Bivariate Associations. In bivariate analyses, women with more or less relative education than her partner were at greater risk of prior year IPV compared to women with equal amounts of schooling. Greater women’s level of schooling, greater partner’s level of schooling, age, formal marriage (vs. cohabitating and previously partnered), and a higher wealth index were protective of IPV risk. Witnessing parental IPV, working outside the home, and greater numbers of children living at home were associated with increased risk in IPV. Place of residence was not significantly associated with prior year IPV. The percent of prior year IPV reported by women showed a modest U-shaped pattern between women’s schooling relative to her partner and prior year IPV with 23.4% of relatively advantaged women and 22.3% of relatively disadvantaged women reporting IPV compared with 19.9% of women with equal schooling attainment (Table 2.3). Otherwise, compared to their counterparts, women who witnessed a father-to- mother beating (28.3% versus 18.8%) more often reported prior year IPV. Previously partnered women reported prior year IPV (36.7%) more often than married women (13.4%) and women living with a partner (22.0%). Women who were working or had worked in the past year more often reported IPV than unemployed women (24.3% versus 18.0%). Multivariate Results A multivariate logistic regression model was run to examine the influence of relative schooling attainment on a woman’s risk for prior year IPV (Table 2.4). Net of all other factors, the log odds of a woman experiencing prior year IPV was 23% higher for women with more grades of schooling than her partner. The negative gradient in prior year IPV with women’s own grades of schooling remained, as did witnessing her father beat her mother. Age was associated with a 3% decrease in log odds of prior year IPV for every year of age. The log odds of prior year IPV were 11% higher for each additional child in the home. Previously partnered women had 172% higher odds of prior year IPV and cohabitating women had 41% higher log odds compared to married women. Women who had worked in the past year had 32% higher log odds of IPV than women who were not working. Household wealth did not remain significantly associated with prior year IPV. Predicted Probabilities of Relative Education and IPV
Bivariate Associations. Women were more likely to experience prior year IPV if the respondent had greater numbers of children at home (crude OR = 1.03, 95% CI = 1.01, 1.06 ), was cohabitating (crude OR = 1.85, 95% CI = 1.71, 1.82) or had previously lived with someone (crude OR = 3.84, 95% CI = 3.52, 4.20), had witnessed her father beat her mother (crude OR = 1.71, 95% CI = 1.60, 1.82), had worked in the prior year (crude OR = 1.45, 95% CI = 1.36, 1.56), and had greater years of schooling than her partner (crude OR = 1.14, 95% CI = 1.07, 1.21). Women were less likely to experience prior year IPV if they were older (crude OR = 0.97, 95% CI = 0.97, 0.98) or had more years of schooling (crude OR = 0.97, 95% CI = 0.96, 0.98).
Bivariate Associations. This project examined the relationship between keratometric astigmatism (in diopters, D) and various demographic and surgical factors. Factors examined included: the treatment applied (contact lens vs. IOL); age at surgery; gender; incision type (clear cornea vs. scleral tunnel); the location of the incision on the eye, given as a “clock hours” location; presence or absence of an extended keratome; the number of sutures needed to close the incision; and suture type (interrupted vs. running). Finally, the differences in keratometric astigmatism between cataractous and fellow eyes was examined. Two sample t tests were performed to compare mean keratometric astigmatism values between various binary factor groups. Means of these factors were compared at both baseline, defined as date of surgery for cataract removal, and the visit at 1 year of age. The relationship between the change in mean keratometric astigmatism between these two time points, defined as the 1 year value minus the baseline value for each individual patient, and various factors was also examined. Some factors were considered as binary for the purposes of this analysis: number of sutures (1-2 sutures vs. 3-4 sutures), and age at surgery (less than 49 days vs. at least 49 days). The cutoff of 49 days was used since this was the age range defined for the stratification of the randomization of treatment. In comparing keratometric astigmatism in treated versus fellow eyes, a paired t-test was used. To further investigate the effects of incision location, a non-binary surgical factor, an analysis of variance with Tukey’s multiple comparisons testing was used. While possible values were given in “clock hours,” giving possible values {1, 1.5, 2, ... , 12.5}, only six values were used: 10; 10.5; 11; 11.5; 12; and 12.5. 21 comparisons were therefore made, with an alpha in an omnibus test of α = 0.05. These comparisons were done at baseline, 1 year, and the change between the two time points.
Bivariate Associations. Age (taken as an interval scale) was associated with number of transitions. Older participants reported to have experienced a larger number of transitions during the Holocaust, and more often lost one or both of their parents; not unexpectedly, age was also related to the length of the persecution period: older participants reported to have experienced a significantly longer period of persecution. On most post-traumatic stress indices we found no association with age, except for a small but significant association with the PDS subscale for re- experiencing. Older participants reported to somewhat more often re-experience their past traumas. SOC was not related to age (see Table 2). Participants with more physical illnesses also reported more post-traumatic stress and displayed a lower SOC. Participants who reported more parental losses and more transitions displayed more traumatic stress symptoms. Lastly, participants with a higher SOC reported fewer post-traumatic stress symptoms in all domains of the PDS (see Table 2). Multivariate analyses We conducted a hierarchical multiple regression analysis predicting current post-traumatic stress (PDS total) from physical illnesses (first step), loss of parents during the war, number of transitions, and period of persecution (second step). The beta weights for the separate predictors in both steps are based on the final regression equation including all predictors (see Table 3). The number of physical illnesses significantly contributed to the prediction of post-traumatic stress, but controlling for physical illnesses (beta .28, p <.01), the number of transitions (beta .16, p = .03), and the loss of parents (beta .15, p = .03) also significantly predicted post-traumatic stress symptoms. More physical illnesses, more transitions, and higher parental losses predicted more post-traumatic stress, see Table 3. Similar results were found for the PDS domain of arousal. For the PDS domains of re- experience and avoidance only physical illnesses appeared to be a significant predictor (see Table 3). In order to examine the role of SOC in predicting post-traumatic stress, we repeated the hierarchical multiple repression predicting current post-traumatic stress (PDS total) from physical illnesses (first step), SOC (second step), and loss of parents during the war, number of transitions, and period of persecution (third step). XXX contributed significantly to the regression (beta -.36, p < .01). Greater SOC predicted fewer pos...
