Calculating and summarizing data
Summarizing data means turning raw tallies, times and trial marks into a single number your supervisor can compare across sessions, such as a rate per minute or a percentage correct. The arithmetic is simple; the skill is picking the calculation that matches the measure and the data sheet.
What calculations does task A.6 include?
Task A.6 asks you to calculate and summarize data. The common calculations are rate (count divided by time), percentage correct (correct trials divided by total trials, times 100), percentage of intervals, mean count per session and mean duration per episode. Your supervisor states which summary each program uses, and you apply it consistently.
| Summary | Formula | Worked example |
|---|---|---|
| Rate | Count / minutes (or hours) observed | 45 requests in 15 minutes = 3 per minute |
| Percentage correct | Correct / total trials x 100 | 18 of 24 = 75% |
| Percentage of intervals | Scored intervals / total intervals x 100 | 12 of 40 = 30% |
| Mean count | Sum of counts / number of sessions | 4, 7, 3, 6, 5 = 25 / 5 = 5 |
| Mean duration | Total minutes / number of episodes | 2, 6, 4, 12 min = 24 / 4 = 6 minutes |
When should you use rate instead of a count?
Use rate whenever observation lengths differ. A count of 24 in 40 minutes and 18 in 20 minutes looks like improvement, but the rates are 0.6 and 0.9 per minute, so the behavior actually became more frequent. Converting to rate puts sessions of different lengths on the same footing so comparisons are fair.
Home example: a learner's rate of requests during a 20-minute snack routine is compared with a 10-minute play routine. Only the rate tells the supervisor which routine produces more requesting per minute.
Write the time unit with every rate. "0.6" means nothing on its own; "0.6 per minute" or "36 per hour" does. Exam options sometimes give the correct number with the wrong unit.
How do you calculate percent correct when prompts are involved?
Follow the data sheet's rule for prompted trials. Most programs score only unprompted correct responses as independent. If a learner answers 6 of 10 independently, 3 more after a prompt and misses 1, percent independent is 60%, not 90%. Counting prompted answers as correct inflates progress and can lead to targets being marked mastered too early.
Summarizing a session, step by step
Worked scenario: a learner tantrums four times in a day for 2, 6, 4 and 12 minutes. Total duration is 24 minutes; mean duration per episode is 6 minutes. Notice that the single 12-minute episode pulls the mean up, which is worth mentioning to your supervisor if it was unusual.
- Check the sheet is complete: start and end times, every trial or interval marked.
- Identify the measure: count, duration, trials, intervals or products.
- Choose the summary the program specifies.
- Do the arithmetic once, then check it a second way, such as multiplying back.
- Round only as the agency directs, usually to one decimal place or a whole percent.
- Record the summary with its unit and note anything that limits it, such as a shortened session.
Which arithmetic mistakes appear on exam questions?
Common errors include dividing time by count instead of count by time, forgetting to multiply by 100 for a percentage, using the wrong total such as correct trials over incorrect trials, averaging percentages from sessions with very different trial numbers without noting it, and counting prompted responses as independent. Many wrong options are the result of one of these slips.
- Rate is responses per unit of time: count on top, time underneath.
- Percentage needs the full total of trials or intervals as the denominator.
- Mean is the sum divided by how many values you added.
What is the technician's role compared with the supervisor's?
Your supervisor decides which summary represents each target, sets mastery criteria and interprets what the numbers mean for treatment. You do the calculations accurately, show your working when asked, flag sessions where the number may mislead, such as a 10-minute session that was cut short, and never adjust a figure to make it look more consistent.
Practice questions
During a 15-minute play session, Isla makes 45 vocal requests. What is her rate of requests per minute?
Theo has four tantrums in a day lasting 2, 6, 4 and 12 minutes. What is the mean duration per tantrum?
On Monday, Sofia bites her nails 24 times in a 40-minute session. On Wednesday, she bites them 15 times in a 20-minute session. Which statement is accurate?
Over 10 trials, Caleb answers correctly and independently on 6, correctly after a prompt on 3, and incorrectly on 1. His plan counts only independent correct responses. What percentage should be recorded?
During a 30-minute session a technician tallies 12 instances of hand-raising. In a 60-minute session the next day she tallies 18. Which measure lets her compare the two sessions fairly?
Questions people ask
Do I need a calculator for the RBT exam?
The arithmetic involved is basic. Check the current BACB and testing-centre rules on what you may bring rather than assuming.
Is a median ever used instead of a mean?
Sometimes, when one extreme value would distort the mean. Your supervisor decides; you report the summary the plan asks for.
Should I round rates?
Follow your agency's rule. Rounding inconsistently between sessions can make small changes look larger or smaller than they are.
How ready are you?
Ten questions, about 3 minutes, no account. Get your readiness score and your weakest domains.
Check my readiness