Computational protocol: Neglect and perceived stigmatization impact psychological distress of orphans in Tanzania

Similar protocols

Protocol publication

[…] To test hypotheses 1) and 2), we conducted two MANOVAs comparing orphans and non-orphans regarding their exposure to maltreatment and their psychological distress. Univariate analyses of variance (ANOVA) were then calculated. In order to avoid α-error inflation due to multiple testing, the Bonferroni-correction was applied. Following West, Finch and Curran (), all outcome variables were normally distributed except the UCLA score and neglect types for non-orphans. Violation of homogeneity was revealed by Levene's test between groups for CDI score, UCLA score and neglect types, and Box-M-Test for variance–covariance matrices in both MANOVAs. We used Pillai–Bartlett's trace for the MANOVAs since it is robust against violation of assumptions when sample sizes are equal (Bray & Maxwell, ). All analyses used a two-tailed α=0.05. For (M)ANOVAs, our metric for a small effect was η 2≥0.01, for a medium effect η 2≥0.06, and for a large effect η 2≥0.14.To test hypotheses 3–5, we conducted multiple sequential regression analyses correlating neglect, abuse, perceived stigmatization, and the interaction of neglect and stigmatization both on the CDI score and the SDQ Total Difficulties Score. Correlates were z-standardized before analysis and significant interactions were analyzed via simple slopes according to Aiken and West (). Data sets were excluded listwise for data missing in the stigmatization score (n=10). All assumptions for linear regression models were tenable. Variance inflation factor (VIF) did not exceed 1.16, indicating no risk of multicollinearity. No outliers were detected. Residuals were normally distributed. Linearity, homoscedasticity, and independence were tenable. For multiple regression analyses, our metric for a small effect was f 2≥0.02, for a medium effect f 2≥0.15 and for a large effect f 2≥0.35. We used IBM SPSS Statistics Version 21 for data analyses. Using G*Power (Erdfelder, ), we calculated the power (1–β) of our analyses post-hoc. […]

Pipeline specifications

Software tools SPSS, G*Power
Application Miscellaneous
Organisms Human immunodeficiency virus 2, Homo sapiens
Diseases Acquired Immunodeficiency Syndrome