Analysis
Descriptive statistics were conducted on all measures with frequencies used to examine the most frequent contact, pathway chain, and the frequencies of each contact along the pathway. All variables were assessed for normality, linearity and homoscedasticity. Overall delay in treatment and number of contacts were significantly skewed, and non-parametric statistics were employed. Spearman’s correlations were conducted to examine the relationship between the number of pathway contacts and overall delay in weeks.
Additionally, three multiple regression analyses were conducted in SPSS (IBM Corp., 2022) one for each dependent variable: 1) onset duration; 2) treatment delay duration; and 3) help-seeking delay. For all regressions the following variables were included: caregiver familiarity (Caregiver MH treatment and Caregiver suicidality), the age of young person, young person’s gender and caregiver gender. Next moderation was conducted for each moderator separately (age and gender) for each of the dependent variables using PROCESS for SPSS (Hayes, 2013).
Prior to interpreting the results of the regression, several assumptions were evaluated, and results indicated that normality, linearity, homoscedasticity of residuals, multicollinearity and assessment for multivariate outliers supported assumptions of the analyses. Descriptive analysis (Table 2) revealed outliers. Therefore, a sensitivity analysis was conducted to determine whether results were affected by including, excluding or adjusting the highest time durations (outlier) in the regression findings. The results showed similar patterns of predictors and statistical significance irrespective of inclusion or exclusion of outliers for the time duration variables. Thus, we report the results of the regression with all duration scores as reported and unchanged (i.e., including outliers).