Several studies have examined change processes using task analysis to link proximal outcomes occurring within or between sessions to specific therapeutic events. In an initial pilot study comparing IBCT with TBCT, Cordova, Jacobson, and Christensen (1998) found that couples in IBCT showed relatively more constructive detachment over the course of therapy, and that these changes predicted couples' gains in relationship satisfaction. Three task-analytic studies of EFCT (Greenberg, Ford, Alden, & Johnson, 1993) showed that: (a) couples receiving EFCT demonstrated more shifts from hostility to affiliative behaviors than waiting-list couples; (b) best sessions as identified by couples were characterized by more depth of experiencing and affiliative and autonomous statements than were sessions identified as poor; and (c) intimate, emotionally laden self-disclosure by one partner was more likely to lead to affiliative statements by the other partner than were other randomly selected responses. A more recent task analysis of four EFCT sessions by Bradley and Furrow (2004) found that specific therapist interventions linked to ''softening events'' (reformulations of partners' hostile or critical comments as expressions of vulnerability) involved intensifying a couple's emotional experience and promoting intrapsychic awareness and interpersonal shifts in attachment-related interactions.
Predictors of Treatment Outcome
Research targeting predictors of couple therapy outcome has overwhelmingly emphasized prognostic rather than prescriptive indicators (Snyder et al., 2006). The former predict response to a particular treatment (or response across treatments, irrespective of specific approach), whereas the latter predict response to one versus another treatment. Although prescriptive indicators are essential to informed selection of specific interventions in tailoring treatment to partner or relationship attributes, empirical findings in this regard are virtually non-existent. Studies examining demographic, individual, and relationship prognostic indicators of treatment outcome are more frequent, but their yield has been modest and mixed (see reviews by Snyder et al., 2006; Whisman, McKelvie, & Chatav, 2005). The most consistent finding (but of limited usefulness in selecting interventions) is that couples having the greatest difficulties in their relationship are less likely to benefit from treatment, with initial levels of relationship distress accounting for up to 46% of the variance in treatment outcome (Johnson, 2002). Lack of commitment and behavioral steps taken toward divorce have been associated with poor treatment outcome to BCT in two studies (Beach & Broderick, 1983; Hahlweg, Schindler, Revenstorf, & Brengelmann, 1984), but not in another (Jacobson, Follette, & Pagel, 1986). Snyder, Mangrum, and Wills (1993) found that poorer response to couple therapy was predicted by lower relationship quality, greater negative relationship affect and disengagement, and greater desired change in the relationship.
The effectiveness of couple therapy in the field outside of research settings has not been researched sufficiently. In two uncontrolled prospective studies, Hahlweg and Klann (1998) and Klann (2002) investigated the effectiveness of marital counseling in Germany. In total, 139 counselors recruited 1,152 clients into the study. The interpretation of the results is difficult because the attrition rate was high: Only about 50% of the clients participated in the post-assessment six months later. Pre-post comparisons resulted in significant improvements in several scales. The effect size for the Global Dissatisfaction Scale (of the Marital Satisfaction Inventory, MSI; Snyder, 1981) was ES = 0.45, a medium effect size. About 26% reported happy relationships after couple counseling - considerably less than in the Christensen, Atkins, Yi, Baucom, and George (2006) study, which leav