Echoes of the Holocaust

Contents
Interviewers' Reactions to Holocaust Survivors' Testimony
Aviva Mazor, Yolanda Gampel, Gilit Horwitz
Content Analyses of the Phenomenological Data
The subjects' responses were analyzed according to Helgson, Shaver and Dyer (1987), and Selman, (1980). In the first step, two experts read all the subjects' descriptions and recorded every emotion, thought, concept, reason, wish, or behavior expressed in response to every question. In the second step, three judges read all these issues collapsed into specific categories. A category was defined based on issues which constituted common or similar meanings for a specific question. For example, "What thoughts and feelings did you recall immediately after the interview - and later on?" All the following responses were collapsed into one category: "I felt admiration and amazement," "A great appreciation of the person who survived," "Again and again I thought, how wonderful it was that despite all the horrors these people went through, they were coping and functioning normally," and "Amazement and admiration at how they could retain their humanity and survive." The title of the category was "Appreciation of the survivor."
In the third step, the frequency of responses was calculated for every category. If at least three subjects (5%) responded similarly to the categories, it was included in the final coding and the data analyses. However, if there were only one or two responses relating to a specific category, they were included in the "most similar" category. The instrument was employed in a pilot study that helped clarify the questions. Three experts read the response lists of the subjects and conducted the second and third steps of the content analysis. Interrater reliability was calculated by correlating the three raters' scores across all the protocols. The correlation was quite high (.90). All disagreements were resolved without any particular difficulties.
Scoring
In order to employ statistical procedures, the scoring for item 1 was ordinal due to the content of the responses. For all the other items (besides 5 and 7) all the nominal categories were changed to dichotomic series, and the subject's answer scored 1 if the response was related to a category, and 0 if not. Regarding items 5 and 7, ordinal categories were devised due to the nature of the responses (Horwitz, 1991).
Results
Data analysis
The first step in data analysis focused on drawing 91 categories of responses, derived from the content analysis of the questionnaires. In the second step of data analysis, a matrix of correlation coefficients was computed using a Principle Component Analysis with rotation of Promax (based on Varimax), which combined different categories into factors. To avoid overfactorization, we used the regular criteria (clear scree or Eigen > 1). Using the cluster analysis of the variables according to SAS (Krzanowski, 1988), the correlation coefficients matrix produced 25 clusters. In the third step, these 25 clusters were defined according to four domains which were built into the questionnaire's rationale. These four domains included (1) the preparation of the interviewer; (2) the influence of the interview on the survivor; (3) feelings and thoughts after the interview; (4) the perception of the interviewer's role.
There were two main purposes for this cluster analysis: one was to characterize the response patterns within each content domain, which refers specifically to different periods of time with respect to the whole interviewer experience. The second purpose was to reduce the number of categories that were greater than the N of the sample. In addition, the cluster analysis aimed at revealing the relationship between the 25 clusters that were combined into 14 clusters which described the main response patterns.
Table 1 describes the contents of the 14 main clusters of interviewers' responses according to the time sequence in the four domains of the questionnaire.

Table 1: Content of the Clusters of the Interviewers' Responses Based on the Interview Process

Time Sequence in Questionnaire Responses Before the Interview Responses During the Interview Responses After the Interview Responses Before & After the Interview

Interview Process in the Questionnaire Interviewer's Preparations
Clusters 1, 7, 12, 13 Interviewer's Experiences
Clusters 2, 6 Interviewer's Reflections
Clusters 3, 8, 9, 10, 11 Interviewer's Reflections 4, 5,14

Cluster 1: Fear responses. Cluster 2: Intellectual structuring of interview process. Cluster 3: Interviewer's emotional strain as well as perceiving the coping processes of survivors. Cluster 4: Interviewers' anxieties about themselves before the interview. After the interview, the survivors were able to express their pain and leave their testimonies.

Cluster 7: Directions & rules of behavior to aid interviewers. Cluster 6: Acceptance, concern and interest towards the survivor. Cluster 8: Interviewer's growing self-awareness and the recognition that the interview brings relief to survivors. Cluster 5: Before and after the interview there were fears of the survivors' emotional responses. There was a need for working through by interviewers.

Cluster 12: Anxiety responses due to personal involvement. Cluster 9: Identification and guilt feelings toward survivors. Later working through process proceeds. Cluster 14: Subjective emotional preparation of interviewer before the interview. After the interview, the interviewer transforms personal involvement into clinical and historical knowledge.

Cluster 13: Curiosity and intellectual interest. Cluster 10: Painful feelings evoked in the survivors. The appreciation of survivors' endurance.

Cluster 11: Feelings of shock, dismay and anxiety regarding survivor and interviewer's family. Long-term effects of painful feelings continue in interviewer's life.
[Page 5 of 9]