1. Does Expected Social Support Buffer Negative Adjustment for Victims of Domestic Violence? Nino Javakhishvili Maka Lortkipanidze Ilia State University December 9, 2015

2. Victims of domestic violence Have been reported to have poor adjustment: PTSD, depression, low self-esteem (Varkovitzky, 2004; Johnson &Ferraro, 2000; Berg, 2004; Mechanic, Weaver & Resick, 2008;). Adjustment is usually measured as an outcome variable Depression, anxiety - negative Life satisfaction – an added value

3. Domestic Violence as an independent variable Is a widespread problem, in Georgia as well 2010 data on prevalence Measured different forms – physical, economical, sexual, psychological 2 forms of psychological violence – added value

4. The current study – forms of violence Psychological violence: Dominance-isolation: controlling victim’s behaviors, isolating from friends and family members, demands for subservience, and rigid observances of traditional sex roles. (Tolman, 1989). Emotional-verbal violence: behaviors related to verbal attacks, attempts to demean the partner, and withholding emotional resources. (Tolman, 1989).

5. Organizational versus Family support The study that considers organizational social support was conducted in Bangladesh with the qualitative results that also support buffer hypothesis (Basu, 2008). Support was provided to victims by the victim support group – self-esteem and interpersonal trust that caused stress coping strategies, such as positive reassessment of stressful situation, planning how to overcome stress, even escape from the perpetrator

6. The buffering hypothesis The buffering hypothesis states that social support plays a buffer role for those who experience stress, trauma or other related problems (Bolger & Amarel, 2007; Uchino, 2009)

7. The buffering hypothesis The buffering hypothesis was checked on victims of domestic violence with findings that social support provides an important resource to decrease, or minimize stress, improve general health to adjust to the violent situation (El-Bassel et al., 2001, Waldrop&Resick, 2004, Panaghi et al.,2012)

8. The buffering hypothesis Social support moderated effects of spouse abuse on health outcomes in a general health questionnaire Social support can decrease the negative physical and psychological influences of spouse abuse on health condition (Panaghi et al.,2012)

9. Proposed model

10. Method - sample 256 women victims of domestic violence were studied From the capital city of Tbilisi - 29.8% from regional towns and villages - 69.3% Age: 19-67, M 36, SD 9.664 69.7% Living with partner/spouse 89.0% have children 40.4% have higher education 65.1% employed – seasonal, or part time

11. Method - sample Personal income – 62.2% Financially dependent on husband – 26.9% Financially dependent on parents – 9.20% Personal income – 62.2%, and only for Enough to survive – 5.9% Snowball principle was used

12. Method - tools 1.Psychological Maltreatment of Women Inventory, Tolman, 1995 – modified version 2.Center for Epidemiologic Studies Depression Scale (CESD) Laurie Radloff, 1977 – validated for this study 3.State-Trait Anxiety Inventory, Spielberger et al. 1983. We used only state anxiety subscale – validated for Georgian sample by Arutunov and Skhirtladze 4.The Satisfaction with Life scale, Pavot & Diener, 1993 – validated for Georgian sample by Skhirtladze 5.Expected Social Support Questionnaire, Sarason et al. 1983 – modified version

13. Method - tools Psychological Maltreatment of Women Inventory, Tolman, 1995 Added 12 questions on physical, economic and sexual violence from World Vision questionnaire on gender violence. Expected Social Support Questionnaire, Sarason et al. 1983 As the questionnaire studies general social support, we took out 5 questions not connected to DV cases, and left only 22 questions

14. Method – tools - CFA CFA of the Center for Epidemiologic Studies Depression Scale (CESD) resulted in a 9 item scale all loading on the depression factor Chi-Square 47.357 (P < 0.0091) CFI 0.911 TLI 0.881 RMSEA 0.055 We had to take out 11 items that did not load on the factor EFA - 11 items loaded on three other factors that did not make any sense

15. Method – tools - CFA CFA of the state Anxiety Scale (Spilberger) resulted in a 18 item scale all loading on the anxiety factor Chi-Square 212.341 (P < 0.0000) CFI 0.941 TLI 0.907 RMSEA 0.068 We had to take out 2 items that did not load on the factor

16. Results Central tendency and distribution for: Forms of violence Outcomes Social support and satisfaction with SS Correlations Regressions Moderation analysis

18. Results - Frequencies 5 forms together – 66% Both forms of psychological violence are present in almost every combination Sexual violence is the least present in combinations

19. Results - violence means Verbal-emotional - 15.92 (Std 3.177) min - 3, max - 18 Dominance-isolation - 34.98 (Std 9.923) min - 8, max - 48 sexual - 8.22 (Std 3.540) min - 2, max -12 economic - 13.98 (Std 4.550) min - 4, max -24 physical - 20.50 (Std 10.072) min - 6, max - 42

21. Results – outcome variables High level of anxiety - 42.16 (Std 4.575) High level of depression - 52.17 (Std 9.609) Low level of life satisfaction - 10.31 (Std 5.197) The data is expected and found in the literature(Varkovitzky, 2004; Ham-Rowbottom, et al. 2005; Liu, Dore & Amrani-Cohen, 2013; Beck, et al. 2014)

22. 253 yes, 3 - no Expected Social Support - Family

23. 74- yes, 182 - no Expected Social Support - Institutions

25. Results - correlations Total social support correlates with both, family (.953 p<.01) and institutional (.364 p<.01) support. Total social support correlates with depression (-.146 p<.05) and life satisfaction (.321 p<.01). No correlation was found with anxiety Family support correlates with depression (-.164 p<.01) and life satisfaction (.303 p<.01). No correlation was found with anxiety Institutional support correlates with anxiety (-.170 p<.01) and life satisfaction (.125 p<.05). No correlation was found with depression

26. Results - correlations Total DV correlates with anxiety (.269 p<.01) and life satisfaction (-.430 p<.01). No correlation was found with depression Physical violence correlates with anxiety (.196 p<.01) and life satisfaction (-.327 p<.01). No correlation was found with depression Economic violence correlates with anxiety (.246 p<.01) and life satisfaction (-.306 p<.01). No correlation was found with depression Emotional-verbal violence correlates with anxiety (.196 p<.01), depression (.124 p<.05) and life satisfaction (-.465 p<.01) Isolation-domination correlates with anxiety (.136 p<.05), depression (.166 p<.01) and life satisfaction (-.313 p<.01) Sexual violence correlates with life satisfaction (-.202 p<.01). No correlation was found with depression and anxiety

27. Discussion expected family support is high expected institutional support is especially low Much less people hope for institutions than for family The correlations show that total SS and family SS behave very much alike Institutional and Family support are both linked with life satisfaction Institutional support is linked with anxiety, while family support is linked with depression Two forms of psychological violence correlate with all outcome variables Physical and economic violence correlates with anxiety and life satisfaction

28. Results – forms of violence controlling for income Physical violence (beta = 0.185, sig. 0.038) and isolation-domination (beta = 0.300, sig. 0.001) are significant predictors in case of depression, despite income Economic violence (beta = 0.260, sig. 0.000) is significant predictor in case of anxiety, despite income Emotional-verbal violence (beta = 0.287, sig. 0.000), physical violence (beta = 0.219, sig. 0.001) and isolation-domination (beta = 0.157, sig. 0.024) are the predictors in case of life satisfaction

29. Discussion Physical and psychological violence are reported to predict depression (Varkovitzky, 2004). This is in line with our data. In addition, our data enables us to narrow down psychological violence to one of its forms – isolation- domination Sexual violence does not predict any variable, probably because of its distribution We could not find data on relations between different types of violence and anxiety or life satisfaction

30. Discussion Anxiety is predicted by economical violence - – why? life satisfaction is predicted by physical violence and both forms of psychological violence – why?

31. Results – forms of social support Social support total (beta = -0.178, sig. 0.008) is significant predictor in case of depression Family support (beta = -0.164, sig. 0.008) is significant predictor in case of depression and (beta = 0.321, sig. 0.000) in case of life satisfaction Institutional support (beta = -0.170, sig. 0.006) is the predictor in case of anxiety

32. Discussion To consider the results in combination with violence regressed on outcomes: Institutional support helps in case of economical violence and reduces anxiety While family support helps in case of physical and psychological violence and reduces depression and increases life satisfaction

33. Moderation analysis was used to test the buffering effect of social support. No effect was found on life satisfaction The effect was found (beta = -0.157 sig. 0.014) on depression. Family support plays the role of moderator The effect was found (beta = -0.405 sig. 0.001) on anxiety. Family support plays the role of moderator Results – buffering hypothesis

34. Discussion – the buffering hypothesis The buffering hypothesis is supported partially The low quality of institutional social support, might be the case (Javkhishvili, et al., 2010, Javakhishvili, Lordkipanidze, Petriashvili, 2011) Correlation analysis shows that total social support and family support vary similarly, while institutional support - does not (institutional support does not correlate with outcome variables)

35. Thank you for your attention! Questions? Comments?