Preference Assessments and Skill Assessments

A preference assessment asks: Which of these items or activities does this person tend to choose, or stay with? A skill assessment asks: What can this person already do, and what might they need to learn? As a technician, you run or help run both kinds of assessment by following your supervisor's written directions. Your main job is to record exactly what happened.

What You Will Be Able to Do

• Explain why a preferred item is not automatically a reinforcer.
• Run and score free-operant, paired-stimulus, MSW and MSWO preference assessments, and contrast them with single-stimulus presentation.
• Record refusals, "no selection" trials and stopped sessions accurately.
• During a skill assessment, tell an independent response from a prompted one, follow the directions for prompts and reinforcement, and report unusual conditions.
• Describe what technicians do not do: diagnose, choose assessment tools or interpret results.

Before you start: You need basic division and percentages. Each calculation is shown step by step.

Preferred Is a Prediction, Not Proof

A preference is what the client picks or engages with when they have options. (The client is the person receiving services.) A reinforcer is a consequence that follows a behavior and makes that behavior more likely in the future. Reinforcement can work in more than one way. In this lesson, the reinforcers we discuss are items or activities the client gets after a behavior.

A preference assessment produces a preference hierarchy: a ranked list of items, from most chosen to least chosen. The list tells you which items may work as reinforcers. It does not prove that they will.

Example: when a toy whistle is freely available, the client picks it every time. Later, the client can earn the whistle by sorting cards, but their card sorting does not increase. The whistle is preferred. It has not been shown to reinforce card sorting.

Preferences also change. A client who has just had a lot of time with an item may want it less for now. This is called satiation. For this reason, your supervisor may have you repeat assessments regularly and also check preferences informally during the day.

The Methods

Every method below answers the same question: what does the client prefer? Each method gets the answer differently and records something different. Your supervisor's protocol sets these details:

• the item list
• where each item is placed
• how long the client gets with a chosen item
• when to stop

Follow that protocol rather than a number you saw somewhere else.

Free-operant observation

The client has free access to a set of items, and nothing is taken away. You record duration: how long the client engages with each item.

Example: in a 10-minute observation, the client spends 6 minutes with blocks and 3 minutes with music. They spend 1 minute not engaged with anything. Record that minute as "no engagement." It is real data. Do not drop it, and do not assign it to an item.

Contrast: single-stimulus. Here you offer one item at a time. You record whether the client takes it and how long they engage. Because only one item is offered, you learn whether the client accepts that item. You do not learn whether they would choose it over another item.

Paired-stimulus

You present two items at a time and record which one the client selects, or "no selection." Every item is eventually paired with every other item.

Worked example with four items (drum, car, book, ball):

• The possible pairs are drum–car, drum–book, drum–ball, car–book, car–ball and book–ball. That makes 6 trials.
• Each item meets each of the other 3 items once, so each item is presented 3 times.
• Switch which item sits on the left and which on the right, as the protocol directs. Otherwise a habit of reaching to one side could look like a preference.

If your protocol scores a percentage, the formula is: selections ÷ times that item was presented × 100.

• Drum: 3 ÷ 3 × 100 = 100%
• Car: 2 ÷ 3 × 100 ≈ 67%
• Book: 1 ÷ 3 × 100 ≈ 33%
• Ball: 0 ÷ 3 × 100 = 0%

Check your work: 3 + 2 + 1 + 0 = 6 selections, one for each of the 6 trials. That match only happens if the client chose something on every trial. If a trial ended with no selection, record "no selection" for that trial. Do not credit that choice to either item.

Multiple-stimulus with and without replacement

You place several items in front of the client, called an array. The client picks one and gets brief access to it. Then the two methods differ:

• MSW (with replacement): The chosen item goes back into the array, and the positions are rearranged. Items chosen more often across trials rank higher. Trials continue until the point your protocol sets, such as a set number of trials, or until the client stops choosing.
• MSWO (without replacement): The chosen item stays out, and the remaining items are rearranged. Items chosen earlier rank higher. The assessment ends when all items have been chosen or when the client stops choosing.

Some MSW protocols also tell you to swap an item that was not chosen for a new item. Do this only if your protocol says to.

MSW worked example: count every trial

Your protocol says: use a 3-item array (slinky, stamp, kazoo), run 6 trials, return the chosen item after access, and rearrange positions each trial. No item swaps.

The client's choices:

1. Slinky
2. Slinky
3. Stamp
4. Slinky
5. No selection
6. Stamp

After every choice, the chosen item goes back in. So all three items are in front of the client on every trial. Count each item's selections:

• Slinky: 3 (trials 1, 2 and 4)
• Stamp: 2 (trials 3 and 6)
• Kazoo: 0
• No selection: 1 (trial 5)

Check your work: 3 + 2 + 0 + 1 = 6, one record for each trial. Slinky ranks highest because it was chosen most often.

The kazoo's 0 is a real zero. It was offered on all 6 trials and never chosen. Trial 5 stays "no selection." It is not a choice of any item.

Tempting mistake: leaving the slinky out after the client picks it. That is the MSWO rule. In MSW, the slinky must go back, or the client can no longer show that they would choose it again.

Checkpoint: Your MSW protocol uses a 4-item array (puppet, bead maze, stacking cups, rain stick), 8 trials, and says to return the chosen item and rearrange positions. On trial 3, the client picks the puppet. (a) What do you do with the puppet before trial 4? (b) Across all 8 trials, the choices are: puppet, cups, puppet, puppet, bead maze, cups, puppet, cups. What is each item's selection count, and which ranks highest?

Answer: Checkpoint 1

(a) After the client's access time, put the puppet back in the array and rearrange all four positions. Keeping it out would turn this into an MSWO assessment.

(b) Puppet: 4 (trials 1, 3, 4 and 7). Stacking cups: 3 (trials 2, 6 and 8). Bead maze: 1 (trial 5). Rain stick: 0. Check: 4 + 3 + 1 + 0 = 8 trials. The puppet ranks highest. The rain stick's 0 is real data because it was offered on every trial.

MSWO: when the client stops choosing

Tempting mistake: in MSWO, if the client stops choosing with items still left, those items are not selected. They are not "tied for last." Record what actually happened.

Checkpoint: An MSWO assessment uses five items. The client picks a puzzle, then stickers, and then stops reaching for anything. A coworker says, "Rank the other three as 3, 4 and 5 in any order." What should you record?

Answer: Checkpoint 2

Record the puzzle as first and the stickers as second. Record the remaining three as not selected, and note that the client stopped making selections. Giving them made-up ranks would turn missing choices into invented data.

Respect Refusal and Safety

• Never force a choice. Do not guide the client's hand to an item.
• If the client walks away or stops participating, let them stop the assessment. Do not physically hold or block them just to finish it. This does not override safety. If leaving the area could put the client in danger, follow the client's approved safety or emergency procedures.
• Do not withhold food, drink or needed comfort to make the items seem more wanted. Do not decide on your own to keep items away from the client to raise their interest.
• Some protocols rotate items or keep certain items out of view, often to prevent satiation. That is part of the supervisor's plan, so follow it. If you are unsure whether limiting access to an item is part of the plan, ask your supervisor.
• If the client shows distress or challenging behavior, stop as the protocol directs and note what happened. Offer the assessment again later only if your supervisor directs it.
• Use only items your supervisor has approved. Approval accounts for risks such as allergies.

Record Observations, Not Interpretations

An observation is what you saw: "Held the drum for 40 seconds." An interpretation is what you think it means: "Loves drums." Record the observation.

Missing is not zero. Suppose a fire drill stops the session after trial 4. Trials 5 and 6 were not run. Recording them as "not selected" would claim the client refused items that were never offered.

Check what each number is out of. 4 minutes of engagement in a 10-minute observation is 40% (4 ÷ 10 × 100). The same 4 minutes in a 5-minute observation is 80% (4 ÷ 5 × 100). The minutes are identical, but the totals they are measured against differ. In the same way, "3 selections" could mean 3 out of 3 presentations or 3 out of 6 trials. Always report the denominator: the total that the number is out of.

Checkpoint: Two reports say "Puzzle: 50%." One is 5 minutes out of 10. The other is 2 selections out of 4 presentations. Do they show the same preference?

Answer: Checkpoint 3

No. One measures how long the client engaged, and the other measures how often they chose the puzzle. Each is out of a different total. Report the method and both numbers, not just the percentage.

Skill Assessments: A Different Question

A preference assessment finds out what the client chooses. A skill assessment finds out what the client can do: their strengths (skills they have) and deficits (skills they have not yet shown). You may help with three kinds:

• Curriculum-based: An assessment built from the client's own curriculum, meaning the skills the program is teaching. It checks what the client can do with those skills. One common form is progress monitoring: short probes repeated over time to track growth toward a goal. Each probe uses a different version of similar difficulty. Example: each week, the client reads a different list of sight words at the same level.
• Developmental: A tool that looks at skills across areas such as communication or self-care. Your supervisor selects and interprets it.
• Social-skill: Observable social tasks that the protocol defines. Example: "When a peer says 'your turn,' the client takes a turn in the game."

Your role is to present items exactly as written and record responses. These assessments are not diagnoses. Even formal screening tools do not produce a diagnosis, and neither do your observations.

Independent versus prompted

An independent response is correct with no help beyond the standard instruction. A prompted response follows extra help, such as pointing, repeating the instruction, hinting or modeling. Even a look toward the correct answer can act as a prompt.

Example: you say, "Point to the cat." The client hesitates, you glance at the cat card, and then they point to it. Your glance may have shown them the answer, so you cannot count the response as independent. Score it the way the assessment directions say, and report the glance.

Giving hints so the client "gets a fair chance" inflates the score. An inflated score can lead the team to skip teaching the client actually needs. Also report anything unusual, such as illness, a new room, a loud environment or a stopped session.

Checkpoint: During a social-skill probe, the peer says "your turn," and the client does not respond. You whisper, "It's your turn," and the client takes a turn. How do you record it?

Answer: Checkpoint 4

Record it as prompted, following the directions, and note the whispered prompt. Recording it as independent would hide the help you gave.

Follow the reinforcement and feedback rules too

Prompts are not the only thing the directions control. They also say what you do after a response: whether you praise, correct, give an item or activity, or say something neutral. Follow those rules exactly, even when adding something feels kind or encouraging.

Example: the directions for a vocabulary probe say to say "Thank you" after every answer, with no praise or correction. The client calls a picture of a dog "cat." You want to say, "Good try! That's a dog." Don't. Telling them the answer could help them on the next item. Then the results would show what they did with your teaching, not what they could already do. Saying only "Thank you" follows the directions.

The rule is "follow the directions," not "never reinforce." Some assessments do direct you to give praise or an item after certain responses. Then you give exactly what is written, when it is written. If the directions are unclear, ask your supervisor before the session.

Checkpoint: A curriculum-based probe's directions say to give no tangible rewards during the probe. The client seems bored, so a coworker suggests giving a sticker after each correct answer "to keep them motivated." What should you do, and why?

Answer: Checkpoint 5

Do not add the stickers. Follow the directions, and if you are concerned about the client's motivation, tell your supervisor and note it. Adding rewards changes the conditions of the assessment. The client might respond differently than they would under the written directions, so the results would not mean what the team thinks they mean. Keeping the directions is not coldness. It is what makes the data trustworthy.

Checkpoint: After a developmental assessment, a caregiver asks you, "So does this mean my child has a disability?" What do you do?

Answer: Checkpoint 6

Do not interpret the results or offer a diagnosis. Kindly say that the supervisor reviews assessment results, and pass the question to your supervisor. Directing questions about your services to your supervisor is part of the RBT Ethics Code.

Practice on Your Own

A paired-stimulus assessment uses three items: bubbles, crayons and a spinner. That makes 3 trials, and each item is presented twice. The results:

• Bubbles vs. crayons: the client chooses bubbles.
• Bubbles vs. spinner: no selection.
• Crayons vs. spinner: the client chooses the spinner.

Give each item's percentage and say what you would report.

Answer: Practice

• Bubbles: 1 ÷ 2 × 100 = 50%
• Crayons: 0 ÷ 2 × 100 = 0%
• Spinner: 1 ÷ 2 × 100 = 50%

Report the one "no selection" trial along with the percentages. Do not treat it as a choice for either item, and do not drop it. The scores only show a preference. Whether bubbles or the spinner actually reinforce a behavior still has to be shown.

Remember

• Preferred items may be reinforcers. Only an increase in behavior shows that an item reinforces that behavior.
• Free-operant and single-stimulus methods record duration. Paired-stimulus and MSW count selections. In MSW, the chosen item goes back in. MSWO ranks items by the order they were chosen, and chosen items stay out.
• Follow the protocol for items, positions, access, item rotation and stopping. Never force a choice. Let the client leave the assessment while still following their approved safety procedures. Never withhold food, drink or needed comfort.
• "No selection," "not run" and "zero" are different records.
• In skill assessments, follow the directions for both prompts and reinforcement, record prompted responses as prompted, report unusual conditions, and leave interpretation and diagnosis to qualified professionals.

Sources

• Vanderbilt Kennedy Center TRIAD. (2024). Preference Assessments: An Introduction for Educators. https://vkc.vumc.org/assets/files/triad/tips/preference_assessment.pdf
• Vanderbilt IRIS Center. Progress Monitoring, Page 2. https://iris.peabody.vanderbilt.edu/module/pmm/cresource/q1/p02/
• Centers for Disease Control and Prevention. Clinical Screening for Autism Spectrum Disorder. https://www.cdc.gov/autism/hcp/diagnosis/screening.html
• Deliperi, P., et al. (2015). Training staff to implement a paired-stimulus preference assessment using video modeling with voiceover instruction. Behavioral Interventions. DOI 10.1002/bin.1421. https://www.caldwell.edu/wp-content/uploads/2021/07/1642cf7ea97ea80ebd5abc5e655934d3.pdf
• Behavior Analyst Certification Board. (2023). RBT Test Content Outline (3rd ed.), B.1–B.2. https://www.bacb.com/wp-content/rbt-outline-3rdEd/
• Behavior Analyst Certification Board. (2021). RBT Ethics Code (2.0), 1.06, 2.01, 2.02, 2.06. https://www.bacb.com/wp-content/rbt-ethics-code

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.