Parsimony

Parsimony reduces unsupported assumptions

When several explanations could fit an observation, begin with accounts grounded in evidence that can be checked. A missed appointment may relate to transportation, unclear reminders, illness, cost, schedule conflict, or a choice to stop. A speculative personality label adds little until those conditions are understood.

The BACB BCBA Test Content Outline, 6th edition lists parsimony among philosophical assumptions underlying behavior analysis. It also covers description, prediction, control, measurement, experimental design, assessment, referrals, and data-based decisions. Those skills help turn parsimony into disciplined inquiry.

Simple means adequate and testable

The shortest explanation can be wrong. Parsimony favors the account that explains the observations with the fewest unsupported additions, while respecting established facts.

A useful explanation should identify evidence, make a prediction, and allow a result that would weaken it. “He lacks motivation” offers few clear tests. “The task instructions are absent on days with delayed starts” suggests checking instruction availability and comparing starts when instructions are present.

A fictional scheduling example

Camila is a fictional teenager who arrives late to a weekly cooking group. Staff initially describe the pattern as avoidance. The team defines on-time arrival as entering within five minutes of the scheduled start and reviews eight sessions.

Camila arrives on time in 3 of 8 sessions. Her accessible text reminder is sent at the agreed time in 4 of 8, and she arrives on time in 3 of those 4. On the four days without a timely reminder, she arrives late every time. Camila also reports that the schedule changes are hard to track.

The reminder account is simpler and better supported than a broad avoidance label. It remains a descriptive association. Transportation, activity preference, fatigue, and other conditions could still matter. The team fixes the reminder process and continues measuring with Camila.

Check health and access early

Pain, sleep, medication, hearing, vision, feeding, seizures, trauma, mobility, and mental health can affect behavior. Parsimony provides no reason to delay an appropriate medical or interdisciplinary referral. A behavioral explanation that ignores known health evidence becomes less adequate, not more parsimonious.

Check communication and AAC access, understandable instructions, accessible materials, environmental noise, staffing, and the person’s ability to decline or request help. System failures often supply a more direct explanation than a presumed client deficit.

Separate description from explanation

First describe the pattern. State the response, setting, timing, opportunities, and source of each observation. Then list plausible explanations and evidence for each.

Interviews can identify variables outside the observation window. Direct observation can test parts of those reports. Experimental comparison can strengthen a causal claim when safe and appropriate. Keep the claim level matched to the method.

Parsimony does not erase complexity

Behavior can reflect interacting variables across biological, learning, social, and cultural histories. A complete account may need several components. Use parsimony to prevent unnecessary invention, rather than to force every case into one immediate contingency.

Add complexity when evidence requires it. Document why each added variable improves prediction or decisions. Remove components that contribute no explanatory or practical value.

Use a competing-hypothesis table

List each explanation, supporting evidence, conflicting evidence, missing information, safe test, and responsible owner. Rank urgent health and safety questions separately from likelihood. A lower-probability medical risk may need faster action than a likely scheduling issue.

Predefine the result expected from each explanation. When the data arrive, revise the ranking. This process protects against settling on the first familiar account.

Common misuses

Avoid equating parsimony with blaming the person, choosing the cheapest intervention, or refusing interdisciplinary input. Avoid treating a diagnosis as a complete cause. Avoid using one observation to dismiss the client’s report.

Parsimony works best with empiricism: begin with a clear, economical account, expose it to observation, and add or remove complexity as evidence changes.

Audit the chosen explanation

Before a decision, ask a colleague to trace every claim to its source. Mark which facts were directly observed, reported by the person, inferred from a pattern, or drawn from research. Remove assumptions that do no work.

After intervention, compare the predicted and observed results. If the account predicted faster starts when instructions were available and starts remain unchanged, preserve that finding. Check implementation, then revise the explanation rather than adding invisible causes.

Use the simplest measurement capable of answering the question. A brief direct count may outperform a long rating scale when the decision concerns occurrences. A complex assessment is justified when a simpler method cannot resolve the clinically important uncertainty.