Data Collection and Analysis Flashcards

A behavior analyst is using a partial-interval recording system across 10-minute sessions divided into 30-second intervals. The target behavior (self-injurious head-hitting) occurs for a continuous 45 seconds starting at the 2-minute mark. How many intervals will be scored as '+' (behavior occurred)?

Answer: 2 intervals

In partial-interval recording, an interval is scored '+' if the behavior occurs at any point during it. The 45-second bout begins at 2:00 and ends at 2:45, spanning the interval covering 2:00–2:30 AND the interval covering 2:30–3:00. Only 2 intervals are marked '+', even though the behavior lasted into a third partial second of the next interval — 2:45 falls within the 2:30–3:00 interval, making exactly 2 complete intervals affected.

When graphing frequency data, a practitioner notices that session lengths have been highly variable (ranging from 15 to 60 minutes) across the baseline phase. Which data transformation is most appropriate before interpreting trends?

Answer: Convert frequency to rate (responses per minute)

When session length varies substantially, raw frequency counts are not directly comparable across sessions because a longer session provides more opportunity for responding. Converting to rate (responses per minute) standardizes the data relative to observation time, making cross-session comparisons valid. Cumulative records do not address variable session length. Log scales address variability in magnitude, not duration. Percentage of opportunities requires a discrete-trial context with a fixed number of opportunities.

A BCBA supervising an ABAT instructs her to collect IOA data using the 'scored-interval' IOA method. On a 20-interval observation, Observer 1 scores 8 intervals as '+' and Observer 2 scores 6 intervals as '+', with 5 intervals scored '+' by BOTH observers. What is the scored-interval IOA?

Answer: 55.6%

Scored-interval IOA focuses only on intervals where at least one observer scored '+'. The formula is: agreements on '+' / (agreements on '+' + disagreements) × 100. Agreements on '+' = 5 (both scored '+'). Disagreements = intervals scored '+' by only one observer = (8 − 5) + (6 − 5) = 3 + 1 = 4. IOA = 5 / (5 + 4) × 100 = 5/9 × 100 ≈ 55.6%. This is stricter than total-interval IOA because it ignores intervals both observers scored as '−'.

A behavior analyst is selecting between momentary time sampling (MTS) and whole-interval recording (WIR) for a behavior she suspects occurs at approximately 40% of the time. Which statement BEST describes the measurement bias expected from each method?

Answer: MTS will yield an accurate estimate while WIR will underestimate true prevalence

Momentary time sampling records behavior only at the precise moment the interval ends; research shows MTS yields relatively accurate prevalence estimates when behavior is not extremely rare or extremely frequent. Whole-interval recording requires the behavior to occur for the ENTIRE interval to be scored '+', which systematically underestimates true prevalence — the behavior could occur for 95% of the interval and still be scored '−'. Partial-interval recording is the method known to overestimate.

An ABAT is using an equal-interval (arithmetic) graph versus a semi-logarithmic (Standard Celeration Chart) to display the same response rate data showing 10-fold growth over 8 weeks. Which statement about interpreting trend lines on these two charts is accurate?

Answer: A straight trend line on the arithmetic chart indicates linear growth; a straight line on the semi-log chart indicates proportional (multiplicative) growth

On an equal-interval (arithmetic) graph, equal vertical distances represent equal absolute changes — so a straight line means the behavior is increasing or decreasing by the same absolute amount each unit of time (additive/linear growth). On a semi-logarithmic chart like the Standard Celeration Chart, equal vertical distances represent equal ratios or proportional changes — so a straight line means the behavior is multiplying by a constant factor each unit of time (multiplicative/exponential growth). This distinction is critical for accurately characterizing learning celeration.

During a functional analysis, a behavior analyst reviews her data and finds that problem behavior occurred at high rates during the alone condition AND the attention condition, but at near-zero rates during the demand and play conditions. Which function does this pattern MOST suggest, and why is the alone-condition data critical to that interpretation?

Answer: Automatic reinforcement function; high rates in the alone condition rule out social mediation as the primary maintaining variable

In a standard functional analysis, the alone condition controls for automatic (sensory) reinforcement by removing all social consequences. High responding during 'alone' indicates the behavior produces its own reinforcement without any social mediator. If behavior were primarily attention-maintained, we would expect low rates during alone (no attention available). The concurrent elevation in the attention condition could reflect that attention is also reinforcing, but the alone-condition pattern is the definitive indicator that automatic reinforcement is the primary maintaining variable.