Research and Literature
What the literature says
Distinguish efficacy (controlled trials, internal validity) from effectiveness (real-world settings, generalizability). Know the evidence base for major models, including cognitive-behavioral and behavioral couple therapy, integrative behavioral couple therapy, emotion-focused couple therapy, structural and strategic approaches for youth externalizing problems, multisystemic and family-based treatments, and solution-focused and narrative applications. Be able to say what a meta-analysis establishes versus what a single trial does.
The common factors literature emphasizes alliance, empathy, goal consensus, client expectancy, and extra-therapeutic factors; the specific-ingredients debate (sometimes framed by the "Dodo bird" verdict) is a standard exam item. Process-outcome research links alliance ruptures and repairs to outcome, which supports routine monitoring. Measurement-based care and feedback-informed treatment have empirical support for preventing deterioration.
Reading research critically
- Significance is not magnitude: read effect sizes and confidence intervals, not p-values alone.
- Moderators (age, culture, severity, presenting problem) determine who benefits; attrition and allegiance effects limit conclusions.
- Publication bias inflates apparent effects; replication matters.
- Systemic research is hard: unit of analysis (individual, dyad, family) and interdependence of family data require multilevel models.
Traps
- Treating an efficacy trial with tight inclusion criteria as proof for every community family.
- Assuming one study settles a question.
- Confusing correlation with causation, or client report with observed behavior.
- Overlooking that most couple and family research samples underrepresent marginalized groups.
Rule of thumb
Ask three questions of any study: compared with what, measured how, and for whom? If a claim cannot answer all three, it is marketing, not evidence.
Worked example: A family presents with adolescent conduct problems. Evidence supports family-based and multisystemic approaches over individual treatment alone; you select a family-based protocol, monitor response, and adjust rather than assuming the model guarantees success.