Empiricism in applied behavior analysis (ABA) is the practice of basing all conclusions and decisions on directly observable, measurable data rather than on opinions, assumptions, or gut feelings. It is one of six attitudes of science that form the philosophical foundation of behavior analysis, and it shapes everything from how practitioners assess behavior to how they decide whether a treatment is working.
Empiricism as a Scientific Attitude
ABA is built on six attitudes of science: determinism, empiricism, experimentation, replication, parsimony, and philosophic doubt. Empiricism is the one that governs how information gets collected. It requires that observations be objective, meaning they are independent of the individual prejudices, tastes, and private opinions of the person doing the observing. If two different people watch the same behavior and use the same measurement system, they should reach the same conclusion. That independence from personal bias is the core idea.
This stands in contrast to approaches that rely on clinical intuition, anecdotal reports, or subjective impressions of progress. A parent saying “he seems calmer this week” is valuable feedback, but it isn’t empirical data. Empiricism in ABA means translating that impression into something countable and verifiable: the number of times a behavior occurred, how long it lasted, or how quickly it started after a prompt.
What Empirical Data Looks Like in Practice
Behavior analysts use several standardized ways to measure behavior, each designed to capture something specific:
- Frequency: A simple count of how many times a behavior occurs during an observation period.
- Rate: Frequency divided by time, which allows comparison across sessions of different lengths.
- Duration: How long a behavior lasts from start to finish.
- Latency: The time between a prompt or instruction and the start of the response.
- Interresponse time: The gap between two consecutive occurrences of the same behavior.
Each of these produces a number. That number can be recorded, graphed, compared across days or weeks, and shared with other professionals who can independently verify it. This is what makes the process empirical. The data exist outside any single person’s interpretation.
How Empiricism Guides Treatment Decisions
Empiricism isn’t just a philosophy that sits in the background. It directly controls how behavior analysts decide what’s working and what isn’t. Once data are collected, they get plotted on graphs, and practitioners use visual analysis to evaluate three main features: level (the average performance during a phase), trend (whether the data are moving up, down, or staying flat), and variability (how much the data points bounce around). A treatment is considered effective when the graph shows a clear, consistent change that lines up with the introduction of the intervention.
For example, if a child’s frequency of disruptive behavior is steady at around 15 times per session during a baseline phase, then drops to 3 or 4 per session after an intervention starts, the visual contrast on the graph gives the analyst empirical grounds to say the intervention is working. If the data stay flat or become highly variable, the analyst has empirical grounds to change course. Decisions are driven by the pattern in the data, not by how the analyst feels about the approach.
Empiricism in Functional Assessment
One of the clearest examples of empiricism in ABA is the functional analysis, a method for identifying why a problem behavior keeps happening. Rather than guessing at the cause, the analyst sets up controlled test conditions. In an attention test, for instance, the therapist withdraws attention and then delivers it briefly each time the problem behavior occurs. In a control condition, the person has free access to preferred items, receives frequent attention, and faces no demands. By comparing the rate of problem behavior across these conditions, the analyst can see which variable is maintaining it.
This is empiricism applied directly to assessment. The analyst doesn’t assume the behavior is “attention-seeking” based on how it looks. Instead, they create conditions that isolate one possible cause at a time, measure the behavior under each, and let the data reveal the function. The result is an objective demonstration of a cause-and-effect relationship, which then informs the treatment plan.
Why It Matters Ethically
The emphasis on empiricism isn’t optional in the field. The Behavior Analyst Certification Board’s ethics code requires practitioners to rely on professionally derived knowledge based on science when making clinical judgments. Clients have a right to effective treatment grounded in the research literature and adapted to their individual needs. Behavior analysts are obligated to collect and graphically display data in a way that supports sound decision-making, and they cannot omit findings that might change the interpretation of their work.
These requirements exist because the consequences of ignoring data are real. A child could spend months in an ineffective program if no one is tracking progress objectively. An intervention could inadvertently make a behavior worse if no one is measuring it carefully enough to notice. Empiricism protects clients by forcing practitioners to let observable evidence, not personal preferences or theoretical loyalty, guide their choices.
How Empiricism Differs From the Other Attitudes
It helps to understand empiricism by seeing where it ends and the other scientific attitudes begin. Determinism is the belief that behavior is lawful and caused by identifiable variables. Experimentation is the process of manipulating those variables to test their effects. Replication asks whether the results hold up when repeated. Parsimony favors the simplest explanation that fits the data. Philosophic doubt means treating all conclusions as tentative and open to revision.
Empiricism sits at the foundation of all of these. You can’t experiment without first having a reliable way to observe and measure. You can’t replicate findings if the original data were based on subjective impressions. You can’t apply parsimony if there’s no objective evidence to evaluate competing explanations against. Empiricism is the attitude that ensures every other scientific practice in ABA has solid ground to stand on.