When to use it
A multiple-stimulus-without-replacement (MSWO) assessment shows the learner several items at once. The learner picks one, and that item is not put back. The learner picks again from the items left, and so on until all items are chosen or the learner stops choosing. Repeating this across a few sessions produces a ranked list of likely reinforcers.
Use an MSWO when:
- The learner can scan an array of several items and make a choice by reaching, pointing, or using another reliable selection response.
- You need a ranked list, not just a single “favorite.”
- You want an assessment that takes less time than a paired-stimulus (paired-choice) assessment with the same number of items.
Consider another method when:
- The learner can’t scan several items or has a position bias. A paired-stimulus or single-stimulus assessment may work better.
- Removing items reliably leads to problem behavior. A free-operant assessment, where no items are removed, may be a better fit.
A preference assessment identifies likely reinforcers. Only a reinforcer assessment, or seeing that responding increases when the item is delivered contingently, confirms that an item works as a reinforcer. Preferences change, so repeat brief assessments often.
How to fill it in
- Choose the items. Most MSWO procedures use 5 to 8 items. Pick them from caregiver interviews, direct observation, or an earlier assessment. Keep edibles and leisure items in separate assessments. Edibles are often chosen first and can hide leisure preferences. Check every edible against the learner’s allergies, diet, and family preferences before offering it.
- Give each item a short label (A, B, C, and so on) and write it in the item key.
- Give pre-session exposure. Before the first session, let the learner briefly sample each item, such as a small taste or about 30 seconds with a toy, so they know what each one is.
- Set up the array. Place all items in a line, evenly spaced, within easy reach. Randomize the starting order.
- Deliver the instruction (“Pick one”) and wait. Many protocols allow about 30 seconds for a selection. Follow your program’s rule.
- When the learner selects an item, record the item in the next trial row and let them eat it or play with it briefly (about 20 to 30 seconds for leisure items, a small bite for edibles). If the learner reaches for two items at once, block gently and represent the array.
- Remove the selected item. Then rearrange the remaining items. A common method moves the item on one end to the other end and slides the rest over. The goal is to reduce position bias.
- Repeat until all items are selected or the learner doesn’t select within the time limit. If they stop, end the session and mark the remaining items as not selected.
- Run several sessions. The original MSWO procedure used 5 sessions. A brief version using 3 sessions is also common. Record each session in its own column.
Common mistakes:
- Putting a selected item back in the array. That makes it a different assessment.
- Always placing a favorite item in the same position, which confuses preference with position bias.
- Using item sets that are too large for the learner to scan.
- Treating the top-ranked item as a permanent reinforcer. Re-assess regularly.
The sheet
Client:Date:Assessor:Setting:
Item type (edible / leisure / other):Selection time limit:Access time per selection:
Item key
| Code | Item |
|---|---|
| A | |
| B | |
| C | |
| D | |
| E | |
| F | |
| G | |
| H |
Selection order by session. Write the item code chosen on each trial.
| Trial | Session 1 | Session 2 | Session 3 | Session 4 | Session 5 |
|---|---|---|---|---|---|
| 1 | |||||
| 2 | |||||
| 3 | |||||
| 4 | |||||
| 5 | |||||
| 6 | |||||
| 7 | |||||
| 8 |
Summary. “Times available” is the number of trials in which the item was on the table.
| Code | Times selected | Times available | Selection % | Overall rank |
|---|---|---|---|---|
| A | ||||
| B | ||||
| C | ||||
| D | ||||
| E | ||||
| F | ||||
| G | ||||
| H |
Summarizing the data
Selection percentage = number of times the item was selected divided by the number of trials in which it was available, times 100. Rank items from highest to lowest selection percentage.
An item chosen first in a session was available for 1 trial in that session. An item chosen fourth was available for 4 trials. An item never chosen in a session that ended early was available on every trial run in that session.
Worked example: 5 items, 3 sessions.
| Item | Session 1 position | Session 2 position | Session 3 position | Selected / available | Selection % |
|---|---|---|---|---|---|
| A | 1 | 1 | 2 | 3 / 4 | 75% |
| B | 2 | 3 | 1 | 3 / 6 | 50% |
| C | 3 | 2 | 3 | 3 / 8 | 37.5% |
| D | 4 | 5 | 4 | 3 / 13 | 23% |
| E | 5 | 4 | 5 | 3 / 14 | 21% |
Item A is ranked highest, followed by B and C. Items D and E are close together and may not be meaningfully different.
Some clinicians rank by average selection position instead (A = 1.33, B = 2.0, C = 2.67, D = 4.33, E = 4.67). The two methods usually give similar orders. Pick one and use it consistently.
Tips
- Use the top two or three items as candidate reinforcers, and rotate them to reduce satiation.
- Re-run a brief MSWO (even a single session) at the start of teaching sessions if preferences seem to change from day to day.
- Watch for changes in motivating operations. An edible chosen first right before lunch may rank lower right after.
- Note the learner’s affect and approach behavior. A learner who chooses an item and then pushes it away is telling you something the ranking alone doesn’t.
- Respect assent. If the learner shows clear signs of not wanting to take part, such as pushing items away, leaving, or protesting, pause and try again later or use a less demanding method.