Independent variable

The analyst changes the independent variable

The dependent variable is the repeatedly measured outcome. The independent variable, or IV, is the condition deliberately changed to test whether that outcome responds. Systematic manipulation may involve introducing, withdrawing, alternating, staggering, or changing a defined value of the condition.

The BACB BCBA Test Content Outline, 6th edition covers independent and dependent variables, single-case design features, procedural integrity, comparative, component, and parametric analyses, and data interpretation. It is examination content rather than a research or clinical protocol.

A naturally occurring feature can predict an outcome without qualifying as a manipulated independent variable. If noise happens to differ across sessions, it is a contextual variable. If the investigator deliberately arranges two defined sound conditions in an approved experiment, sound level becomes an independent variable with specified levels.

A treatment name is too broad for replication

“Visual support,” “caregiver training,” “reinforcement,” and “usual care” can describe many different procedures. Define the steps that make each condition distinct:

• materials, cues, and environmental arrangement
• implementer actions and timing
• response criteria and partner consequences
• prompts, feedback, and error procedures
• session length, opportunities, intensity, and schedule
• start, stop, transition, and exception rules
• ordinary supports held constant across conditions

A single-case intervention-development review identifies replicable precision, systematic manipulation under investigator control, and overt fidelity measurement as quality features for independent variables.

Describe the comparison condition with the same care. “Baseline” might mean ordinary practice, no planned intervention, or another specified condition. Record what actually occurs there. An undefined comparison makes the contrast unclear and can hide active support.

The intended and delivered variables can differ

The protocol describes the intended independent variable. Procedural fidelity records what implementers delivered. If a three-step intervention occurs correctly in only half the sessions, the data do not represent clean exposure to the written procedure.

The SCRIBE 2016 reporting guideline calls for describing the intervention and control condition in each phase, how and when they were actually administered, and how procedural fidelity was evaluated. It also calls for the completed sequence, trials per session, adverse events, raw outcomes, and limitations. SCRIBE guides reporting rather than choosing a design.

A review of procedural-fidelity reporting defines fidelity as the extent to which an independent variable is implemented according to its predetermined procedure. Fidelity is separate from the outcome and from interobserver agreement. Staff can implement a procedure accurately while the dependent variable stays flat; observers can agree on outcome scoring while implementation drifts.

Packages, components, and levels make different variables

A multicomponent intervention can be the independent variable when the research question concerns the package as a whole. The resulting claim belongs to that combination. Identifying the contribution of a checklist, prompt, feedback step, or reinforcement procedure requires a component analysis.

A parametric analysis changes a quantitative value of one variable, such as 5-, 10-, and 20-second delays. Each value is a level of the independent variable. A comparative analysis contrasts separate interventions. Label conditions according to what was actually manipulated.

The 2024 practitioner's guide to procedural fidelity recommends defining package components individually when building a fidelity system. A single total percentage can conceal one consistently omitted element. Component-level scores reveal which condition was delivered.

Confounds obscure what the variable caused

Features meant to stay constant can change with the IV. If a new intervention always arrives with a new staff member, easier task, quieter room, longer session, or medication change, each offers an alternative explanation. The evidence may support the combined package while failing to isolate the named procedure.

Plan the contrast before data collection. Balance or randomize eligible condition order when appropriate, match materials, preserve session exposure, and log context changes. In staggered designs, collect comparison-tier data concurrently when possible. Record departures rather than relabeling them after the result appears.

Manipulation also needs enough separation to be detectable. Two conditions that differ only on paper may produce the same actual experience. Conversely, different colors, devices, partners, or rooms can become part of the condition. Report those correlated features.

A fictional independent-variable example

Priya is a fictional adult who uses speech and AAC during a preferred job-training routine. She chooses to test a predictable partner-response procedure. AAC and all ordinary supports stay available in every condition. The intervention requires the partner to pause for ten seconds after a naturally occurring help opportunity, recognize Priya's speech, AAC, sign, or defined gesture, and provide the available item or a clear answer within 15 seconds.

The three partner actions form the independent variable. Across 18 baseline opportunities under documented usual practice, Priya independently communicates in 6 of 18, or 33.3%. Across 18 intervention opportunities, she communicates in 14 of 18, or 77.8%. Staff complete all three planned actions in 17 of 18 opportunities, so opportunity-level fidelity is 17 of 18, or 94.4%. The missed opportunity remains in the denominator.

These values describe the intended procedure, its delivery, and the outcome separately. One baseline-to-intervention comparison cannot establish a functional relation by itself. The team reviews raw session order, context, preference, comfort, partner response, and a stronger design before making a causal claim.

Clinical authority sets the experimental boundary

An investigator controls only variables that can be lawfully, safely, and ethically manipulated. Medication remains under the authorized prescriber. Essential AAC, food, water, bathroom access, mobility, prescribed care, pain care, rest, emergency help, and safety protections remain available. Consent withdrawal, assent withdrawal when applicable, distress, or changing risk can stop a planned comparison.

The current BACB Ethics Code applies to BCBA and BCaBA certificants and people who completed an application. It addresses competence, client involvement, informed consent and assent when applicable, assessment-based intervention, risk, data, and evaluation. Experimental control never expands a person's professional or legal authority.