Research Cluster

Quantitative Models and Data Fit

This cluster shows how to check if math rules really match real behavior. It tells BCBAs the best ways to count responses and pick the right numbers to study. When the numbers fit well, we can predict how choices will change and make better plans. The papers give simple steps to test models so practitioners can trust their graphs and decisions.

161articles
1959–2026year range
5key findings
Key Findings

What 161 articles tell us

  1. Using response rate alongside percentage correct reveals implementation drift that percentage summaries alone can conceal.
  2. Mastery criteria based on a learner's individual baseline are more accurate than flat standards like 80% or 100%.
  3. Relational databases allow behavior analysts to store, query, and reproduce analyses in ways spreadsheets cannot support.
  4. Q-learning algorithms fed real-time operant data can predict a person's next response with very high accuracy.
  5. Equivalence classes formed in lab conditions may not hold up in natural, noisy environments, so probe untrained relations in context.
Free CEUs

Get 60+ CEUs Free in The ABA Clubhouse

Live CEU every Wednesday — ethics, supervision, and clinical topics. Always free.

Join Free →

Frequently Asked Questions

Common questions from BCBAs and RBTs

A quantitative model is a math equation that describes how behavior relates to variables like reinforcement rate or effort. Common examples include the matching law, demand curves, and mastery criteria formulas. Models help you predict what will happen when you change a variable.

If the model does not fit, your predictions will be wrong. For example, a demand curve that fits poorly will give you inaccurate estimates of how sensitive behavior is to changes in cost. Always check your model fit before reporting results.

Start by measuring baseline accuracy for the skill. Then set a mastery threshold that represents a meaningful improvement over that baseline. Free software like MIEBL does this calculation for you and adjusts based on your target mastery level.

Procedural fidelity is how closely a session matches the written protocol. You can measure it as a percentage of steps completed correctly, but adding response rate data gives a more sensitive measure that catches problems percentage alone misses.

AI tools that predict client responses are mainly research-stage right now, but they are becoming more accessible. For everyday practice, relational databases and structured data collection are practical ways to improve the quality of your analysis.