Despite the existing differences in the ability to generate changes among psychotherapists, there is still no robust evidence to explain what those results should be related to. Therefore, the aim of this thesis project is to find out whether these differences are linked to the use of language and communication skills (hereafter CS) by psychotherapists in clinical sessions. Because clinical settings by their nature are difficult to control, it is planned to use a convergence design, which suggests that the evidence found from different methodological designs can be considered robust evidence. Thus, in this project it is intended to use an analogous experiment and a study in natural setting of communication skills of psychotherapists. The sample in clinical settings will consist of six experts therapists (with more than 10 years of continuous experience in clinical sceneries) and six novice therapists (less than two years of continuous experience in clinical contexts) divided into two models (cognitive and systemic). Each of these 24 therapists must provide a case of one of depressive problems and anxiety disorders, taking a total of 48 cases analyzed. The analysis will be performed using the coding instrument SICOLENTE language.

Accordingly, the size of the sample (the conversational turns) will exceed the number of 250000 subjects. Regarding the analogous experiment, the sample will be split with the same criteria of clinical experience. It will take form by using a computer task where the subject will interact with a video of a professional actor representing a case under two conditions: a condition in which the client reacts to the answers of the participants collaboratively and other conditions in which the client reacts uncooperatively. The answers must be justified and selected from the options proposed by researchers. The task will be interactive so that, the video will be constantly changing the sketches to show, depending on the communication skills chosen by the participants. Possible outcomes would allow to develop linguistic patterns based on the experience of therapists, from which will generate predictors of success in therapy and develop new frameworks for training in psychotherapy.