To assess high-risk skills without creating avoidable danger, use records, client report, interview, safe simulation, component probes, and qualified interdisciplinary evidence before any live exposure. Define health, setting, staffing, access, consent, assent, stop, and emergency gates. Never test performance by creating uncontrolled traffic, fire, water, medication, elopement, violence, or other hazards solely to obtain data.

Define the hazard and decision

Describe the exact activity, foreseeable harm, current protection, decision, and qualified roles. Separate assessment of knowledge, recognition, communication, component performance, and full real-world performance. They need different evidence.

Use an evidence hierarchy

Begin with Luz's report, records, incident history, caregiver or staff examples, medical or specialist evidence, and safe component review. Add tabletop or simulation methods that preserve protection. Live observation occurs only when naturally available and safely authorized.

Build readiness gates

Verify current health information, communication and AAC, mobility, environment, staffing, competence, consent, assent, privacy, protective equipment when appropriate, emergency contact, and rescue capacity. A failed gate creates a hold.

Set observable stop rules

Define pain, distress, confusion, assent withdrawal, environmental change, equipment failure, staffing loss, procedural deviation, and risk thresholds plus immediate response. Everyone present may stop for imminent danger without routine approval.

Avoid false realism

Simulation can assess selected components under arranged conditions. It does not establish safe real-world performance. Record which cues, hazards, distances, people, noise, time pressure, and consequences were absent.

Route qualified questions

Medical, swallowing, mobility, mental-health, driving, water, fire, medication, or other specialized questions require the relevant profession and authority. ABA data can contribute bounded observations without replacing those decisions.

Protect Luz's access and participation

For Luz, preserve AAC and other effective communication, privacy, ordinary supports, food, water, bathroom, mobility, prescribed care, pain response, rest, relationships, and emergency help. Verify health and safety before probing performance. Use applicable consent and assent processes, honor refusal and withdrawal during nonemergency participation, and route medical, mental-health, educational, communication, mobility, or other specialist questions to qualified roles.

Build Luz's skills-evidence ladder

Begin with Luz's accessible report, priorities, strengths, and ordinary routines. Add source-labeled interview, records, natural observation, structured opportunities, work samples, published measures, or interdisciplinary evidence only when the decision needs them. Keep skill, typical adaptive performance, access, opportunity, support, prompt, diagnosis, treatment recommendation, authorization, and outcome separate.

Explain the assessment to Luz

Use Luz's preferred language and communication mode to explain the purpose, activities, settings, tools, people, supports, prompts, recording, privacy, risks, stop route, possible decisions, and uncertainty. Invite questions, correction, refusal, pause, and withdrawal where applicable. Assessment participation never requires speech, eye contact, suppression of harmless traits, or independence from useful supports.

Build Luz's high-risk skill assessment plan

Create one versioned high-risk skill assessment plan for the community-safety review. Include question, client priority, strengths, source, setting, activity, opportunity, response, access, health, ordinary supports, prompts, refusal, no response, invalidity, sampling, methods, tools, qualifications, integrity, raw data, missingness, uncertainty, authority, decision, referral, client communication, revision, and recheck trigger.

Work through Luz's example

Luz's plan contains 12 safety-related components. Eight can be assessed through records, accessible interview, task analysis, or safe simulation. Three require medical, mobility, or community-safety specialist input and remain held. One low-risk component can be observed naturally with existing protections. Report 8/12 safely assessable, 3/12 referred, and 1/12 natural observation. No category authorizes a dangerous live test. Preserve all planned and valid units, raw counts, denominators, source labels, support and prompt states, access conditions, refusals, invalidity, corrections, versions, decisions, and open work. This fictional example demonstrates one assessment control. It supplies no universal cutoff, diagnosis, deficit, medical-necessity finding, legal conclusion, treatment effect, or outcome guarantee.

Use Luz's evidence for a bounded decision

Identify the exact decision and qualified owner before expanding assessment. For Luz, document supporting and conflicting evidence, client priority, strengths, access, health, opportunity, support, alternatives, burden, further evidence, chosen action, implementation check, accessible explanation, disagreement route, and review trigger. Choose the least burdensome adequate path and stop collection that cannot change action.

Review Luz's data without causal shortcuts

For Luz, show raw performance and denominators, representative opportunities, direct and proxy sources, ordinary supports, prompts, access, health context, refusals, invalidity, method integrity, missingness, alternatives, and uncertainty. Observed performance describes the sampled conditions. It never proves why the response occurred, what happens everywhere, or what intervention will work.

Address Luz's main assessment risk

A realistic street-crossing probe could expose Luz to moving traffic for the sake of assessment. A clinic simulation may also overstate performance in the community. The plan uses layers of evidence and names the transfer limit. Review definitions, client meaning, access, health, opportunity, context, support, prompt, method fit, manual adherence, sampling, causal scope, and authority separately. A total score cannot repair inaccessible administration or an irrelevant referral question. Technical labels never override Luz's current communication or right to useful support.

Choose Luz's next assessment action

Luz chooses a virtual route rehearsal, reviews the emergency stop message, and waits for mobility input before any community component changes. Record qualified owners, current safeguards, method and tool versions, evidence or referral tasks, due dates, validation, client communication, uncertainty, decision, and next review. Preserve original observations when the account changes. A later assessment adds an auditable version instead of rewriting what Luz did under earlier conditions.

Apply current clinical-assessment sources to Luz

For Luz, the CASP public summary supplies high-level context for ABA treatment of autistic people. The BCBA outline includes strengths, skill needs, skills assessments, measurement, cultural variables, referral, and client-informed goals as examination content. The current BACB Ethics Code addresses competence, understandable communication, client involvement, consent and assent when applicable, assessment, referral, risk, documentation, and evaluation for covered people. BACB has no separate organizational jurisdiction.

Keep adaptive and proprietary tools in scope for Luz

AAIDD describes adaptive behavior within its intellectual-disability framework as conceptual, social, and practical skills learned and performed in everyday life. The public Pearson Vineland-3 page identifies interview, parent or caregiver, and teacher forms and three broad domains. For Luz, those sources explain scope only. A published tool follows its current manual, qualifications, licensing, administration, accommodation, scoring, norms, interpretation, and security rules.

Protect access and assent in Luz's assessment

Breaux and Smith offer assent-focused practice guidance in an evolving evidence base, and ASHA says AAC users should always have access to their tools or devices. For covered public accommodations, the DOJ Title III overview describes equal opportunity, reasonable modifications, effective communication, and physical access subject to the law's standards and defenses. Preserve Luz's accessible communication, refusal, pause, correction, and ordinary supports.

Separate school evaluation rules for Luz

When Luz's work is part of an IDEA evaluation, 34 CFR 300.304 requires the public agency to use a variety of tools and strategies, avoid a single measure as the sole criterion, use assessments for their validated purpose, administer them through trained personnel under producer instructions, and meet its language and nondiscrimination requirements. That rule governs covered IDEA evaluations; it is not a universal ABA protocol or authority for other settings.

Close Luz's skills review

Review the high-risk skill assessment plan with Luz, the responsible clinician, access and operational owners, and the specialists named in the manifest. Verify that the question stays meaningful, strengths and supports remain visible, methods and tools stay within scope, invalidity and missingness remain explicit, interpretations avoid deficit framing, and decisions remain within authority. Keep this page draft and noindex until every required review is complete.

Related resources

Sources