Do high-risk couples profit more or less from couple relationship education programs than low-risk couples? Room for improvement and vulnerability effects.

In recent years it has been discussed whether high-risk couples benefit more from Couple Relationship Education programs (CREs) than low-risk couples due to larger room for improvement, or profit less due to greater vulnerability. Pertinent response prediction studies yielded inconclusive results. Careful review suggests this may be due to: statistical handling (not disentangling room for improvement and vulnerability effects), time frame analyzed (not disentangling opposing effects during intervention and follow-up), sampling, and selection of risk factors. We used an analytic strategy that maximized odds for replicability and tested two hypotheses: (1) room for improvement: pre-intervention relationship dissatisfaction predicts gain in satisfaction during intervention, and decline during follow up, and (2) vulnerability: when adjusted for room for improvement (pre-intervention relationship dissatisfaction), risk factors show negative or negligible, but no positive associations with gain in satisfaction. Actor-Partner Interdependence Modeling (APIM) was employed in 79 self-referred (SR) couples and 50 clinician-referred (CR) couples who had completed the 'Hold me Tight' program, a CRE based on Emotionally Focused Couple Therapy. Our findings supported both the room for improvement hypothesis, with pre-intervention dissatisfaction predicting more gain during intervention (both samples) and decline during follow-up (SR sample, for the CR sample the effect was negligible), and the vulnerability hypothesis, as several negative, but no positive effects of risk factors were observed during intervention and follow-up. Specific risk factors did not replicate between samples. To promote replicable results in future research, we advocate disentangling room for improvement and vulnerability effects, separately testing effects during intervention and follow-up, purposeful sampling, and studying a large set of risk factors including partner variables.