The Effect of the High‐Probability Instructional Sequence Across Foods With Varying Levels of Participant Interaction
High-p warm-ups can spark eating in some autistic kids yet flop for others, so probe first and stay ready to add reinforcers.
01Research in Context
What this study did
Sheppard et al. (2026) tested the high-probability (high-p) instructional sequence on three autistic kids.
They used foods that needed different levels of child work: touching, smelling, or chewing.
Each session started with three easy requests the child always obeyed, then one hard food request.
What they found
One child ate every food type after the high-p warm-up.
The other two children still refused most bites, even after the easy tasks.
The team called the outcome “mixed” because it helped some but not all.
How this fits with other research
Older feeding studies looked rosier. Penrod et al. (2012) and Meier et al. (2012) both got autistic kids to swallow new foods with the same high-p trick and reported clear wins.
Rosales et al. (2021) gives a clue why results clash. They ran high-p for compliance, not eating, and also saw only two of three kids improve. Their takeaway: have differential reinforcement ready when high-p alone stalls.
Put together, the picture is “works for some, not for others.” The 2026 paper simply confirms that rule in the lunch room.
Why it matters
Do not assume the high-p sequence will fix food refusal. Run a quick probe with each child and watch the data daily. If acceptance stays flat after a few sessions, layer in extra reinforcers or try sequential presentation as shown by Davis et al. (2023). Always let the data, not the protocol, guide your next bite.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Start lunch with three easy requests your client always follows, then present one tiny bite of target food—and graph hit or miss to see if high-p is enough.
02At a glance
03Original abstract
ABSTRACT We evaluated the utility of the high‐probability (high‐ p ) instructional sequence to increase bite acceptance and consumption of foods to which three children with autism spectrum disorder exhibited varying levels of interaction in a pre‐assessment. We first conducted food preference assessments to identify each participants' most preferred foods. Next, to identify high‐interaction, medium‐interaction, and low‐interaction foods, we conducted a food interaction assessment, which consisted of assessing participants' cooperation with instructions to touch, smell, and consume foods. We then incorporated the identified high‐preference foods into high‐ p instructions (i.e., “Take a bite of high‐p food ”) and evaluated their utility to increase cooperation with three low‐probability (low‐ p ) instructions (i.e., “Take a bite of low‐p food” ) for each participant. The low‐ p instructions varied according to the level of participant interaction with the food as identified in the food interaction assessment (i.e., foods presented in the low‐ p instruction were either high‐interaction, medium‐interaction, or low‐interaction foods). Results show that the high‐ p instructional sequence was effective to increase bite acceptance and consumption across all three levels of low‐probability instructions for one participant, but ineffective to increase acceptance and/or consumption for the other two participants. We discuss the utility of the high‐ p instructional sequence to increase food consumption.
Behavioral Interventions, 2026 · doi:10.1002/bin.70085