual relationship with the deceased, general biopsychosocial functioning, and traumatic perception of the loss. For all the factors, higher scores indicate higher levels of grief, more problematic adjustment and unhealthy attachment to the deceased. The reliability values were ranging from .75 to .94. Rubin et al. (2009) also reported that according to the higher order analysis, five factors loaded under two tracks; Track I was assessing the functioning of the bereaved (general biopsychological functioning and traumatic perception factors) and Track II was assessing the relationship to the deceased (RAG, close and positive relationship and conflictual relationship factors). Authors suggested different ways of scoring; using a total 212 AJESI - Anadolu Journal of Educational Sciences International, 2018; 8(2): 204-231 DOI: 10.18039/ajesi.454571 score for the whole questionnaire, using any of the tracks or factors. TTBQ was adapted into Turkish by Ayaz, Karancı, and Aker (2014) and 5-factor structure was suggested with few items loading under different factors from the original structure. They also confirmed the two track structure with higher order analysis. Overall, the reliability values were ranging from .65 to .91 (Ayaz et al., 2014). In the current study, RAG subscale assessing the influences of maladaptive grieving, and biopsychosocial functioning subscale exploring the dysfunctioning of the bereaved were used to test the criterion-related validity of the SCS. Results This section includes the translation process, content validity, construct validity and convergent validity of the Social Constraints Scale. Relationships between social constraints and demographic variables such as age and gender are also given in this section. Finally, criterion-related validity and internal consistency reliability of the Social Constraints Scale are presented. Translation Process and Content Validity of SCS After gathering the author permission to use and adapt the scale, five PhD students proficient in both Turkish and English (one with English Language Teaching background, three from psychological counseling field, and one from educational studies field) translated the items. After the comparison of translations, back translation of the chosen items were conducted to ensure the consistency of meaning. Finally, three academicians from psychological counseling field were asked to provide feedback and so the translation was finalized. During the whole process, guidelines of International Test Commission (2010) were 213 AJESI - Anadolu Journal of Educational Sciences International, 2018; 8(2): 204-231 DOI: 10.18039/ajesi.454571 followed. Criterion-Related Validity and Internal Consistency Reliability of SCS Pearson correlation coefficients among SCS, RAG and biopsychosocial functioning were calculated to test the criterion-related validity. SCS scores had significant positive correlations with both RAG (r = .22, p < .01) and biopsychosocial functioning (r = .37, p < .01), as seen in Table 1 below. Cronbach’s alpha coefficient was computed to examine the internal consistency of SCS. Reliability value was α = .72 for the 5-parcel structure. Table 1 Descriptives for Variables and Pearson’s Correlations between SCS and Other Variables M SD Skewness Kurtosis α Pearson’s r RAG 36.85 11.89 .76 .09 .88 .22** Functioning 20.02 5.77 .94 1.85 .70 .37** Variable Functioning: Biopsychosocial Functioning. ** p < .01. Construct and Convergent Validity of SCS One-factor structure was tested with Confirmatory Factor Analysis (CFA) via Amos 18. Before running CFA, sample size and missing data, normality, outliers, linearity, and multicollinearity assumptions (Kline, 2011) were checked. No violations of the assumptions were observed. Normality test results showed that the distribution was moderately non-normal since skewness values were smaller than 3 and kurtosis values were smaller than 8 (Kline, 214 AJESI - Anadolu Journal of Educational Sciences International, 2018; 8(2): 204-231 DOI: 10.18039/ajesi.454571 2011). Thus, maximum likelihood (ML) estimation was used and hypothesized model was tested via bootstrapping method with 2000 bootstrapped samples and 95% confidence interval (CI) to handle with the potential effects of non-normal data. In order to interpret the results of CFA, fit indexes as Root Mean Square of Error of Approximation (RMSEA), The Bentler Comparative Fit Index (CFI), Tucker-Lewis Index (TLI), and Standardized Root Mean Square Residual (SRMR) were used with model chisquare (χ2) and chi-square/ degrees of freedom ratio (χ2/df-ratio) values. Before the results, the cut points of the fit indexes are summarized to have a better understanding. Chi-square result defined to be small (i.e., less than 3) and non-significant in the perfect fit (Kline, 2011). In terms of RMSEA criteria, Browne and Cudeck (1993) suggest that, RMSEA < .05 indicates good fit, and RMSEA < .08 indicates reasonable fit. Moreover, MacCallum, Browne, and Sugawara (1996) suggested RMSEA values ranging between .08 and .10 indicate mediocre fit and values above .10 indicate poor fit. MacCallum et al. (1996) also suggest reporting RMSEA values with the confidence intervals (CI) and indicate that the confidence intervals should be around RMSEA value. About the confidence intervals (CI), Kline (2011) suggests the cut-off points as lower bound of CI ≤ .05 and upper bound of CI ≤ .10. In addition to confidence intervals, AMOS also calculate the closeness of fit (pClose) and according to Jöreskog and Sörbom (1996) this value should be non-significant, i.e., pClose > .05. Both CFI and TLI values range from 0 to 1 and .95 is the cut-off value for both of the fit indices (Hu & Bentler, 1999) and .90 is acceptable for both of the indices (Schumacker & Lomax, 2010). SRMR values ranges from 0 to 1, and smaller values show better fit. A SRMR value less than .08 (Hu & Bentler, 1999) is acceptable. Results of CFA showed poor fit with 10-item structure (χ2(35) = 271.606, p < .05, χ2/df-ratio = 7.76, CFI = .79, TLI = .73, RMSEA = .15 [90% CI = .13,.16], pClose < .05, and SRMR = .09), and suggested freely estimating the error terms of the repeated items asked for 215 AJESI - Anadolu Journal of Educational Sciences International, 2018; 8(2): 204-231 DOI: 10.18039/ajesi.454571 the most significant one of the bereaved and other people in the social network (i.e., items 16, 2-7, 3-8, 4-9, and 5-10). Thus, item parceling technique (Kline, 2011) was used by adding the repeated items to create a total score and five parcels were formed. Results of CFA for the parceled structure showed good fit (χ2(5) = 12.98, p < .05, χ2/df-ratio = 2.596, CFI = .98, TLI = .97, RMSEA = .07 [90% CI = .02,.12], pClose > .05, and SRMR = .03). All standardized estimates were above .40, except from one (i.e., parcel2) and all the regression weights were significant. The items of the parcels and results of the CFA for the parceled structure of the scale are presented in Table 1 below. The standardized estimates given in Table 2, shows the loadings of the parcels as indicators of convergent validity of SCS. According to Kline (2011), when all indicators of a factor have relatively high loadings (e.g., > .70), the results indicate convergent validity. Results shows that one parcel has moderate (.59) and one parcel has low (.16) loading, indicating doubtful convergent validity. However, poor convergent validity suggests that the measure has “too few factors” (Kline, 2011, p. 240). Table 2 Skewness and Kurtosis Values, Standardized Estimates and Regression Weights of the Parcels Standardized Regression Parcel Items Skewness Kurtosis Estimates Weights parcel1 1+6 1.56 2.27 .59 p < .001 parcel2 2+7 -.28 -1.01 .16 p < .01 parcel3 3+8 1.80 3.23 .73 p < .001 parcel4 4+9 2.03 5.05 .87 p < .001 parcel5 5 + 10 1.69 3.85 .78 p < .001 Moreover, as a second indicator of convergent validity, Average Variance Extracted (AVE) and Composite Reliability (CR) were computed. AVE was found as .45, which was below the threshold of .50 and insufficient to represent convergence. According to Huang, 216 AJESI - Anadolu Journal of Educational Sciences International, 2018; 8(2): 204-231 DOI: 10.18039/ajesi.454571 Wang, Wu, and Wang (2013), AVE values higher than .45 could be acceptable when the CR value is higher than .60. For the current study, the CR was calculated as .53, thus indicating poor convergence. Relationships between Social Constraints and Age and Gender In order to gain deeper understanding on the concept of social constraints, and point out its possible relations with age and gender, multi-group comparisons were conducted by running ANOVA. Since the sample sizes for each group are small (i.e., less than 200), the power of multi-group CFAs would be limited (Kline, 2011), thus a 4X2 ANOVA was conducted to evaluate the effects of age groups (1 = early adulthood, 2 = middle adulthood, 3 = late adulthood, and 4 = old age) and gender on social constraints scores. Before running ANOVA, Levene’s test showed that the homogeneity of variance assumption is not violated (F(7,315) = 1.33, p > .05). ANOVA results indicated no significant interaction between gender and age groups (F(3,315) = 1, p > .05) and no significant main effect for both gender (F(1,315) = .24, p > .05) and age groups (F(3,315) = 1.66, p > .05). That is, there were no significant group differences for neither gender nor age in terms of their social constraints scores. The results are presented in Figure 1 and Table 3 below. 217 AJESI - Anadolu Journal of Educational Sciences International, 2018; 8(2): 204-231 DOI: 10.18039/ajesi.454571 Figure 1. Effects of age groups and gender on social constraints scores (N = 323). Table 3 ANOVA Results for the Effect of Age Groups and Gender on Social Constraints Scores Source SS df MS F η2 Gender 9.74 1 9.74 .24 .001 Age group 198.79 3 66.26 1.66 .016 Gender * Age group 119.46 3 39.82 1 .009 Error 12565.64 315 39.89 Total 12898.83 322 As seen from the Figure 1, except for the early adulthood group (age group 1), males reported higher levels of social constraints than females. All means and standard deviations are presented in Table 4 below. 218 AJESI - Anadolu Journal of Educational Sciences International, 2018; 8(2): 204-231 DOI: 10.18039/ajesi.454571 Table 4 Means and Standard Deviations of Social Constraints Scores for Age and Gender Groups M SD Age group 1 (18-25) 20.05 6.56 Age group 2 (26-40) 18.05 7.03 Age group 3 (41-60) 20.09 4.94 Age group 4 (61-beyond) 19.03 5.45 Age group 1 (18-25) 18.49 6.54 Age group 2 (26-40) 19.09 5.30 Age group 3 (41-60) 21.63 8.30 Age group 4 (61-beyond) 19.88 4.26 Females Males Discussion Several studies revealed the significance of social support on human functioning and well-being (Lepore, 2001; McGloshen & O’Bryant, 1988; Pennebaker & Harber, 1993; Riley et al., 2007; Sheridan et al., 2010; Vachon & Stylianos, 1988), and the adverse effect of lack of support (Lepore et al., 1996; Schuster et al., 1990). Social constraints were also found to have more powerful influences on well-being than social support (Lepore, 1992, 2001; Lepore & Ituarte, 1999). Especially after negative life events such as loss, individuals search for support and when not found, adaptation becomes problematic (Lepore et al., 1996; Sheridan et al., 2010; Vachon & Stylianos, 1988). Therefore, measuring and monitoring the perceptions of social constraints is important as well as the social support. While social support is commonly tested, further studies should also consider measuring the constraints simultaneously to better understand the influences of social environment. Also, testing the cross-cultural psychometric properties of SCS will help the researchers and practitioners gain a deeper acknowledgement on social mechanisms. Thus, current study aimed to explore the 219 AJESI - Anadolu Journal of Educational Sciences International, 2018; 8(2): 204-231 DOI: 10.18039/ajesi.454571 relationships between social constraints and functioning after the loss of a loved one in a Turkish sample, and test the reliability and validity of SCS. The significant relationship between social constraints and adaptation to the loss that was found in the current study has important implications for mental health field. Although the design of the study is correlational and do not indicate causal associations among the variables, results are still indicators of the relationships between problematic bonds with the deceased and difficulties in functioning of the bereaved and perceptions of the support or constraints from the social network. Also, considering the social differences in help-seeking behaviors and emotional expressivity, further studies should validate the cross-cultural structure of the social constraints. Chinese (You & Lu, 2014) and Greek (Koutrouli et al., 2015) cancer survivor samples showed that social constraints are negatively correlated with quality of life and positive affect, and positively correlated with intrusive thoughts and negative affect. Similarly, current study provides supportive evidence for cross-cultural validation by indicating the negative associations between social constraints and functioning after the loss. Current study made an important contribution with exploring the relationships among social constraints, age, gender and adaptation to loss, and tested the initial psychometric properties of SCS in a Turkish sample. Results of ANOVA showed no significant age and gender differences in SCS scores. However, males reported higher levels of constraints than females for all age groups except for the youngest group (i.e., 18-25 ages). Despite being nonsignificant, these findings were consistent with the early study results, revealing male gender and old age groups as disadvantaged groups in terms of social constraints (Farberow et al., 1992; Fry, 2001; Lubben, 1988; Stelle & Uchida, 2004; Stroebe et al., 1999). These results can be explained by the fact that males have lower levels of emotional abilities and expressivity (Lindinger-Sternart, 2015; Vaswani, 2011), and tend to seek less help from 220 AJESI - Anadolu Journal of Educational Sciences International, 2018; 8(2): 204-231 DOI: 10.18039/ajesi.454571 others (Mackenzie et al., 2008; Oliver et al., 2005). Higher levels of social constraints of males may be explained by gender roles in Turkish culture, too. Studies on Turkish samples show that males have higher levels of suppression and masking their emotions, and have lower levels of emotional expressivity and impulse strength than females (Tunay-Akan & Barışkın, 2016), and are less likely to express their happiness, surprise, sadness and fear than females, and think that expressing their happiness is more appropriate than expressing their sadness and fear (Bolak-Boratav et al., 2011). These gender roles on emotional expressivity may play a critical role in social communication; assuming that males will not express their emotions, the social network may constrain the bereaved. However, these gender differences were tested by ANOVA in the current study. In order to better explore the gender and age group differences, multi-group analyses should be conducted. A limitation of the current study was the unequal and relatively small group sizes which enabled to control the invariances. Thus, further studies should be planned accordingly to reach for balanced groups and test the invariances. Results of the current study revealed significant associations between social constraints and adjustment to the loss of a loved one. Acknowledging gender and age differences, this study provides beneficial information for the mental health field. Monitoring the resources and risk factors of the bereaved clients, support groups should be promoted. Furthermore, mental health professionals should promote and support bereaved clients to express their loss-related thoughts and feelings. Bereaved individuals who express their thoughts and feelings in the safe, understanding, and caring atmosphere of the counseling session will satisfy their needs for sharing their experiences, being listened and approved by others. In terms of the psychometric properties of the Turkish SCS, internal consistency was calculated with Cronbach alpha coefficient and found as .72. Additionally, to control the 221 AJESI - Anadolu Journal of Educational Sciences International, 2018; 8(2): 204-231 DOI: 10.18039/ajesi.454571 construct validity, two CFAs were run. Depending on CFA results, item parceling technique (Kline, 2011) was used and the one factor structure was confirmed. Results also showed initial evidence for the validity and reliability of the Turkish SCS. Since the measure uses total score to evaluate the level of constraints originally, item parceling technique did not change the evaluation criteria or the factor structure (Kline, 2011). One limitation of the study was the low standardized estimate of the 2nd parcel (i.e., parcel created by adding the 2nd and 7th items), including the reversed-coded items, asking for perceptions of support. In order to evaluate the low loadings of these items, the results should be further tested with different samples and evaluated carefully. Another limitation of the current study was its insufficient convergent validity values (i.e., both the standardized estimates and average variance extracted values). In order to deal with this problem, further studies should test different factor structures, as Kline (2011) suggests that low values of convergent validity indicate possible additional factors. Moreover, to gain a better understanding of the validity and reliability of the Turkish Social Constraints Scale, further studies should also test the divergent validity and test-retest reliability. Construct validity of the scale was also tested with Relational Active Grieving and Biopsychosocial Functioning subscales of TTBQ, and SCS showed significant correlations with both of them, showing the association between the perceptions of social constraints and difficulties in adaptation to the loss of a loved one. Results of the current study are also supporting former studies exploring the relationship between social constraints and psychological distress (Clark, 1993; Lepore, 1992; Lepore, Fernandez-Berrocal, Ragan, & Ramos, 2004; Lepore et al., 2000), health related stress and quality of life (Eton, Lepore, & Helgeson, 2001, 2005; Koutrouli et al., 2015; Lepore, 2001), and discrimination (Henson, Derlega, Pearson, Ferrer, & Holmes 2013; Lewis et al., 2006). These studies showed that when individuals have social support, they cope better with difficult life events such as stress, 222 AJESI - Anadolu Journal of Educational Sciences International, 2018; 8(2): 204-231 DOI: 10.18039/ajesi.454571 life threatening illness, discrimination and stigmatization. However, when they perceive constraints, they report higher levels of stress and lower abilities of coping. While supporting the former findings, results of the current study showed significant but low correlations. Thus, the relationships among the related variables should be considered carefully. Current study assessed the grief related stress level of the participants by using TTBQ, however, grief related stress could be correlated with other variables, such as emotion regulation and social support. The low correlation coefficients could be interpreted as an indicator of mediator variables on the relationship between social constraints and difficulties in adaptation to the loss of a loved one. Therefore, further studies should consider these low coefficients and use other variables to test the construct validity. Current study explored social constraints only from the perspectives of the bereaved, which is another limitation. The nature of the social constraints includes social interaction with others, thus additional information resources should be added into the research (You & Lu, 2014). Further studies should reach to a comparison group of caregivers and significant others, or synthesize self-report and observatory data to gain a better understanding of the social constraints mechanism. References Agustsdottir, S., Kristinsdottir, A., Jonsdottir, K., Larusdottir, S. O., Smari, J., & Valdimarsdottir, H. B. (2010). The impact of dispositional emotion expressivity and social constraints on distress among prostate cancer patients in Iceland. British Journal of Health Psychology, 15, 51-61. doi: 10.1348/135910709X426148. Ayaz, T., Karancı, N., & Aker, A. T. (2014). İki boyutlu yas ölçeğinin Türkçe formunun geçerlilik ve güvenirlilik çalışması [The reliability and validity studies of the Turkish version of Two Track Model of Bereavement Questionnaire]. Türk Psikiyatri Dergisi 223 AJESI - Anadolu Journal of Educational Sciences International, 2018; 8(2): 204-231 DOI: 10.18039/ajesi.454571 [Turkish Journal of Psychiatry], 25(4), 253-263. Retrieved from http://www.turkpsikiyatri.com/en/default.aspx?modul=article&id=958. Bischof, L. J. (1976). Adult psychology. New York: Harper & Row. Bolak-Boratav, H., Sunar, D., & Ataca, B. (2011). Duyguları sergileme kuralları ve bağlamsal belirleyicileri: Türkiye’de üniversite öğrencileri ile yapılan bir araştırma [Emotional display rules and their contextual determinants: An investigation with university students in Turkey]. Turkish Journal of Psychology, 26(68), 90-103. Retrieved from http://www.turkpsikolojiyazilari.com/PDF/TPD/68/06.pdf. Braitman, A. L., Derlega, V. J., Henson, J. M., Robinett, I., Saadeh, G. M., Janda, L. J., Hixon, M., & Miranda, J. (2008). Social constraints in talking about diabetes to significant others and diabetes self-care: A social-cognitive processing perspective. Journal of Social and Clinical Psychology, 27(9), 949-969. doi: 10.1521/jscp.2008.27.9.949. Browne, M. W., & Cudeck, R. (1993). Alternative ways of assessing model fit. In K. A. Bollen, & J. S. Long (Eds.) Testing Structural Equation Models (pp. 136-162). Beverly Hills, CA: Sage. Clark, L. F. (1993). Stress and the cognitive-conversational benefits of social interaction. Journal of Social and Clinical Psychology, 12(1), 25-55. doi: 10.1521/jscp.1993.12.1.25 Eton, D. T., Lepore, S. J., & Helgeson, V. S. (2001). Early quality of life in patients with localized prostate carcinoma: An examination of treatment-related, demographic, and psychosocial factors. Cancer, 92(6), 1451-1459. Retrieved from http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2610318/. 224 AJESI - Anadolu Journal of Educational Sciences International, 2018; 8(2): 204-231 DOI: 10.18039/ajesi.454571 Eton, D. T., Lepore, S. J., & Helgeson, V. S. (2005). Psychological distress in spouses of men treated for early-stage prostate carcinoma. Cancer, 103(11), 2412-2418. doi: 10.1002/cncr.21092. Farberow, N. L., Gallagher-Thompson, D., Gilewski, M., & Thompson, L. (1992). The role of social supports in the bereavement process of surviving spouses of suicide and natural deaths. Suicide and Life-Threatening Behavior, 22(1), 107-124. doi: 10.1111/j.1943278X.1992.tb00479.x. Fry, P. S. (2001). Predictors of health-related quality of life perspectives, self-esteem, and life satisfactions of older adults following spousal loss: An 18-month follow-up study of widows and widowers. The Gerontologist, 41(6), 787-798. doi: 10.1093/geront/41.6.787. Groff, E. C. (2014). The relationship between social constraints, social support, loss-related factors, depression, and PTSD in a treatment-seeking suicide bereaved sample (Unpublished Doctoral Dissertation). Retrieved from http://gradworks.umi.com/37/ 37/3737784. Html. Henson, J. M., Derlega, V. J., Pearson, M. R., Ferrer, R., & Holmes, K. (2013). African American students’ responses to racial discrimination: How race-based rejection sensitivity and social constraints are related to psychological reactions. Journal of Social and Clinical Psychology, 32(5), 504-529. doi: 10.1521/jscp.2013.32.5.504. Hu, L. T., & Bentler, T. M. (1999). Cut off criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6, 1-55. doi: 10.1080/10705519909540118. Huang, C. C., Wang, Y. M., Wu, T. W., & Wang, P. A. (2013). An empirical analysis of the antecedents and performance consequences of using the Moodle platform. 225 AJESI - Anadolu Journal of Educational Sciences International, 2018; 8(2): 204-231 DOI: 10.18039/ajesi.454571 International Journal of Information and Education Technology, 3(2), 217-221. doi: 10.7763/IJIET.2013.V3.267. Hunter, K. (2007). “Sleep on dear Ernie, your battles are o’er”: A glimpse of a mourning community, Invercargill, New Zealand, 1914-1925. War in History, 14(1), 36-62. doi: 10.1177/0968344507071040. International Test Commission. (2010). The ITC guidelines for translating and adapting tests. Retrieved from http://www.psyktestbarn.no/cms/ptb_mm.nsf/lupgraphics/ ITC%20guidelines.pdf/$file/ITC%20guidelines.pdf. Jöreskog, K. G., & Sörbom, D. (1996). LISREL 8 User's reference guide. Chicago: Scientific Software. Kline, R. B. (2011). Principles and practice of structural equation modeling. New York: The Guildford Press. Koutrouli, N., Anagnostopoulos, F., Tsikkinis, A., Papastylianou, D., & Lepore, S. (2015). Psychometric properties of the Greek version of the Social Constraints Scale in a sample of women with breast cancer. Women & Health, 23, 1-15. doi: 10.1080/03630242.2015.1101746. Lepore, S. J. (1992). Social conflict, social support, and psychological distress: Evidence of cross-domain buffering effects. Journal of Personality and Social Psychology, 63(5), 857867. doi: 10.1037/0022-3514.63.5.857. Lepore, S. J. (2001). A social-cognitive processing model of emotional adjustment to cancer. In A. Baum & B. Anderson (Eds.), Psychosocial Interventions for Cancer (pp. 99118). Washington, DC: American Psychological Association. Lepore, S. J. (2003, November 1). Social constraints construct & measurement. Retrieved from http://www.ibrarian.net/navon/paper/Social_Constraints_Construct___Measure ment. pdf?paperid=1814194. 226 AJESI - Anadolu Journal of Educational Sciences International, 2018; 8(2): 204-231 DOI: 10.18039/ajesi.454571 Lepore, S. J., Fernandez-Berrocal, P., Ragan, J., & Ramos, N. (2004). It’s not that bad: Social challenges to emotional disclosure enhance adjustment to stress. Anxiety, Stress, and Coping, 17(4), 341-361. doi: 10.1080/10615800412331318625. Lepore, S. J., & Ituarte, P. H. G. (1999). Optimism about cancer enhances mood by reducing negative social relations. Cancer Research Therapy and Control, 8(1-2), 165-174. Retrieved from http://astro.temple.edu/~slepore/Lepore_CRTC_Itu_1999. Lepore, S. J., Ragan, J. D., & Jones, S. (2000). Talking facilitates cognitive-emotional processes of adaptation to an acute stressor. Journal of Personality and Social Psychology, 78(3), 499-508. doi: 10.1037//0022-3514.78.3.499. Lepore, S. J., Silver, R. C., Wortman, C. B., & Wayment, H. A. (1996). Social constraints, intrusive thoughts, and depressive symptoms among bereaved mothers. Journal of Personality and Social Psychology, 70(2), 271-282. doi: 10.1037/0022-3514.70.2.271. Lewis, R. J., Derlega, V. J., Clarke, E. G., & Kuang, J. C. (2006). Stigma consciousness, social constraints, and lesbian well-being. Journal of Counseling Psychology, 53(1), 48-56. doi: 10.1037/0022-0167.53.1.48. Lindinger-Sternart, S. (2015). Help-seeking behaviors of men for mental health and the impact of diverse cultural backgrounds. International Journal of Social Science Studies, 3(1), 1-6. doi: 10.11114/ijsss.v3il.519. Lubben, J. E. (1988). Assessing social networks among elderly populations. Family & Community Health, 11(3), 42-52. doi: 10.1097/00003727-198811000-00008. MacCallum, R. C., Browne, M. W., & Sugawara, H. M. (1996). Power analysis and determination of sample size for covariance structure modeling. Psychological Methods, 1(2), 130-149. doi: 10.1037/1082-989X.1.2.130. Mackenzie, C. S., Scott, T., Mather, A., & Sareen, J. (2008). Older adults' help-seeking attitudes and treatment beliefs concerning mental health problems. The American 227 AJESI - Anadolu Journal of Educational Sciences International, 2018; 8(2): 204-231 DOI: 10.18039/ajesi.454571 Journal of Geriatric Psychiatry, 16(12), 1010-1019. doi: 10.1097/JGP.0b013e31818cd3be. McGloshen, T. H., & O’Bryant, S. L. (1988). The psychological well-being of older, recent widows. Psychology of Women Quarterly, 12, 99-116. doi: 10.1111/j.14716402.1988.tb00930.x. Mojaverian, T., Hashimoto, T., & Kim, H. S. (2013). Cultural differences in professional help seeking: A comparison of Japan and the US. Frontiers in Psychology, 3:615. doi: 10.3389/fpsyg.2012.00615. Nagel, J. K. (1988). Unresolved grief and mourning in Navajo women. American Indian and Alaska Native Mental Health Research, 2(2), 32-40. doi: 10.5820/aian.0202.1988.32. Nordanger, D. (2007). Discourses of loss and bereavement in Tigray, Ethiopia. Culture, Medicine and Psychiatry, 31, 173-194. doi: 10.1007/s11013-007-9050-6. Oliver, M. I., Pearson, N., Coe, N., & Gunnell, D. (2005) Help-seeking behavior in men and women with common mental health problems: Cross-sectional study. British Journal of Psychiatry, 186, 297-301. doi: 10.1192/bjp.186.4.297. Ono, Y., Araki, N., Mitani, M., & Yoshimura, K. (1999). Help-seeking behavior across different age and culture groups. In Y. Ono, A. Janca, M. Asai, & N. Sartorius (Eds.), Somatoform Disorders: A Worldwide Perspective (pp. 132-140). Tokyo: SpringerVerlag. doi: 10.1007/978-4-431-68500-5_13. Pennebaker, J. W., & Harber, K. D. (1993). A social stage model of collective coping: The Loma Prieta Earthquake and the Persian Gulf War. Journal of Social Issues, 49(4), 125-145. doi: 10.1111/j.1540-4560.1993.tb01184.x. Pressman, D. L., & Bonanno, G. A. (2007). With whom do we grieve? Social and cultural determinants of grief processing in the United States and China. Journal of Social and Personal Relationships, 24(5), 729-746. doi: 10.1177/0265407507081458. 228 AJESI - Anadolu Journal of Educational Sciences International, 2018; 8(2): 204-231 DOI: 10.18039/ajesi.454571 Riley, L. P., LaMontage, L. L., Hepworth, J. T., & Murphy, B. A. (2007). Parental grief responses and personal growth following the death of a child. Death Studies, 31, 277299. doi: 10.1080/07481180601152591. Rime, B., Mesquita, B., Philippot, P., & Boca, S. (1991). Beyond the emotional event: Six studies on the social sharing of emotion. Cognition and Emotion, 5(5/6), 435-465. doi: 10.1080/02699939108411052. Rubin, S. S., Bar-Nadav, O., Malkinson, R., Koren, D., Goffer-Shnarch, M., & Michaeli, E. (2009). The Two-Track Model of Bereavement Questionnaire (TTBQ): Development and validation of a relational measure. Death Studies, 33, 305-333. doi: 10.1080/07481180802705668. Sanders, C. M. (1988). Risk factors in bereavement outcome. Journal of Social Issues, 44(3), 97-111. doi: 10.1111/j.1540-4560.1988.tb02079.x. Santrock, J. W. (2011). Life-span development. New York, NY: McGraw-Hill. Schmidt, J. E. (2002). The role of social and dispositional variables associated with emotional processing in adjustment to breast cancer (Unpublished Master’s Thesis). Retrieved from http://uknowledge.uky.edu/gradschool_theses/386. Schumacker, R. E., & Lomax, R. G. (2010). A beginner s guide to structural equation modeling. New York, NY: Routledge. Schuster, T. L., Kessler, R. C., & Aseltine, R. H. (1990). Supportive interactions, negative interactions, and depressed mood. American Journal of Community Psychology, 18(3), 423-438. doi: 10.1007/BF00938116. Sheridan, M. A., Sherman, M. L., Pierce, T., & Compas, B. E. (2010). Social support, social constraints, and affect in spouses of women with breast cancer: The role of cognitive processing. Journal of Social and Personal Relationships, 27(1), 5-22. doi: 10.1177/0265407509347836. 229 AJESI - Anadolu Journal of Educational Sciences International, 2018; 8(2): 204-231 DOI: 10.18039/ajesi.454571 Sousa, V. D., & Rojjanasrirat, W. (2011). Traslation, adaptation and validation of instruments or scales for use in cross-cultural health care research: A clear and user-friendly guideline. Journal of Evaluation in Clinical Practice, 17, 268-274. doi: 10.1111/j.1365-2753.2010.01434.x. Stelle, C. D., & Uchida, M. (2004). The stability and change in the social support networks of widowers following spousal bereavement. The Journal of Men’s Studies, 13(1), 85105. doi: 10.3149/jms.1301.85. Stroebe, M. S., & Schut, H. (1999). The dual process model of coping with bereavement: Rationale and description. Death Studies, 23, 197-224. doi: 10.1080/074811899201046. Stroebe, W., Stroebe, M. S., & Abakoumkin, G. (1999). Does differential social support cause sex differences in bereavement outcome? Journal of Community and Applied Social Psychology, 9, 1-12. doi: 10.1002/(SICI)1099-1298(199901/02)9:1<1::AID- CASP478>3.0.CO;2-X. Tunay-Akan, Ş., & Barışkın, E. (2016). Kültür ve cinsiyet bağlamında Berkeley Duygu İfadesi Ölçeğinin geçerlilik ve güvenilirlik ölçütleri [Reliability and validity indicators of Berkeley Expressivity Questionnaire in the context of culture and gender]. Türk Psikiyatri Dergisi [Turkish Journal of Psychiatry], 27. doi: 10.5080/u13378. Ullrich, P. M., Lutgendorf, S. K., & Stapleton, J. T. (2002). Social constraints and depression in HIV infection: Effects of sexual orientation and area of residence. Journal of Social and Clinical Psychology, 21(1), 46-65. doi: 10.1521/jscp.21.1.46.22402. Vachon, M. L., & Stylianos, S. K. (1988). The role of social support in bereavement. Journal of Social Issues, 44(3), 175-190. doi: 10.1111/j.1540-4560.1988.tb02084.x. 230 AJESI - Anadolu Journal of Educational Sciences International, 2018; 8(2): 204-231 DOI: 10.18039/ajesi.454571 Vaswani, N. (2011). Encouraging help-seeking behavior among young men: A literature review. Report for the Child Protection Committee, Glasgow City Council. Retrieved from http://www.glasgow.gov.uk/chttphandler.ashx?id=5252. You, J., & Lu, Q. (2014). Social constraints and quality of life among Chinese-speaking breast cancer survivors: A mediation model. Quality of Life Research, 23(9), 1-8. doi: 10.1007/s11136-014-0698-4. 231