Measurements

Systemic Therapy Inventory of Change (Pinsof et al., 2009) is an assessment and feedback system in which clients fill out a questionnaire before each therapy session. Via electronic devices, the clients evaluate their response to treatment, including progression, and alliance to the therapist (Pinsof et al., 2009; Tilden et al., 2010). Client evaluations are processed into a report that is returned to the therapists who can use this information as the basis for understanding and hypotheses in the clinical assessment of the current client. The response options in STIC are on a five-level scale from worst to best option. Modules are added depending on the therapy mode and number of clients in the therapy system. The questions cover six subscales, individual problems and strengths (IPS), FOO, relationship with partner, family/household, children’s problems, and strength and relationship with child or children. On average, it takes 45 min to complete the STIC questions before the first treatment session. Before every subsequent session, the clients complete a short version of STIC that takes 7–8 min to fill out. Because this study is cross-sectional by making use of data before the first session only, the intersession data were not included. STIC has good internal reliability (Pinsof et al., 2009; Zinbarg et al., 2018; Zahl-Olsen et al., in review) with a Cronbach’s alpha as high as 0.94 for the different subscales.

The STIC consists of several subscales that further contain factors and items, and some of these address the topic for this study. The response options to those questions were, not at all/never, rarely, sometimes, often, and all of the time. The relationship with partner (RWP) scale has one item addressing physical violence between the members of the couple: “We get into shoving or hitting each other when we fight.” The family household (FH) scale has two items addressing physical violence exerted within the family: “Someone in my family is physically abusive to other family members” (item 1) and “There is someone in my family who pushes other family members around physically to get his or her way” (item 2). We combined these two items to one family violence item in our analysis. We hypothesized that both the CV item and the family violence item would load from the same latent variable Physical Violence. Further, the IPS scale, the RWP scale, and the FOO scale all have variables we, based on the presented theory and the literature review above, hypothesize are predictors for the underlying latent variable – Physical Violence. We also modeled that perceived anger toward the partner and level of expectation of household chores were predictors of Physical Violence. We hypothesized two negative predictors of Physical Violence: Self-control of thoughts, impulses, and rage as the first one, and sexual satisfaction within the relationship as the second. See Table 1 for a full list of items included in the variables.²

List of items used to model physical violence.