Common use of Sample Characteristics Clause in Contracts

Sample Characteristics. The majority of women reported low involvement in household decision making (N=1,796, 44.3%) with fewer women reporting medium involvement (N=1,257, 31.0%) and high involvement (N=999, 24.7%). The individual decision making indicators indicate that joint decision making between the respondent and the husband occurs most frequently followed by the husband’s exclusive decision making. Women’s exclusive decision making was in the majority only for daily household purchases. On average, women had 2.8 children (range 0-9). Relative to their community’s fertility- related gender norm, 1,593 (39.3%) women were categorized as normative. Similar numbers of women were classified as deviant below (30.7%) and deviant above (30.0%) their community norm. A majority of women had at least one living child (89.4%), and a majority of women had at least one living son (73.3%). Approximately one-quarter of respondents reported experiencing IPV within the year preceding the BDHS. Respondents’ mean age was 30.3 years. Respondents had completed 4.2 years of schooling on average, and the majority of women had less education than their spouse (N=2,951, 72.8%). Husbands were absent from approximately 9% of respondents’ homes (N=352). Women were married a mean 15 years. Approximately 60% were from rural settings, and about half (55.9%) maintained some form of economic involvement via employment or women’s development organizations. Additional respondent characteristics are available in Table 1. Results of the three mediation analyses are presented in Table 2. Odds ratios e|β coef.| (OR) are presented to facilitate interpretation of each predictors’ association with the dependent variable. Step 1 and Step 3 were modeled using generalized ordered logistic regression. In each step, a few variables violated the proportional odds assumption indicating that the effect of these variables was not constant across all levels of the outcome, women’s involvement in household decision making. Accordingly, these parameters were left unconstrained and two ORs are presented; The first odds ratio (with superscript ‘a’) predicts medium or high involvement in decision-making as opposed to low, and the second odds ratio (with superscript ‘b’) predicts high involvement compared to medium and low involvement. Each model is discussed step by step. In Model 1, Step 1, the quadratic individual fertility term was the only fertility measure with a significant association with women’s empowerment. Since the OR is less than one (0.977), it indicates women’s individual fertility has an inverse relationship, concave downward, with all levels of decision making. Although significant, an odds ratio so close to one indicates fertility has only a modest association with involvement in household decision making. None of the measures of conformity to fertility-related gender norms were significant. In Step 2, again only one fertility measure was significant: the linear measure of women’s individual fertility. For every one child increase in a woman’s fertility in the 1+ years before the BDHS, she was 1.215 times more likely to experience IPV in the past year. As in Step 1, none of the measures of conformity to fertility-related gender norms were significantly associated with IPV. Model 1, Step 3: Since the IPV measure violated the assumption of proportional odds, two ORs are presented, but the association is only significant for high decision making. Specifically, women who experienced IPV in the past year were 1.326 times more likely to achieve high levels of decision making. Also, in Step 3, the quadratic measure of individual fertility was found to be significantly associated with involvement in household decision making. Model 1, Step 4 indicated that IPV did not mediate the relationship between women’s fertility and their empowerment. Women’s fertility (quadratic term) was significantly associated with empowerment; however, after the addition of IPV to the regression model in Step 3, there was no change in the OR or statistical significance of women’s fertility (OR=0.977, p<=0.05). In addition, the nominal measure of women’s conformity to fertility-related gender norms was never found to be significantly associated with women’s empowerment. Thus, there was no association established for IPV to mediate between gender norm conformity and empowerment.5 Model 2, Step 1: having at least one living child was significantly associated with women’s empowerment. Specifically, women with at least one child were 43% more likely to have increased involvement in household decision making relative to childless women. In Step 2, having at least one child was not significantly associated with IPV. In Step 3, IPV violated the proportional odds assumption and two ORs are presented. Similar to Model 1, IPV is only significantly associated with high involvement in household decision making; women who reported experiencing IPV in the past year were 28% more likely to report high involvement in decision making. Also, in Step 3, having at least one child was significantly associated with women’s increased involvement in household decision making. Model 2, Step 4 indicated that IPV did not mediate the relationship between having at least one child and empowerment; there was no change in the OR or statistical significance of this aspect of women’s fertility (OR=1.423, p<=0.05). Model 3, Step 1: having at least one living son was not associated with women’s empowerment. Additionally, in Step 2, having at least one son was not associated with IPV. In Step 3, IPV again violated the proportional odds assumption and two ORs are presented. Similar to Models 1 and 2, IPV is only significantly associated with high involvement in household decision making. With IPV in the Step 3 regression model, having at least one son remains insignificant. Model 3, Step 4: 5 An additional mediation analysis was explored with contraceptive use. Similar to the nominal relative fertility variable, normative contraceptive use was established by community (categories were consistent non-user, consistent user, inconsistent user) using data from the BDHS contraceptive calendar. Women whose contraceptive behavior matched the norm in their community were considered to conform to the norm, and women who did not were categorized as deviant. In Step 1 of mediation analysis, controlling for fertility and demographic predictors, no association was found for IPV to mediate between relative contraceptive use and decision making. Results made available on request. Having at least one son was never found to be significantly associated with women’s empowerment. Since there was no association established between having at least one son and empowerment for IPV to mediate, there is no evidence of mediation.

Sample Characteristics. The majority of women reported low involvement in household decision making (N=1,796N=4,749, 44.346.84%) with fewer ). Smaller percentages of women reporting reported medium involvement (N=1,257, 31.0%) and high involvement (N=99929.83% and 23.34%, 24.7%respectively). The individual An investigation of each of the four decision making indicators indicate questions individually indicates that joint decision making occurs most often between the respondent husband and the husband occurs most frequently respondent jointly followed by the husband’s husband alone. Women reported exclusive decision making. Women’s making in the majority for daily household purchases only; exclusive decision making was minimal in the majority only for daily household purchasesother three decision making indicators. On average, women Women had 2.8 an average of 2.43 living children (range 0range=0-912). Relative to their community’s fertility- related gender norm, 1,593 (39.3%) The mean fertility across communities ranged from 1.21-4.40. Slightly more than a third of women were categorized as normativenormative relative to their community mean fertility (N=3,687, 36.34%). Similar numbers Another third were categorized as having fertility deviant above their community mean fertility; 20.48% were deviant above and 8.95% were extremely deviant above the mean. The remaining 29.64% of women were classified categorized as deviant below (30.7%) the mean, and 4.59% were extremely deviant above (30.0%) below their community normmean. A majority of women had at least one living child (89.4%), and a majority of women had at least one living son (73.3%). Approximately one-quarter of respondents reported experiencing IPV within the year preceding the BDHS. Respondents’ Participant mean age was 30.3 years29.96 (median=29, l.s.e.=0.113). Respondents Women had completed 4.2 an average of 4.38 years of schooling on averageschool (range=0.17, l.s.e.=0.091), and the majority of women had less education than their spouse (N=2,951N=7,188, 72.870.85%). Husbands were absent A small percentage of women reported that their husband was living away from approximately 9% of respondents’ homes home (N=352N=1,123, 11.07%). Women The median length of marriage was 13 years (range=0-39.5 years). Most women were married currently living in a mean 15 yearsrural setting (N=6,342, 62.51%). Approximately 60% were from rural settingsa third of participants reported employment in the last twelve months (N=3,065, 30.22%), and about half involvement in a women’s organization (55.9N=3,739, 36.83%) maintained some form of economic involvement via employment or women’s development organizationssuch as a microcredit program. Additional respondent characteristics are available presented in Table 1. Results of from the three mediation analyses partial proportional odds model are presented provided in Table 2. Odds ratios (e|β coef.| (ORcoef.|) rather than coefficients are presented to facilitate interpretation of each predictors’ association with independent variable’s impact on decision making. The proportional odds assumption was violated for both the linear and quadratic fertility term, community mean fertility, and the control variable husband away from home. When the proportional odds assumption is not met, the effect of a predictor is not constant across all transitions of the dependent variable. Step 1 and Step 3 were modeled using generalized ordered logistic regression. In each step, a few variables violated the proportional odds assumption indicating that the effect of Parameters for these variables was not constant predictors are left unconstrained to vary across all levels of the outcomedependent variable, women’s involvement in household decision making. Accordinglyand therefore, these parameters were left unconstrained and two ORs odds ratios are presented; . The first odds ratio (with superscript ‘a’) predicts medium or high involvement in decision-making as opposed to low, and the second odds ratio (with superscript ‘b’) predicts high involvement compared to medium and low involvement. Each model is discussed step by step. In Model 1, Step 1, Both the linear and quadratic individual fertility terms were significantly associated with decision making. Odds ratios (ORs) were greater than one for the linear term was the only fertility measure with a significant association with women’s empowerment. Since the OR is (OR=a1.133, b1.234) and less than one for the quadratic term (0.977)OR=a0.980, it indicates women’s b0.965) in both comparisons indicating a predominately positive concave downward relationship between individual fertility has an inverse relationship, concave downward, with and all levels of decision making. Although significantThe odds of transitioning from low to at least a medium level of involvement in decision making were less extreme than those of transitioning to high level of involvement. Therefore, an as individual fertility increases, the probability of having at least medium level of decision making involvement increases but eventually declines. The same trend is true for having high opposed to low or medium level of decision making involvement, but the increase and the subsequent decline are more dramatic. Community mean fertility also significantly predicted at least a medium level of decision making involvement. Specifically, a one unit increase in community mean fertility corresponded to a 22% decrease (OR=0.782) in the odds ratio of having at least a medium level of involvement in decision making. Community mean fertility did not significantly predict a woman’s high level of involvement in decision making. Fertility relative to the norm did not violate the proportional odds assumption, so close this parameter was constrained. Women deviant above and extremely above their community’s fertility were approximately 30% less likely to one indicates have increased decision making involvement relative to women with fertility has only a modest association normative to their community (OR=0.715, 0.699 respectively). Decision making was not significantly different between fertility normative women compared to women with fertility deviant below their community’s mean. A number of socio-demographic variables were significantly associated with women’s involvement in household decision making. None The odds of increased involvement increased minimally with age (OR=1.022) and respondents’ years of education (OR=1.039). Higher odds were noted for women of longer marital duration relative to those of shorter duration (OR=1.287), women with membership in a women’s organization such as Proshika or Grameen Bank (OR=1.175), and for women employed within the measures of conformity to fertility-related gender norms were significantprevious twelve months (OR=1.275). In Step 2, again only one fertility measure was significant: the linear measure The strongest predictor of women’s individual fertility. For every one child increase increased involvement in a woman’s fertility decision making was absence of the husband in the 1+ years before the BDHShome. Relative to women with husbands at home, she was 1.215 women with absent husbands were 2.451 times more likely to experience IPV in the past year. As in Step 1, none have at least a medium level of the measures of conformity to fertility-related gender norms were significantly associated with IPV. Model 1, Step 3: Since the IPV measure violated the assumption of proportional odds, two ORs are presented, but the association is only significant for high decision making. Specifically, women who experienced IPV in the past year were 1.326 making and 5.698 times more likely to achieve have high levels level of involvement in decision making. AlsoWomen with more years of education than their husband, in Step 3relative to those with less, the quadratic measure of individual fertility was found to be significantly associated with involvement in household decision making. Model 1, Step 4 indicated that IPV did not mediate the relationship between women’s fertility and their empowerment. Women’s fertility (quadratic term) was significantly associated with empowerment; however, after the addition of IPV to the regression model in Step 3, there was no change in the OR or statistical significance of women’s fertility (OR=0.977, p<=0.05). In addition, the nominal measure of women’s conformity to fertility-related gender norms was never found to be significantly associated with women’s empowerment. Thus, there was no association established for IPV to mediate between gender norm conformity and empowerment.5 Model 2, Step 1: having at least one living child was significantly associated with women’s empowerment. Specifically, women with at least one child were 4322% more less likely to have increased involvement in household decision making relative making. Women living in a rural setting were 27% less likely to childless women. In Step 2have increased involvement in decision making.1, having at least one child was not significantly associated with IPV. In Step 3, IPV violated 2 Based on the partial proportional odds assumption and two ORs are presented. Similar to Model 1model estimates, IPV is only significantly associated with high predicted probabilities were generated for each level of involvement in household decision making; making by individual fertility. Estimates are presented in Figure 1 for women who reported experiencing IPV extremely deviant above and normative to their community’s mean fertility in three community settings of low, medium, and high fertility levels. Several trends may be noted from the past year were 28% more likely to report high predicted probability graphs in Figure 1. Individual fertility maintains a quadratic relationship with decision making as evidenced by the concave lines in each graph. The probability of low involvement in decision makingmaking is highest at fertility extremes: when women have no children or high numbers of children. AlsoCorrespondingly, in Step 3, having at least one child was significantly associated with women’s increased the probability of medium levels of involvement in household decision making. Model 2, Step 4 indicated that IPV did not mediate making is the relationship between having at least one child and empowermentopposite; there was no change in the OR probability is lowest when women have low or statistical significance of this aspect of women’s fertility (OR=1.423, p<=0.05). Model 3, Step 1: having at least one living son was not associated with women’s empowerment. Additionally, in Step 2, having at least one son was not associated with IPV. In Step 3, IPV again violated the proportional odds assumption and two ORs are presented. Similar to Models 1 and 2, IPV is only significantly associated with high involvement in household decision making. With IPV in the Step 3 regression model, having at least one son remains insignificant. Model 3, Step 4: 5 An additional mediation analysis was explored with contraceptive use. Similar to the nominal relative fertility variable, normative contraceptive use was established by community (categories were consistent non-user, consistent user, inconsistent user) using data from the BDHS contraceptive calendar. Women whose contraceptive behavior matched the norm in their community were considered to conform to the norm, and women who did not were categorized as deviant. In Step 1 of mediation analysis, controlling for fertility and demographic predictors, no association was found for IPV to mediate between relative contraceptive use and decision makingis highest at a mid-level of fertility. Results made available on request. Having at least one son was never found to be significantly associated with women’s empowerment. Since there was no association established between The probability of women having at least one son and empowerment for IPV to mediate, there is no evidence of mediation.high