Predictors of treatment dropout from a spouse abuse abatement program.

McSweeney et al. (2000) tracked men in a spouse-abuse counseling program. They wanted to know who would quit early.

The team looked at police records, drinking denial, and personality traits. They used simple before-and-after tests.

Only three flags raised the risk of dropout: lots of police contact, saying 'I don't have a drinking problem,' and paranoid or borderline traits.

Even together, these signs predicted just seven percent of the variance. That is too weak to guide real decisions.

Bélanger et al. (2017) ran a similar check on panic-group CBT. They also found that severe anxiety and marital conflict forecast quitting, showing the same weak-link pattern across a different problem.

Greer (1982) had already warned that we barely measure dropout in abuse programs. K et al. answer that call, yet the low power shows we still need better tools.

McIntyre et al. (2017) looked at carer traits, not client traits, and found younger, lower-SES carers drove dropout in an ID trial. Together the papers hint that both client and carer variables matter, but each explains only a slice.

For BCBAs running court-linked programs, do not rely on a short screener to guess who will stay. Instead, plan for high attrition: use reminder calls, flexible scheduling, and rapid engagement in the first two sessions. Keep extra seats open so you can over-book by twenty percent and still finish with a full cohort.