Page 135 - GHES-3-3

Global Health Economics and
Sustainability
Polyvictimization and suicidal behaviors

this study, the term polyvictimization refers to college and has shown good reliability and validity in previous
students who have had adverse childhood experiences studies (Osman et al., 2014; Zimet et al., 1990).
(e.g., physical abuse/neglect, emotional abuse, contact The covariates for this study included age, gender (i.e.,
sexual abuse, or exposure to household violence) and male, female, and no preference), and racial demographics
subsequently experienced peer victimization during (i.e., White, African American, Hispanic, and others
the past school year. To construct the polyvictimization [Asian, American Indian, Eskimo, or mixed races]).
variable, college students who had adverse childhood
experiences were first selected; then, those who had 2.3. Statistical analyses
experienced bullying victimization among them were Descriptive statistics and bivariate correlation analyses
identified. Sample questions from the ACE-IQ reflecting were conducted using Statistical Package for the Social
st
experiences during the 1 18 years of life included the Sciences 29.0 to explore the demographic characteristics
following: “How often did your parents/guardians not give and relationships of all variables. Models 4 and 5 from
you enough food even when they could easily have done the PROCESS Macro 4.2 (Hayes, 2022) were employed to
so?” and “Did someone make you touch their body in a test the mediation and moderation effects. First, Model
sexual way when you did not want them to, many times?” 4 was used to assess the indirect effect of resilience on
Regarding peer victimization, the questionnaire items the association between polyvictimization and suicidal
included the following: “Punched me” and “Made other thoughts/behaviors. Second, Model 5 was employed to
people not talk to me.” The ACE has 10 categories, each analyze whether perceived social support moderated
with two levels. The ACE score is calculated as the sum the association between polyvictimization and suicidal
of these 10 categories, with a total possible score of 10. To thoughts/behaviors. A bootstrap analysis with 5,000
establish polyvictimization, the response options for ACE repetitions determined the significance of the mediation
scores and bullying victimization were coded as binary and moderation effects, presented with a 95% confidence
responses of No (0) and Yes (1). The Cronbach’s alpha for interval (Hayes & Scharkow, 2013). In addition, for
this scale in the current sample was 0.89. visually recognizable moderating effects and accessible
interpretation, a simple slope analysis was conducted.
2.2.3. Resilience
Resilience, as a mediator, was assessed using the Connor- 3. Results
Davidson Resilience Scale (CD-RISC-25; Connor & Table 1 displays the descriptive statistics (i.e., numbers
Davidson, 2003), which has 10 items. Sample items [%], means, and standard deviations [SD]) and correlation
included the following: “I am able to adapt when changes coefficients among the study variables. The mean scores for
occur” and “I am able to handle unpleasant or painful polyvictimization, suicidal thoughts/behaviors, resilience,
feelings, such as sadness, fear, and anger.” The response and perceived social support were 0.26 (SD = 0.44), 3.81
options for these items were “Extremely unlikely” (1), (SD = 7.56), 36.24 (SD = 7.37), and 62.78 (SD = 16.42),
“Somewhat unlikely” (2), “Neither likely nor unlikely” respectively. Suicidal thoughts/behaviors were positively
(3), and “Somewhat likely” (4). The Cronbach’s alpha for correlated with polyvictimization (r = 0.44, p < 0.001) but
this scale in the current sample was 0.90. The scale has negatively correlated with age (r = −0.15, p < 0.01), White
demonstrated good reliability and validity in previous demographic (r = −0.19, p < 0.001), male gender (r = −0.20,
studies (Cronbach’s α = 0.85 – 0.89; Campbell‐Sills & Stein, p < 0.001), and female gender (r = −0.18, p < 0.001).
2007; Connor & Davidson, 2003; Gonzalez et al., 2016). Nonpolyvictimization was correlated with age (r = −0.09,

2.2.4. Perceived social support p < 0.01), White demographic (r = −0.13, p < 0.001), female
gender (r = −0.15, p < 0.001), and no preferred gender (i.e.,
Perceived social support, as a moderator, was assessed using the sum of nonbinary/third gender and prefer not to say;
the Multidimensional Scale of Perceived Social Support r = −0.09, p < 0.05). Resilience was positively correlated
(Zimet et al., 1988), which consists of 12 items related to with African American demographic (r = 0.13, p < 0.001),
support from family, friends, and special persons. Sample Hispanic demographic (r = 0.08, p < 0.05), male gender
items included the following: “There is a special person (r = 0.12, p < 0.01), and nonpolyvictimization (r = −0.10,
who is around when I am in need,” “My friends really try p < 0.01) but negatively correlated with others (i.e., the sum
to help me,” and “My family is willing to help me make of Asian, Native, Eskimo, or mixed races; r = −0.08, p < 0.05)
decisions.” Response options ranged from “Very strongly and suicidal thoughts/behaviors (r = −0.29, p < 0.001).
disagree” (1) to “Very strongly agree” (7). The Cronbach’s Perceived social support was positively correlated with
alpha for this scale in the current sample was 0.94. This resilience (r = 0.24, p < 0.001), nonpolyvictimization
scale is a widely used measure of perceived social support (r = −0.22, p < 0.001), White demographic (r = 0.19,

Volume 3 Issue 3 (2025) 127 https://doi.org/10.36922/ghes.3737