Every experimental question in behavior analysis has the same shape: if I change this, does that behavior change? The thing you change is the independent variable. The behavior you measure is the dependent variable. Naming each one correctly is the first step in designing, reading, or critiquing any single-case study.
Why it matters
You cannot show that an intervention works until you are clear about what the intervention is and what it is supposed to change. If the independent variable is vague ("we worked on behavior"), nobody can replicate it. If the dependent variable is not measured directly and repeatedly, nobody can see whether it changed.
When a caregiver asks whether a new bedtime routine is "working," they are asking an experimental question: the routine is the independent variable, and minutes to fall asleep is the dependent variable. Vignettes on this task ask you to pick out each variable, spot a second variable that changed at the same time, or explain why an effect cannot yet be credited to the intervention.
Key ideas
- The independent variable (IV) is what the analyst manipulates. In applied work it is usually an intervention or environmental change: a visual schedule, a token system, a change in reinforcer magnitude, or a staff-training package.
- The dependent variable (DV) is the behavior that is measured. More precisely, it is a measurable dimension of the target behavior, such as rate, duration, latency, or percentage correct. The DV "depends" on the IV if the intervention has an effect.
- Manipulating the IV means controlling when and how it is present. The analyst introduces it, withdraws it, or changes its value (for example, two minutes of break versus five) while measuring the DV repeatedly.
- A functional relation is the goal. When the DV changes reliably and predictably each time the IV is manipulated, and other variables are held as constant as possible, you have evidence that the IV controls the behavior.
- Extraneous variables compete for credit. Anything else that might affect the DV, such as a new medication or a change of teacher, is an extraneous variable. When one changes along with the IV, it becomes a confound, and you cannot tell which produced the change.
- Change one variable at a time. If you add a new reinforcer and a new prompting procedure in the same phase, any change in behavior can be credited only to the package as a whole, not to either part.
- The IV has to be delivered as planned. If staff implement only part of the procedure, the IV that was actually tested is not the one on paper. Measuring treatment integrity tells you which IV your data describe.
Definitions
- Independent variable: the intervention or environmental condition that the analyst systematically manipulates to see whether it affects behavior.
- Dependent variable: the measured dimension of the target behavior, recorded repeatedly to detect any change produced by the independent variable.
- Functional relation: a demonstration that a specific change in behavior can be reliably produced by manipulating the independent variable and is unlikely to be the result of other factors.
- Extraneous variable: any variable other than the independent variable that could influence the dependent variable.
- Confounding variable: an extraneous variable that changes along with the independent variable, so its effects cannot be separated from those of the intervention.
- Treatment integrity: the extent to which the independent variable is implemented as it was designed (also called procedural integrity or procedural fidelity).
Examples from a session
Naming the variables
- Clinic: A BCBA tests whether a first-then board shortens transitions for Amara, 5. IV: the first-then board. DV: seconds from the transition instruction until Amara arrives at the next station (total transition time).
- School: A teacher and consultant add a self-monitoring checklist for Devon, 13, during independent seatwork. IV: the self-monitoring checklist. DV: percentage of 10-second intervals scored on task.
- Home: A family tries a bedtime pass for Rosa, 8. IV: the bedtime pass procedure. DV: number of times Rosa leaves her room after lights-out each night.
- Adult services: A BCBA at a day program uses video modeling to teach Mr. Daniels, 42, to prepare a simple lunch. IV: video modeling. DV: percentage of task-analysis steps completed independently.
- Supervision: A BCBA trains RBTs with behavioral skills training. IV: the training package. DV: the RBTs' percentage of teaching steps implemented correctly. The clients' progress is a separate question unless it is also measured.
- Non-example: In Amara's study, staff give the same transition instruction in baseline and with the first-then board. The instruction is not the IV: it is held constant across phases, so it is part of the setting, not the variable being tested.
A quick way to sort any vignette: ask "What did the analyst change on purpose?" (IV) and "What did the analyst count, time, or score?" (DV). If the answer to the second question is not a behavior, look again.
Common mistakes / exam traps
- Calling the behavior the IV. The behavior is what you measure, so it is the DV. The IV is something the analyst does to the environment.
- Mixing up the DV with the goal. "Improve Devon's focus" is a goal. The DV is the specific measure, such as percentage of intervals on task.
Exam trap
IV vs. confound. Both change at the phase line; the difference is intent. The IV is the one thing the analyst deliberately introduced or varied. A confound is anything else that changed at the same time, such as a new medication. If one is present, the effect cannot yet be credited to the intervention.
Exam trap
Ineffective procedure vs. low treatment integrity. No change in the DV supports "the procedure does not work" only if integrity data show it was delivered as designed. If staff implemented it correctly on, say, half of opportunities, the data describe a weaker IV. Fix implementation first, then judge the procedure.
Quick check
Jonah, a BCBA, wants to know whether adding choice of task order changes how many math problems Lina, 10, completes in a 15-minute work period. Which is the dependent variable?