Best For
Teams this article is built to help
Category: FBA & Data Collection
Evidence
What backs this guide
This entry reads as practice guidance rather than a source-cited research summary.
Materials
What you can leave with
- Condensed key takeaways
- Primary downloadable resource
- 3 bonus downloads
- Interactive self-check quiz
You're counting tantrums. But should you be timing them instead? A tantrum that happens once but lasts 45 minutes is very different from 5 tantrums lasting 2 minutes each. Yet simple frequency data shows them as "1 incident" vs "5 incidents", making the less severe situation look worse.
Why Measurement Type Matters
According to Ledford & Gast (2018), the equation is simple but critical:
Wrong measurement = Incorrect interpretation = Failed intervention
Example: The Hidden Problem
Tantrum frequency decreased from 3/day to 1/day. Success? Maybe not.
Before
3 tantrums ร 5 min each = 15 min total
After
1 tantrum ร 45 min = 45 min total
Frequency shows improvement. Duration reveals the problem got 3x worse.
The Three Core Dimensions
๐ข
Frequency
How OFTEN
โฑ๏ธ
Duration
How LONG
โก
Latency
How QUICKLY
The Research-Based Rule
Your measurement must match the behavior's most problematic feature.
Frequency Data: When to Use It
Definition: Count of how many times behavior occurs.
Best For (Pustejovsky et al., 2022):
- Brief behaviors (under 1-2 minutes)
- Clear beginning and end
- Equal opportunity to occur
- The problem is it happens too often
Perfect for Frequency
- โ Hitting
- โ Calling out
- โ Throwing objects
- โ Swearing
- โ Getting out of seat
Poor Choice for Frequency
- โ Tantrums (may last 30 sec or 30 min)
- โ "Off-task" (no clear start/stop)
- โ Arguing (when does one end?)
- โ Crying episodes
- โ Task engagement
The Equal Opportunity Rule
Behavior must have equal chance to occur across sessions. Otherwise, calculate RATE.
Example: 6 call-outs in 30-min math vs. 2 in 10-min bathroom break
Looks different? Calculate rate: Both = 0.2 per minute. Same rate.
Duration Data: When to Use It
Definition: How long behavior continues (start to stop).
Best For:
- The problem is how long it lasts, not how often
- Continuous or extended behaviors
- Intensity correlates with duration
- Goal is shortening, not eliminating
Tantrums: The Classic Duration Example
The large majority of tantrum studies use duration measurement.
Why: A 2-minute tantrum vs. 45-minute tantrum = vastly different problems. Frequency misses this distinction entirely.
Duration Subtypes
Total Duration
Add up all instances
Ex: 5min + 12min + 3min = 20min
Best for: Comparing days/weeks
Average Duration
Total รท frequency
Ex: 20min รท 3 = 6.7min avg
Best for: Typical episode length
% of Session
(Total รท session) ร 100
Ex: 20min in 60min = 33%
Best for: Impact on learning
Other Duration-Appropriate Behaviors
- Time-out effectiveness - Duration predicts success
- Crying episodes - Standard measurement
- Task engagement - How long sustained attention lasts
- Stereotypy/stimming - Continuous behavior, duration shows severity
Latency: The Overlooked Measurement
Definition: Time between stimulus (demand/prompt) and behavior response.
Best For:
- Measuring compliance/non-compliance
- Assessing skill fluency
- Evaluating prompt effectiveness
- Intervention goal is faster responding
Following Directions: Why Latency Matters
Latency-to-comply can predict long-term outcomes.
"Clean up toys"
Student A: Starts in 2 seconds
Student B: Starts in 5 minutes
Both eventually comply. But latency shows severity: a longer delay before complying can signal more entrenched non-compliance.
Latency vs. Duration Confusion
Common mistake: Confusing these two measurements.
Latency
How long to START
Teacher says "sit down" โ 15 sec until student starts moving
Duration
How long behavior LASTS
Once seated โ how long student remains seated
The 5-Second Rule
If child initiates compliance within 5 seconds = Compliance. Over 5 seconds = Non-compliance (even if they eventually comply). Why: The delay itself is the problem behavior.
Other Latency-Appropriate Behaviors
- Transition time - "Line up" โ Time until in line
- Requesting help - When frustrated, how long until asks for help?
- Homework initiation - Assignment given โ Time until starts working
Time Sampling Methods
For behaviors with no clear start/stop or that are too frequent to count:
Momentary Time Sampling
Observe at fixed intervals; record if behavior is occurring at THAT MOMENT.
Best for: On-task, stereotypy, fidgeting
Accuracy: Shorter intervals are substantially more accurate than longer ones.
Partial Interval Recording
Mark if behavior occurred AT ANY POINT during interval.
Best for: High-frequency behaviors you want to reduce
Note: Tends to overestimate the occurrence of behavior.
Whole Interval Recording
Mark only if behavior occurred for ENTIRE interval.
Best for: Behaviors you want to increase (engagement, on-task)
Note: Tends to underestimate the occurrence of behavior.
Intensity Scales
When severity varies more than frequency:
Example: Tantrum Intensity Scale
1: Whining, mildly upset
2: Loud crying, pacing
3: Throwing items, kicking furniture
4: Physical aggression toward objects/self
5: Aggression toward others, property destruction
Decision-Making Flowchart
This decision tree helps guide you to an appropriate measurement choice.
Start: What makes this behavior a problem?
Problem: Happens too often
โ Is behavior brief (<2 min)? โ YES: Use FREQUENCY
โ NO: Consider duration
Problem: Lasts too long
โ Use DURATION
โ Also track frequency for complete picture
Problem: Takes too long to respond
โ Use LATENCY
โ Also track frequency of eventual compliance
Problem: Continuous, no clear start/stop
โ Use TIME SAMPLING
โ Momentary (estimate duration), Partial (catch all), Whole (conservative)
Problem: Severity varies greatly
โ Use INTENSITY SCALE
โ Also track frequency to see how often severe vs. mild
Not sure?
โ Start with FREQUENCY for 2 weeks
โ Review data, add second measurement if needed
Common Mistakes to Avoid
Mistake #1: Measuring the wrong dimension
Fix: Ask "What makes this behavior a problem?" Measure THAT dimension.
Mistake #2: Inconsistent session lengths
Fix: Always calculate RATE (per minute, per hour). Rate is the gold standard.
Mistake #3: Not defining behavior boundaries
Fix: Operational definition with start/stop criteria. Rule of thumb: 30-second gap = new instance.
Mistake #4: Starting with multiple measurements
Fix: Start with ONE dimension for first 10-15 data points. Add only if needed (Pustejovsky et al., 2022).
Making It Practical
The teacher reality: Can't carry stopwatch, can't stop teaching to record, can't remember exact times later.
Frequency
Low-tech: Tally on index card, golf counter
Digital: One-tap Quick Log
Time needed: 3-5 seconds per instance
Duration
Low-tech: Note start/stop times (hard to remember)
Digital: Start/stop timer with one tap
Time needed: 5 sec to start, 3 sec to stop
Latency
Low-tech: Nearly impossible while teaching
Digital: Timer starts at demand, stops at response
Time needed: 10 seconds total
Time Sampling
Low-tech: Timer vibrates every 5 min, mark yes/no
Digital: Auto-reminders, one-tap yes/no
Time needed: 2 seconds per interval
The Efficiency Research
Hott et al. (2021): Digital data collection is 72% faster than paper. More importantly: data is immediately available for analysis vs. waiting to "have time" to review paper logs.
The Bottom Line
Choosing the right measurement type is as important as collecting data at all. Match measurement to the behavior's most problematic dimension.
Frequency isn't always the answer (despite being the most common default). Duration reveals severity for extended behaviors. Latency uncovers non-compliance hidden by eventual compliance.
When in doubt: Start with frequency, add measurements only if needed. Modern tools make all measurement types practical for busy teachers.
References
Ledford, J. R., & Gast, D. L. (2018). Single case research methodology: Applications in special education and behavioral sciences (3rd ed.). Routledge. https://doi.org/10.4324/9781315150666
Cooper, J. O., Heron, T. E., & Heward, W. L. (2020). Applied behavior analysis (3rd ed.). Pearson.
Rooker, G. W., DeLeon, I. G., Borrero, C. S., Frank-Crawford, M. A., & Roscoe, E. M. (2015). Reducing ambiguity in the functional assessment of problem behavior. Behavioral Interventions, 30(1), 1-35. https://doi.org/10.1002/bin.1400
Pustejovsky, J. E., & Swan, D. M. (2022). Single-case effect sizes: Developments and directions. Psychological Methods, 27(1), 1-22. https://doi.org/10.1037/met0000406
Hott, B. L., Jones, B. A., Rodriguez, J., & Brigham, F. J. (2021). Technology-based progress monitoring for special education: A systematic review. Journal of Special Education Technology, 36(2), 116-127. https://doi.org/10.1177/0162643419893702
Put This Into Practice
Turn the article into action with ready-to-use materials. Downloads stay open; emailed resources and quiz results are opt-in.
Key Takeaways
- Wrong measurement = incorrect interpretation = failed intervention (Ledford & Gast, 2024)
- Frequency is best for brief, discrete behaviors; duration for extended behaviors where length matters
- Latency (time to respond) reveals non-compliance that frequency data misses
- 96% of tantrum studies use duration measurement, not frequency (Gongola & Daddario, 2022)
- Following a systematic decision tree results in appropriate measurement 94% of the time
Complete Measurement Selection Bundle
A comprehensive toolkit including: Decision flowchart for selecting measurement methods, Direct Observation Methods Quick Reference Table (all 8 methods), IOA Calculation Worksheets with practice problems, and Pocket Cards for quick reference.
Optional email delivery is only for sending this resource and educator updates if you choose them.
Bonus Materials
Clean downloads to pair with this article
Download resources directly or send a copy to your inbox. Marketing updates are optional.
Direct Observation Methods Reference
IOA Calculation Worksheets
Measurement Selection Pocket Cards
Frequency, Duration, or Latency: Do You Know Which One to Use?
Find out if you're measuring behaviors correctly or accidentally making interventions look like failures (or successes, oops).
Tags:
For classroom teams
Put this into practice with your own data
Log behaviors as they happen, see the pattern on a graph, and bring a clear summary to your next team meeting.
Try All Measurement Types FreeFree for up to 3 students. No credit card required.
About the Author
Practical guides from the Classroom Pulse team for educators and behavior support teams who collect and use behavior data. Classroom Pulse was founded by Bobby Hillman, M.Ed., LBA, BCBA.
About Classroom PulseRelated Articles
Writing IEP Progress Reports for Behavior Goals: What the Data Should Say
A six-part narrative for IEP progress reports on behavior goals: baseline, current level, trend, measurement, whether progress is sufficient, and what changes next, with examples.
Before the FBA Referral: A Four-Week Data Packet Classroom Teams Can Build
What a classroom team can gather in about four weeks so an FBA referral starts from evidence: a definition, baseline, ABC entries, a time-of-day grid, and what was tried.
From 'He Always Melts Down After Lunch' to Usable ABC Data
Five real-sounding hallway anecdotes translated step-by-step into ABC log entries, and how a few weeks of honest entries turn instinct into a testable pattern.
