Narrative psychological content analysis in studies of therapeutic impact  

Péley B.
University of Pécs Institute of Psychology, Pécs, Hungary

Impact of psychotherapy is a long disputed issue in clinical psychology. Westen and Weinberger (2004) depict the studies along the coordinates "Method of Aggregation" x "Type of Informants". Method of aggregation can be statistical vs. informal, whereas the source of data can be either the patient (self report) or the clinician (clinicians report). The two coordinates give four quadrants. The authors claim that whereas two quadrants are heavily loaded, "Until recently, virtually no research has addressed the quadrant, which crosses clinical observation with statistical aggregation". There are also missing studies working with informal data collection from the patients’ self reports. This lack of research is partly due to the reluctance of researchers for dealing with interpreted data statistically. One of the major advantage of the narrative psychological content analysis is that it has the capacity to give diagnostic judgments on the patient’s psychological states, which are based on her informally occurring narrative discourse, and these data are apt to statistical aggregation. A further advantage is that whereas the frequency of using objective tests, which are easily amenable to statistical analysis, is limited in the course of a therapy, discourse samples are available in a non-intrusive way, and content analysis works as a quasi on-line methodology. The lecture presents the results of a pilot study, which compared efficiency of the cognitive scheme therapy with that of the short dynamic therapy along the changes in the therapeutic discourse. Narrative patterns that were quantified in this study included activity-passivity (agency), mentalization, intentionality, evaluation, and characters' psychological functions. Activity-passivity dimension was measured by an algorithm based on a dictionary of active verbs (e.g. do, construct, etc.) and passive verbs (e.g. sleep, stand, etc.) and on local grammars specifying the use of these verbs with meaning of activity and passivity. The algorithm involves partial syntax for identifying passive voice. Similar grammars were built for intentionality. The intentionality algorithm distinguishes between wishes (e.g. hope) and intentions (e.g. want) as well as between "musts" (e.g. ought, must, should) or possibilities (e.g. may). The mentalization algorithm includes three subscales: emotionality, intentionality and cognitive processes. Each subscale is constructed with the logic described above. The emotion algorithms of course handle with adjectives and adverbial forms. Samples of therapeutic discourse were analyzed with the above tools. Therapists gave parallel evaluations on the sessions and on the progress of the patients. In both therapy types systematic changes could have been observed in the course of the therapy; e.g. at sessions, where therapists perceived progress, frequency of activity and intentionality increased.