R version 2

Analyses were performed using R version 2.15.1 and SPSS 21 with an alpha level of 5%. The alpha level was set at 0.1% for multiple group comparisons using Bonferroni adjustment (see Criterion-referenced concurrent validity section). The factor structure as well as the internal consistency and item-total correlation were assessed in all study samples. The test–retest was used to assess the reliability of the scale in the whole study sample 1. Furthermore, predictive validity was investigated in a symptom provocation paradigm (study sample 1; see Predictive validity section), convergent validity was examined between the Shut-D and the DES (study samples 2 and 4), criterion-referenced concurrent validity was obtained by comparisons of different diagnostic groups (study samples 1–4) and point-biserial correlates with the symptom spectrums depression and PTSD symptom severity (study sample 1). To avoid global correlations, the associations were assessed in the patient sample.

Analyses were performed using R version 2.15.1 and SPSS 20.0. Per hypothesis (1) the impact of ACE multiplicity (or the dose-dependent effect) on symptom severities was examined for each symptom dimension (PTSD, shutdown dissociation and depression) by univariate analyses of variance (ANOVA) with the between-subjects factor multiplicity, comparing subgroups according to 5 different levels of multiplicity (0 = no childhood adversities (n = 13), 1 = one type of MACE MULTI (n = 29), 2 = two types of MACE MULTI (n = 20), 3 = three types of MACE MULTI (n = 18) and 4 = four to ten types of MACE MULTI (n = 49; see [12 (link)], for a similar approach). In addition, the relationship between the symptom severities and global measures of the MACE duration, multiplicity (MACE MULTI), overall severity (MACE SUM), and the trauma load in childhood and adulthood were evaluated by Pearson correlations. Similarly, correlations determined the relationship between symptom severity and the severity score of each type of ACE. In order to correct for multiple testing the Benjamini-Hochberg procedure was applied [40 ], resulting in a corrected significance threshold of α = 4.3 % for hypothesis 1 and α = 3.6 % for the hypothesis 2.

Analyses were performed using R version 2.15.1 and SPSS 20.0; alpha level was set at .05. To compare the demographic and clinical data between the PTSD and Non-PTSD group, t-tests were used for continuous dependent variables and χ² tests were applied for nominal variables (using SPSS). Correlations and t-tests of the minimum-norm data were calculated with R. Rosenthal’s r was calculated as effect size [35 ].