Many responses are shaped up by the natural consequences that follow behavior however; when there is a skill deficit it is often necessary to prompt the critical responses and explicitly strengthen them.  Prompts are usually referred to as stimuli that are effective in promoting the response in the appropriate context.  For example, a child that is being taught to request a desired toy through handing a picture of the toy to a caregiver can be taught with physical assistance, through vocal prompts by the caregiver, the caregiver can model the response (and provide the consequence), they can be shown a video of a picture exchange, or via a vast array of similar strategies.  Prompting in this context is typically referred to as response prompting and the primary purpose of response prompting is to guide the child's response during instruction.

There are several types of response prompts that have been used for teaching socially significant behavior. Though I'm discussing response prompting here in the context of focused teaching, prompts are everywhere and occur naturally without explicit plans on the part of the caregiver.  One of the most commonly encountered prompts in any child's life is vocal prompting.  Vocal prompts are very important and many children with ASDs do not effectively learn with them.  It certainly should be a goal of any instructional programming to promote behavior that is in contact with vocal prompting but many children will require intense instruction before that is feasible.  On the other hand, if vocal prompting is effective then it should be capitalized on.  Several studies have shown that children with ASDs will sometimes learn important social skills with vocal prompts.  For example, Odom and Strain (1986) worked with children with ASDs who infrequently initiated social interactions with peers.  They used vocal prompts such as, "I'd like you to play with Jimmy today" and found that this simple prompting was effective for increasing social interactions.

Textual prompts, or written words/instructions, are another form of prompting that can sometimes be effective.  Krantz and McClanahan (1998) used written textual scripts to promote initiating a conversation for children with ASDs who rarely initiated.  Not only did initiations with caregivers increase but there was also a significant level of generalization to novel conversational topics that was noted to have occurred.  Pictorial prompts (or pictures) have also been found to systematically produce behavior change for person with ASDs.  The work of Andy Bondy and Lori Frost is generally well known by educators who work with children with limited or no vocal communication skills.  However, the most common form of prompting with persons with limited communicative skills is probably physical prompting.  Again, an important goal is to move to more naturalistically occuring prompts but I'd like to talk a little bit more in depth about physical prompting.  I'm also glossing over some other types of prompts like gestural prompts and modeling but I plan to do an extensive post on video modeling in the near future.

Physical prompting is sometimes unpleasant to the person being prompted.  In these cases, it is most desirable to use another form of prompting during instruction.  Even some responses taught to typical children through prompting, like tooth brushing, can be taught using other prompts (modeling and video prompts are often very effective).  On the other hand, many people enjoy physical prompts, or the reinforcement that follows being prompted to respond.  This sometimes leads to dependence on prompting (though this is by no means limited to physical prompts). 

Many complex and important skills are efficiently taught through physical prompts.  At the New England Center for Children, we have had a long history of preferring to implement physical (and other) prompts in a manner that avoids the learner making an error.  This is sometimes referred to as errorless teaching (or most-to-least prompting).  However, we noticed that some children learn more efficiently when providing them with minimal physical prompts in more of a trial-and-error fashion (sometimes referred to as least-to-most prompting).  A study we published last year (the lead author was Myrna Libby, a valued colleague who passed away unexpectedly last year) involved a comparison of errorless and trial-and-error teaching procedures (Libby et al., 2009).  We taught several children  a number of solitary play skills (putting together duplo lego structures) and each child was exposed to each teaching procedure.  We found that several of the students learned more rapidly with the trial-and-error approach despite the fact that they made more errors with this procedure while learning.  They also learned the skills taught with the errorless procedure.  However, there were also several children that did not learn at all with the trial-and-error approach.  They only learned with the errorless approach. 

One useful lesson we took from this study was that we would benefit from developing rapid assessment procedures for identifying the teaching procedures that each child learned with most rapidly.  Though our preferred strategy of errorless instruction was always effective, we may actually slow learning down if we stuck to this procedure for all children.  In full disclosure, we have many different teaching procedures that are used across our school and our goal has always been to tailor instruction to each student's needs.  Also, we feel that there is a great need to further explore other subtle nuances of instruction to advance our understanding of how to teach.