188 SCIENTIFIC ARTICLE Table 2. Mean group ratings for the group processes Workshop Phase Overall Average: Group 1 Group 2 Group 3 Group 4 Group 5 Group 6 Group 7 2.9 2.6 2.6 2.4 2.8 2.6 2.9 Table 3. Ratings and inter-rater agreement for open-ended questions. Positive ratings Inter-rater agreement • • • • What did you learn about yourself in this workshop? What did you like about this workshop? What did you not like about this workshop? 16 94% 16 94% 17 100% 17 100% 16 94% 17 100% 8 47% 10 59% 15 88% What did you learn about yourself in this workshop? What did you like about this workshop? What did you not like about this workshop? Is there anything you would appraise or anything else you would like to say? T O R T U R E Vol um e 1 9 , Nu m be r 3 , 20 0 9 Two raters read all the answers to these questions separately, classified the responses as either positive or negative, and then the agreements between the raters were calculated (total number of agreements divided by the number of agreements plus the number of disagreements × 100). This gave an overall inter-rater reliability of 96%, which is highly acceptable. As can be seen from Table 3, there was a strong trend for all questions to give positive answers, even for the question about what was not liked by the participants, where nearly half still gave positive responses. Thus, it can be concluded overall that the workshop process was felt to be of value by all participants, at least at the end of the workshop. There were some complaints, but these were mostly about the quality of the food at one of the venues used. However, enjoying a workshop is not the same as the workshop having some longer term impact, which was assessed on follow-up through the repeat of the initial assessment. Improvement due to the workshop A total of 110 individuals were initially assessed for their suitability for the workshops, but only 73 were finally selected. As indicated above, the rationale for selection was whether the candidate was a primary victim of torture, and hence those rejected were largely secondary victims. The followup sample had significantly higher SRQ-20 scores than the selected sample.e Follow-up was only possible for 33 individuals, since this is a highly mobile population under significant threat and forced to move frequently. As can be seen from the table below, there was a decided mean shift downwards in the SRQ-20 scores for the follow-up group. The difference in the pre and post-SRQ scores was statistically significant, indicating a real improvement in the psychological state of the participants. Additionally, 36% showed a drop in the post-scores below the threshold for “caseness”, which indicates a return to psychological health. More complete data was available for 19