Assessment and Identification
What this domain tests
Multimethod, multisource, multisetting evaluation; functional behavioral assessment methods; and the data metrics you must compute or interpret correctly.
Evaluation architecture
- Pre-referral: universal screening, MTSS tiers, documented intervention fidelity before referral.
- Multifactored evaluation: no single measure decides eligibility; instruments must be technically adequate, given in the child's language and mode of communication, and validated for the purpose used.
- Triangulate: records, interviews, rating scales, direct observation, and work samples.
FBA methods
- Indirect: record review, functional assessment interview, rating scales (e.g., Motivation Assessment Scale, Functional Analysis Interview, QABF).
- Descriptive: ABC narrative recording, scatterplots, conditional probability analysis.
- Experimental functional analysis: alternating conditions (attention, escape/demand, tangible, alone/automatic) against a control condition. Strongest evidence of function, highest cost and risk.
Measurement metrics
Event/frequency (count per unit time), duration, latency, whole- and partial-interval recording, and momentary time sampling. Partial interval overestimates duration; whole interval underestimates; momentary time sampling is the best unbiased estimate of duration.
Worked example
Two observers code a 20-minute session in 10-second partial intervals (120 intervals). They agree on 108. IOA = agreements / (agreements + disagreements) x 100 = 108/120 = 90 percent, acceptable. The teacher then reports the student is off task 90 percent of the time, but the data show off-task in 78 of 120 intervals, or 65 percent of intervals. Because partial interval overestimates, the team should switch to duration or momentary time sampling before writing a baseline.
Traps that cost marks
- Accepting a rating scale score as a diagnosis or as sufficient evidence of function.
- Baseline shorter than the behavior's variability warrants; aim lines built on a two-day sample.
- IOA below 80 percent treated as reliable.
- Missing setting events (sleep, medication changes, bus incidents) that explain day-to-day swings.