Antecedent Strategies and Differential Reinforcement

Some behavior plans work by changing what happens before a behavior. Others work by changing what happens after it, so that reinforcement depends on meeting a rule the plan sets. As an RBT (Registered Behavior Technician), you will carry out both kinds of plan. Your supervisor decides which procedure to use. Your job is to run it exactly as written, record what you see, and ask when the plan doesn't cover a situation.

What you will be able to do

• Recognize and carry out three antecedent strategies: NCR, high-probability request sequences and demand fading.
• Tell DRA, DRI, DRO, DRL and FCT apart by the question each one answers, and carry out each one from a written plan.
• Follow a written plan's timing and criteria, including simple time and count arithmetic.
• Record observations rather than interpretations, and report gaps instead of filling them in.

Before you start. A few terms you will need:

• Reinforcement is a consequence that makes a behavior more likely in the future.
• An antecedent is what happens just before a behavior.
• A behavior's function is what it usually gets the person, such as access to an item, or attention, or escape from a task.

The team estimates the function through a functional behavior assessment (FBA). An FBA shows patterns, such as grabbing usually being followed by getting the toy. That pattern supports a hypothesis about the function. It does not prove it.

Your role: run the plan, don't redesign it

Every example below uses one fictional learner, Riley, and a written plan. If the plan leaves out a rule you need, stop and ask your supervisor. Examples of missing rules are what to do if the behavior happens at a scheduled moment, or what to do if you missed part of the session.

Do not add things the plan doesn't say. That includes:

• a timer length
• extinction, which means stopping the consequence that has been keeping the behavior going. Extinction is not the same as ignoring. If yelling gets Riley out of a task, ignoring the yelling while still letting Riley leave the task is not extinction.
• holding back a reinforcer to make it more powerful

Never withhold food, water, bathroom access, medical care or safety measures.

Antecedent strategies: change what comes first

Noncontingent reinforcement (NCR)

Plain meaning: Riley gets the reinforcer on a clock, no matter what Riley is doing.

"Noncontingent" means the reinforcer does not depend on any behavior. Suppose the team hypothesizes that Riley calls out to get attention. If Riley gets attention every few minutes anyway, there is less reason to call out for it.

Demonstration. The plan says: "Start the timer at 2:00. Make the first delivery after 3 minutes, then give 30 seconds of play-attention every 3 minutes until 2:15."

• 2:00 is when the clock starts, not a delivery time.
• From 2:00 to 2:15 is 15 minutes. 15 ÷ 3 = 5, so you make 5 deliveries.
• They happen at 2:03, 2:06, 2:09, 2:12 and 2:15.
• Riley doesn't have to earn any of them.

Try: At 2:09, Riley is calling out. The plan says nothing about this situation. What do you do?

Answer: Ask before it happens

By definition, NCR is not tied to behavior. Whether a plan delays a delivery during the behavior is still the supervisor's decision. Raise the question before the session if you can. Don't make up a rule in the moment.

High-probability (high-p) request sequence

Plain meaning: First give a few easy requests the learner usually follows. Then give the harder request.

The easy requests are called high-probability because Riley is likely to do them. The harder request is low-probability. Following the easy ones builds momentum toward the harder one.

Demonstration. The plan says: three high-p requests, each praised, then the low-p request.

1. "High five." Riley does it. "Nice!"
2. "Touch your head." Riley does it. "Great!"
3. "Hand me the car." Riley does it. "Thanks!"
4. Then the low-p request: "Put the blocks in the bin."

Demand fading, and why it isn't prompt fading

Plain meaning: The task starts small and grows step by step toward the full goal.

Riley's plan starts with "put 2 blocks in the bin." Then it moves to 5, then 10, then all of them. Each move happens only when the supervisor's criterion is met.

Prompt fading is different. The task stays the same, and the help you give shrinks. For example, you might go from guiding Riley's hand to just pointing.

The way to tell them apart:

• If the amount or difficulty asked changes, it is demand fading.
• If the amount of help changes, it is prompt fading.

Try: Riley is on the 5-block step. Your supervisor says, "Stop pointing at the bin." Which kind of fading is that?

Answer: Prompt fading

The task is still 5 blocks, and only the help changed. Moving from 5 blocks to 10 blocks would be demand fading.

Differential reinforcement: reinforce only what meets the plan's rule

"Differential" means the reinforcer depends on whether a rule is met. The rule can be one of three kinds:

• Riley does a chosen replacement behavior (DRA, DRI and FCT).
• The target behavior stays absent for a set time (DRO).
• The target behavior stays at or below a limit (DRL).

Responses that don't meet the rule don't earn the reinforcer. In every type below, the plan says what happens when the target behavior occurs. That might be extinction, or it might be redirection. Use only what the plan specifies.

DRA, DRI and FCT: replacement behaviors

Riley grabs blocks from a peer. The team's hypothesis is that the function is access to the blocks.

DRA (alternative behavior). You reinforce a different, acceptable behavior. For example, you give Riley extra blocks when Riley builds with their own set. Riley could still grab at the same time. DRA doesn't require that the two behaviors be impossible together.

DRI (incompatible behavior). You reinforce a behavior that physically cannot happen at the same moment as the target behavior. Suppose Riley leaves the table during work. Reinforcing sitting in the chair is DRI, because Riley can't sit and wander at the same time. Reinforcing hand-raising would not be DRI, because Riley can raise a hand while standing. That would be DRA.

FCT (functional communication training). You teach a way to communicate that gets the same outcome the behavior was getting. Riley says or signs "my turn," and Riley gets the blocks. If "my turn" got praise but no blocks, it wouldn't serve the same function.

Early on, a plan will often reinforce every correct replacement behavior. Later it thins reinforcement gradually. Follow the schedule the plan gives you.

DRO (other behavior): reinforce the absence of the target behavior

Plain meaning: You deliver a reinforcer when the target behavior does not occur for a set period.

No particular replacement behavior is required. The only requirement is that the target behavior didn't happen.

Demonstration. The plan says: "4-minute DRO starting at 10:00. If grabbing occurs, reset the timer." Riley grabs at 10:03.

• The timer restarts at 10:03.
• 10:03 + 4 minutes = 10:07, so the next possible reinforcer is at 10:07.

Change one fact: the plan instead says "fixed intervals; if grabbing occurs, wait for the next interval."

• The intervals end at 10:04, 10:08, and so on.
• Riley grabbed at 10:03, so there is no reinforcer at 10:04.
• The next possible reinforcer is at 10:08.

Both versions are the same type of procedure, but they produce different times. That is why you follow the plan's exact rule.

DRL (low rates): a lower acceptable rate

Plain meaning: You reinforce when the behavior occurs at or below a set limit.

DRL is used when some of the behavior is acceptable and the goal is less of it. The plan aims for a lower rate, not for elimination, so it doesn't require zero. But zero still counts: 0 is below any limit, so an interval with no occurrences earns the reinforcer under an "at or below" rule. DRL is not a choice for dangerous behavior unless a specialist has decided it is appropriate.

Demonstration. The plan says: "Call-outs: limit of 2 per 10-minute interval."

Which question does each one answer?

• DRA: Did Riley do the chosen alternative behavior?
• DRI: Did Riley do a behavior that can't happen at the same time as the target behavior?
• FCT: Did Riley communicate in a way that gets the same outcome?
• DRO: Did the target behavior stay absent for the whole period?
• DRL: Did the count stay at or below the limit?

Record what you saw

Write observations, not interpretations.

• An observation: "Riley pushed the bin and said 'no' twice."
• An interpretation: "Riley was frustrated."

A gap in observation is not a zero. Suppose you looked away for a minute during a DRO interval. You can't record "no grabbing" for that minute. Note the gap and follow your supervisor's direction about whether to deliver the reinforcer.

Check what a number is measured against. "2 per 10 minutes" is not the same as "2 per 30-minute session."

• 2 ÷ 10 = 0.2 call-outs per minute.
• 2 ÷ 30 ≈ 0.07 call-outs per minute.

The per-session version is about three times stricter.

Practice on your own

Part 1: Classify each example.

a) Riley gets a sticker if 5 minutes pass with no spitting. b) Riley gets a reinforcer if they hum 3 times or fewer per 15 minutes. c) During circle time, the target behavior is lying on the floor (back or stomach on the floor). Riley gets reinforced for sitting on the carpet square with their bottom on the square and back upright. d) The team's FBA hypothesis is that Riley's hitting during clean-up gets Riley out of clean-up. The plan teaches Riley to tap a "break" card, which gets a short break from clean-up. e) Riley gets a reinforcer every 4 minutes, no matter what they are doing. f) A plan increases Riley's math sheet from 3 problems to 6.

Answer: Classification

a) DRO. Only the absence of spitting is required. b) DRL. The goal is a lower acceptable rate. The plan doesn't require zero hums, though zero would still meet the limit. c) DRI. As defined, Riley can't be lying on the floor and sitting upright on the square at the same moment. The exact definitions are what make this DRI. A vaguer pair, such as "holding a ball" versus "hand-flapping," might not be truly incompatible. d) FCT. The team hypothesizes that hitting gets escape from clean-up, and the card gets that same outcome. Without that stated function, a card that gets a break would not be enough to call it FCT. e) NCR. Delivery is based on time, not behavior. f) Demand fading. The amount of work grows. The help doesn't change.

Part 2: Run the plans.

1. DRO. 3-minute reset DRO starting at 1:00. The behavior occurs at 1:02 and again at 1:04. When is the first possible reinforcer?
2. DRL. Limit of 1 per 5-minute interval. The counts are 0, 2 and 1. Which intervals earn the reinforcer?
3. High-p. Riley refuses the second easy request. The plan doesn't say what to do. What do you do?
4. DRA vs. DRI. A plan says "reinforce asking for help" to reduce tearing paper. Is this DRI?

Answer: Plans

1. 1:07. The behavior at 1:02 resets the timer, so the earliest possible time becomes 1:05. The behavior at 1:04 resets it again: 1:04 + 3 minutes = 1:07.
2. Intervals 1 and 3. 0 and 1 are at or below the limit of 1. 2 is above it.
3. Don't make up a response. Note what happened and ask your supervisor.
4. No. Riley can ask for help while tearing paper, so the two behaviors aren't incompatible. This is DRA, or FCT if asking for help gets the same outcome the tearing did.

Part 3: Carry out complete plans. Each plan below includes the rule you need. Decide what you do next.

1. NCR. The plan says: "Start the timer at 9:00. Make the first delivery after 5 minutes, then give 1 minute of tablet time every 5 minutes until 9:20. Deliver on schedule no matter what Riley is doing, including whining." At 9:10, Riley is whining. List every delivery time, and say what you do at 9:10.

Answer: NCR plan

From 9:00 to 9:20 is 20 minutes, and 20 ÷ 5 = 4 deliveries: 9:05, 9:10, 9:15 and 9:20. At 9:10, you give the tablet time on schedule. This plan covers whining, so you follow it. In the earlier checkpoint, the plan was silent, so you had to ask. The difference is whether the plan supplies the rule.

2. High-p. The plan says: "Give three easy requests, praising each one. Then give the low-p request: 'Wash your hands.' If Riley refuses an easy request, start again from the first easy request." Riley claps and stands up when asked, but refuses "Touch your nose." What do you do next? If Riley then follows all three easy requests, what comes next?

Answer: High-p plan

Start the sequence again from the first easy request, because this plan gives that rule. If Riley follows all three, give the low-p request: "Wash your hands." Don't skip ahead to the low-p request after the refusal. The plan builds momentum through three completed easy requests first.

3. Demand fading. The plan says: "Riley sorts socks into pairs. Steps are 3 socks, then 6, then 10, then the full basket. Move to the next step after 2 sessions in a row where Riley finishes the step without shouting. Point to the basket at every step." Riley finished 3 socks without shouting on Monday and again on Tuesday. On Wednesday, Riley shouted once but finished. What step do you use on Wednesday, and what step on Thursday? Does your pointing change?

Answer: Demand-fading plan

Wednesday is 6 socks: Monday and Tuesday make 2 sessions in a row at the 3-sock step. Thursday stays at 6, because Wednesday's shouting means Riley hasn't met the criterion at 6. The plan doesn't say to move back a step, so don't. If you think a step back is needed, ask. Your pointing doesn't change: the plan keeps the help the same and grows only the task.

4. DRA. The plan says: "Target behavior: grabbing a peer's markers. Alternative behavior: coloring with Riley's own markers. After each full minute of coloring with own markers, give praise and one extra marker. If Riley grabs, return the marker to the peer and point to Riley's own set. Grabbing never earns a marker." Riley starts coloring with their own markers at 10:00, is still coloring at 10:01, and grabs a peer's marker at 10:01:30. What do you do at 10:01 and at 10:01:30?

Answer: DRA plan

At 10:01, give praise and one extra marker, because Riley completed a full minute with their own markers. At 10:01:30, return the marker to the peer, point to Riley's own set, and give no marker. This is DRA, not DRI, because Riley could hold their own marker and grab with the other hand.

5. DRI. The plan says: "Target behavior: lying on the floor during circle time (back or stomach on the floor). Incompatible behavior: sitting on the carpet square with bottom on the square and back upright. Check at minutes 2, 4 and 6. Give a token at each check where Riley is sitting this way. If Riley is lying down at a check, say 'Sit on your square' and point, with no token for that check." At minute 2, Riley is sitting upright. At minute 4, Riley is lying on their stomach. At minute 6, Riley is sitting upright. How many tokens do you give, and what do you do at minute 4?

Answer: DRI plan

Two tokens, at minutes 2 and 6. At minute 4, say "Sit on your square," point, and give no token. The definitions make this DRI: Riley can't have their stomach on the floor and be sitting upright on the square at the same moment.

6. FCT. The plan says: "The team's FBA hypothesis is that Riley's screaming during math worksheets gets escape from the worksheet. Replacement: Riley hands you a 'break' card. Each time Riley hands you the card, give a 1-minute break away from the worksheet right away, then return to the worksheet. If Riley screams, keep the worksheet in place and point to the break card." Riley screams. Ten seconds later, Riley hands you the card. What do you do after each one?

Answer: FCT plan

After the scream, keep the worksheet in place and point to the card. Screaming no longer gets escape. After the card, give the 1-minute break right away, then return to the worksheet. This is FCT because the card gets the same outcome, escape, that the team hypothesizes screaming was getting. If you notice that Riley often screams just before handing over the card, record what you saw and tell your supervisor. Don't change the plan yourself.

What to remember

• Antecedent strategies change what comes before a behavior.
  • NCR delivers reinforcement on a clock. Check when the clock starts and when the first delivery is due.
  • High-p sequences start with easy requests.
  • Demand fading grows the task, while prompt fading shrinks the help.
• Differential reinforcement delivers the reinforcer only when the plan's rule is met.
  • DRA reinforces an alternative behavior.
  • DRI reinforces a behavior that truly can't happen at the same time, as the plan defines both behaviors.
  • FCT reinforces communication that gets the same outcome the team identified for the behavior.
  • DRO reinforces absence.
  • DRL targets a lower acceptable rate rather than requiring absence. Zero may still meet an at-or-below limit.
• Extinction means stopping the consequence that has kept a behavior going. It is not the same as ignoring. Use it only when the plan includes it.
• Implement, record and ask. Your supervisor selects the procedure. You follow its exact rules. Gaps get reported, not guessed.

Sources

• Sam, A., & AFIRM Team (2024). Antecedent-Based Interventions. UNC Frank Porter Graham Child Development Institute.
• Savage, M., & AFIRM Team (2024). Differential Reinforcement. UNC FPG. (DRA, DRI, DRO and DRL definitions; interval DRL "equal to or below" criterion; extinction glossary definition.)
• Griffin, W., & AFIRM Team (2025). Functional Communication Training. UNC FPG. (Same-function replacement; blocking access to the outcome after the interfering behavior.)
• Morin, K., & AFIRM Team (2024). Extinction (NCR definition). UNC FPG.
• Penrod, B., Gardella, L., & Fernand, J. (2012). Journal of Applied Behavior Analysis, 45(3), 527–537 (high-p and demand-fading definitions).
• Behavior Analyst Certification Board. RBT Test Content Outline (3rd ed.); RBT Ethics Code (2.0).

Independent study resource

This study resource is not affiliated with, sponsored by, or endorsed by the Behavior Analyst Certification Board (BACB). BACB, Registered Behavior Technician and RBT are registered trademarks of the Behavior Analyst Certification Board. Source citations do not imply endorsement by their authors or publishers.