Selecting and Implementing Interventions
The decision framework
Select interventions by (1) the function identified in assessment, (2) the evidence base, (3) social validity - acceptability to the client and stakeholders, (4) feasibility and resources, (5) least restrictive and least intrusive options, and (6) risk–benefit. A technically elegant plan the teacher will never run fails criterion 4.
Implementation
- Treatment integrity must be measured, not assumed; low integrity invalidates conclusions about effectiveness.
- Progress monitoring with graphed data, IOA, and prespecified decision rules (for example, change if no improvement over a stated number of sessions).
- Generalization and maintenance planned from the outset: natural contingencies, fading of prompts and schedules, follow-up probes.
- Collaboration and consultation: identify implementers, train them with behavioral skills training, and build systems-level supports.
- Data-based modification: distinguish a poorly designed intervention from one that was poorly implemented before abandoning it.
Traps
Choosing an intervention because staff prefer it despite a contradictory functional assessment; changing course after two data points, or never changing with flat data; ignoring maintenance after discharge; treating caregiver agreement as informed consent without documentation; confusing social validity with treatment integrity.
Worked example
A parent requests time-out for tantrums, but assessment shows attention maintains them. The behavior analyst agrees to a plan that teaches an appropriate attention request, richly reinforces it, withholds attention for tantrums, and trains the parent through modeling, rehearsal, and feedback while measuring integrity weekly. This matches function, honors the parent's participation, and includes social validity and maintenance probes.