Comparison of Prompting Techniques to Teach Children With Autism to Ask Questions in the Context of a Conversation

Abstract

An adapted alternating treatments design was used to compare the efficacy of echoic and textual prompts to teach three students with autism (ages 8–15) to ask questions related to two pre-selected topics of conversation. Participants were first required to answer questions related to the topics to determine whether accurate responses were within their repertoire. This was followed by a transfer of stimulus control procedure to teach participants to ask relevant questions to the experimenter on the same topics. Probes with a novel conversation topic were conducted in the natural environment with a peer, and follow-up probes were conducted after training. Results indicate questions taught using an echoic prompt were acquired in fewer trials to criterion than questions taught using a textual prompt for two of the three participants. Limitations and implications for future research will be discussed.

Keywords

prompt comparison conversation skills question-asking autism

Programming focused on the development of language is often a major focus of treatment for children with a diagnosis of autism spectrum disorder (ASD). In the last decade, there has been an increased interest and focus on applications of B. F. Skinner’s (1957) analysis of verbal behavior to teach elementary verbal operants to children with ASD. In his seminal text, Skinner identified six elementary verbal operants, four of which are defined and discussed in further detail below.

A mand is a verbal operant in which a response is reinforced by access to specific reinforcement (i.e., saying “Milk, please” results in access to milk); and is under the control of relevant conditions of deprivation or aversive stimulation (Skinner, 1957). The second operant, a tact, is a response evoked by a particular object, event, or property of an object or event. Thus, a speaker names things that he or she has direct contact with (e.g., saying “Look, a plane” or “Milk”) when the item is present, and this results in access to nonspecific social reinforcement (i.e., "Yes, that is a plane!"). Third, an echoic is the repetition or mimicking of a speaker’s verbal behavior, which is also consequated with nonspecific social reinforcement. Fourth, an intraverbal is an operant under the control of a verbal stimulus with no point-to-point correspondence or formal similarity with the response, and is also maintained by nonspecific social reinforcement (Skinner, 1957). Common examples of elementary intraverbals include completing fill-in-the-blank statements (e.g., “Ready, set, ____!”), filling in words to songs or nursery rhymes, and answering questions.

Several procedures have been examined to teach these verbal operants to children with ASD. One such procedure is referred to as transfer of stimulus control, which consists of presenting a prompt (e.g., textual, picture, or echoic) and gradually fading these prompts until the learner responds independently when presented with a discriminative stimulus. For example, a number of studies have evaluated and compared prompting strategies to teach individuals with ASD to answer questions related to functions of objects (i.e., “What do you do with scissors?”), which is considered an intraverbal response (Finkel & Williams, 2001; Ingvarsson & Hollobaugh, 2011; Ingvarsson & Le, 2011; Vedora, Meunier, & Mackay, 2009). Findings from this line of research are contradictory, with some studies indicating visual (text or picture) prompts result in more efficient learning (Finkel & Williams, 2001; Ingvarsson & Hollobaugh, 2011; Vedora et al., 2009) while other studies indicate auditory (echoic) prompts are more efficient (Ingvarsson & Le, 2011). Results of a study by Coon and Miguel (2012) suggest the rate of acquisition for a target response may be a function of the history of reinforcement associated with a specific prompt type rather than the prompt itself.

A second area of research that has received recent attention is the role of motivation in teaching children to ask as opposed to respond to questions (Koegel, Bradshaw, Ashbaugh, & Koegel, 2014; Lechago, Carr, Grow, Love, & Almason, 2010; Shillingsburg & Valentino, 2011; Sundberg, Loeb, Hale, & Eigenheer, 2002). This skill may be considered a mand for information. A common feature of this line of research is the specific manipulation of motivation (e.g., contrived establishing and abolishing operations) to teach participants to ask questions that puts them in direct contact with a desired tangible reinforcer (Lechago et al., 2010; Shillingsburg & Valentino, 2011; Sundberg et al., 2002). At least one study also incorporated the use of different prompts to evoke the desired response (echoic or textual; Shillingsburg, Bowen, & Valentino, 2014).

There is also some existing research on procedures to teach children and adolescents how to ask questions maintained by social reinforcement (i.e., engaging in a conversation with another individual). Procedures that have been evaluated to date include cue card/written script (Charlop-Christy & Kelso, 2003), small group instruction (Palmen, Didden, & Arts, 2008), self-management (Doggett, Krasno, Koegel, & Koegel, 2013), and pivotal response treatment (Koegel et al., 2014). No studies to date have directly compared common prompting techniques that have demonstrated success in teaching other verbal operants to establish what may be conceptualized as mands for information maintained by social reinforcement.

Teaching individuals with ASD to ask questions that are maintained by social reinforcement may be considered a starting point to successful social interactions with peers (e.g., basic greetings). Teaching to ask questions that are not directly maintained by access to a tangible item may also help promote independence (e.g., “Is it cold outside today?” can help provide general information about the type of clothing one should wear) or safety skills (e.g., “What street am I on?”). In a recent literature review on this topic, Raulston and colleagues (2013) summarized the results of 21 studies that evaluated teaching strategies to prompt question-asking behavior in individuals with ASD. Results indicated the most common procedures utilized to teach this target behavior include echoic prompting and fading procedures, positive reinforcement (e.g., providing access to information or requested item), and manipulating establishing operations to motivate question-asking behavior. Despite the positive outcomes reported to date, no studies have systematically compared prompting procedures to evaluate their relative effectiveness. Therefore, the purpose of the present study was to extend previous research by systematically comparing two procedures (echoic and textual prompts) that have demonstrated efficacy in teaching other verbal responses to this population.

Method

Participants, Setting, and Materials

Three students were recruited from a therapeutic center in Northeast Ohio for children with developmental disabilities. Steve was a 12-year-old Caucasian male with a diagnosis of ASD, Wayne was a 15-year-old Caucasian male with a co-morbid diagnosis of ASD and obsessive compulsive disorder, and Ivan was an 8-year-old African American male with a co-morbid diagnosis of ASD and attention deficit hyperactivity disorder (ADHD). Diagnoses were confirmed from each student’s individualized education plan (IEP). Steve and Wayne did not receive any prescribed psychotropic medications throughout the course of the study. Ivan was prescribed Focalen™ once daily as an extended release capsule at the start of the study, but stopped taking this medication after the third baseline session was conducted and did not take other medications for the remainder of the study.

Each participant was assessed using the Verbal Behavior Milestones Assessment and Placement Protocol (VB-MAPP; Sundberg, 2008). All participants scored in Level 2 for the mand, tact, echoic, and intraverbal subsections. They demonstrated a basic reading repertoire of five to eight word sentences and were fluent in answering WH- questions (e.g., what, who, when, where, and why), but presented deficits in spontaneous question-asking. In addition, all participants had a goal on their IEP related to the use of appropriate social skills, including asking questions during a conversation.

Sessions were conducted in a quiet area of the building in a classroom adjacent to the student’s regular classroom and lasted approximately 5 min. The area was equipped with a table and several chairs. Each participant’s classroom was comprised of 8 to 15 peers diagnosed with ASD, one teacher, and at least one teacher’s aide. One session, which included both prompt types, was conducted per day for each participant. Baseline, training, and follow-up sessions were conducted individually with each participant, the experimenter. A second observer was also present during sessions in which interobserver agreement (IOA) and procedural integrity (PI) data were collected. Generalization sessions were conducted with one participant at a time, but a peer was always present, as described below.

Three sets of four questions each were created based on three different topics of conversation. All questions in each set contained five to eight words (see Table 1) to help control for the level of difficulty across sets of questions. Questions and topics were selected based on their applicability to the natural environment and typical conversation with peers. Some questions allowed for variability in answers (e.g., “What did you eat for breakfast?”), while others did not (e.g., “What school do you go to?”). Textual prompts were presented on the plain side of an index card (7.62 cm × 12.7 cm) with text typed in Times New Roman 36-pt. black colored font.

Table 1.

List of Stimulus Set Questions.

Stimulus set	Question
1. Food	1. What are your favorite foods?
	2. What did you eat for breakfast?
	3. Which restaurants do you like to eat at?
	4. What do you like to cook?
	5. What are you having for lunch/dinner today?
2. Games	1. What is your favorite board game/video game?
	2. Who do you play games with?
	3. Where do you play games?
	4. What sports do you play?
	5. What is your favorite sports team?
3. School	1. What school do you go to?
	2. What is your teacher’s name?
	3. What is your favorite subject?
	4. What do you want to be when you grow up?
	5. What do you like about school?

Experimental Design, Dependent Variable, and IOA

An adapted alternating treatments design (Sindelar, Rosenberg, & Wilson, 1985) was implemented to compare the efficacy of echoic and textual prompts on the acquisition of questions asked by participants in the context of a conversation. This design controls for major threats to internal validity (e.g., history, maturation, instrumentation) and demonstrates experimental control when differentiation between conditions is evident (e.g., data paths of the conditions do not overlap). We presented the echoic and textual prompt conditions in a random order to help eliminate the possibility of temporal cycles of behavior.

Baseline sessions were implemented followed by training with each prompt type for a specific stimulus set (i.e., set of questions). One session with each stimulus set was conducted per day, with the order of presentation alternated daily. Stimulus sets were randomly assigned to a prompt type via a coin flip and these were counterbalanced across participants. For example, Stimulus Set 1 (food) was assigned to the echoic prompt condition and Stimulus Set 2 (games) was assigned to the textual prompt condition for Steve, and these assignments were switched for Wayne and Ivan. Stimulus Set 3 (school) was not directly trained, but utilized in the novel stimulus set probe phase (see details below). In addition, the questions presented by the experimenter and those that were prompted were randomized within each training session by shuffling index cards that listed the questions from each stimulus set. This was done in an attempt to produce a novel conversation and eliminate memorization of a script.

The dependent variable was the number of questions asked independent of prompts during each session. Three responses were scored: (a) correct response, defined as a vocalization that was grammatically correct, consisted of more than one word, and corresponded to the selected topic of conversation; (b) incorrect response, defined as a vocalization that consisted of just one word, a repetition of a previously answered question within that session, a response that was not grammatically correct, or a nonsensical response (unrelated to the selected topic of conversation, or echolalia); and (c) no response, defined as no emitted vocalization within the allotted time (0–8 s).

Interobserver agreement data were collected by a second trained observer and calculated by dividing the number of agreements by the number of agreements plus disagreements and multiplying this number by 100. An agreement was scored when both the primary and secondary observer recorded the same participant response during each trial (e.g., correct, incorrect, or no response). Interobserver data were collected during 32% of all sessions for Steve (M = 97.6%, range = 50%–100%), 60% of all sessions for Wayne (M = 97.83%, range = 50%–100%), and 42% of all sessions for Ivan (M = 96.21%, range = 50%–100%). The variability in IOA was partially attributed to disagreements in whether a participant emitted a complete sentence during the initial IOA sessions. We conducted additional training and review of the operational definitions for the secondary observer and noted an increase in IOA for the remaining sessions throughout the study.

Procedural integrity data were also collected by a second trained observer and calculated by summing the number of correctly implemented responses performed by the experimenter on checklists created for the purpose of this study (available from the second author upon request), by the total number of possible responses, and multiplying that number by 100. Procedural integrity data were collected during 22.5% of all sessions for Steve, 45% of all sessions for Wayne, and 42.8% of all sessions for Ivan, and scored at 100% across all sessions.

Procedure

Pre-Training

A list of questions related to each of the three selected topics of conversation (food, games, and school) were probed with each participant to determine if appropriate accurate responses were within the participants’ repertoire prior to beginning of data collection. For example, if the experimenter presented the question, “What’s your favorite game?” an appropriate and accurate response was “My favorite game is SORRY” (or any other game). If an accurate response was not emitted, a new question was substituted until a question with an appropriate response was recorded. This continued until each stimulus set contained a total of five questions answered correctly by all participants. This phase was important because participants were required to respond to all the questions they were taught to ask during training to produce a back-and-forth exchange with the experimenter. In addition, during pre-training, we ensured that all participants reliably followed echoic prompts and read textual prompts without difficulty.

Prompt Exposure

Following the pre-training phase, 10 responses to questions that were unrelated to the topics of conversation, but based on general knowledge that the participants did not have within their repertoire were taught using echoic and textual prompts. Five questions were assigned to the echoic prompt condition, and five questions were assigned to the textual prompt condition. Questions were presented in a quasi-random fashion in 10 trial blocks until participants demonstrated a mastery criterion of 80% (8 of 10) correct independent responses across 2 trial blocks. Correct responses were defined as complete and accurate answers to the question and were reinforced with descriptive praise (e.g., “You’re right! Columbus is the capital of Ohio!”). Incorrect responses were defined as incomplete, inaccurate, and unrelated or no response to a question and were followed by corrective feedback and either an echoic prompt (e.g., “Say _____”) or textual prompt (e.g., “Say” + presentation of the textual stimulus on an index card as described above). This phase was conducted to provide a similar history of reinforcement for each prompt type across all participants prior to the start of the study.

Baseline

During baseline, the experimenter sat at a table across from the participant and presented the instruction, “I am going to ask you some questions about (topic) and I want you to answer and then ask me four questions about (topic) today.” The experimenter then asked the participant one of five pre-selected questions based on two topics of conversation (e.g., games and food). Following each participant response, the experimenter presented additional small talk (e.g., elaborations, statements, or acknowledgment) to simulate, as close as possible, natural conversation (see Table 2 for an example). A 5-s delay was incorporated following the small talk to provide an opportunity for the participant to ask the experimenter a related question and continue the conversation. If a correct response was emitted, the experimenter answered the question and incorporated small talk, but did not provide any additional feedback. This was followed by another 5-s delay, and the same sequence was repeated until the participant produced up to four correct responses. If an incorrect response was recorded, the experimenter stated, “We are talking about (food/games) right now,” ended the trial, and presented a new question on the same topic. If no response was recorded after 5 s, the trial was terminated, and a new question was presented by the experimenter on the same topic. This incorrect/no response sequence continued until four questions from the stimulus set had been presented to the participant. In sum, four trials or opportunities to ask questions were presented per session regardless of response accuracy.

Table 2.

Example Conversations During Baseline.

Trial	Conversation
No response	Experimenter: “What is your favorite game to play?” (Question)
	Participant: “My favorite game is Super Mario.” (Answer)
	Experimenter: “Cool, I play Mario.” (Small talk)
	Time delay (up to 5 s)
	Trial ends
Incorrect response	Experimenter: “What is your favorite game to play?”
	Participant: “My favorite game is Super Mario.”
	Experimenter: “Cool, I play Mario.”
	Time delay (up to 5 s)
	Participant: “Do you like pigs?”
	Experimenter: “We are talking about games right now.”
	Trial ends
Correct response	Experimenter: “What is your favorite game?” (Question)
	Participant: “My favorite game is Super Mario.” (Answer)
	Experimenter: “Cool, I play Mario.” (Small talk)
	Time delay (5 s)
	Participant: “Do you play baseball?”
	Experimenter: “I play baseball sometimes.”
	Experimenter: “I prefer basketball.” (Small talk)
	Trial ends

Note. Conversation starter/directions: Experimenter: “I am going to ask you some questions, and I want you to ask me four questions about (topic) today.”

Prompt Comparison

During the prompt comparison phase, the same instructions and format provided during baseline were implemented, with the exception that specific verbal praise was provided following all correct independent responses (i.e., “That’s a good question!” See Table 3 for an example). A stimulus set in either the echoic or textual prompt condition was considered mastered when the participant emitted four correct responses across two consecutive sessions.

Table 3.

Example Conversations During Training.

Trial	Conversation
No response	Experimenter: “What is your favorite game to play?” (Question)
	Participant: “My favorite game is Super Mario.” (Answer)
	Experimenter: “Cool, I play Mario.” (Small talk)
	Time delay (up to 5 s)
	Participant: No response scored
	Experimenter: “Say . . . where do you play baseball?” or “Say” + presentation of textual prompt
	Participant: “Where do you play basketball?”
	Experimenter: “I play basketball at school.”
	Experimenter: “Lots of people play there.”
	Trial ends
Incorrect response	Experimenter: “What is your favorite game to play?”
	Participant: “My favorite game is Super Mario.”
	Experimenter: “Cool, I play Mario.”
	Time delay (up to 5 s)
	Participant: “Do you like pigs?” (Incorrect response scored)
	Experimenter: “We are talking about games right now. Say . . . where do you play baseball?” Or “Say” + presentation of textual prompt
	Participant: “Where do you play basketball?”
	Experimenter: “I play basketball at school.”
	Experimenter: “Lots of people play there.”
	Trial ends
Correct response	Experimenter: “What is your favorite game?” (Question)
	Participant: “My favorite game is Super Mario.” (Answer)
	Experimenter: “Cool, I play Mario.” (Small talk)
	Time delay (5 s)
	Participant: “Do you play baseball?” (Correct response scored)
	Experimenter: “What a great question! I play baseball sometimes.”
	Experimenter: “I prefer basketball.” (Small talk)
	Trial ends

Note. Conversation starter/directions: Experimenter: “I am going to ask you some questions, and I want you to ask me four questions about (topic) today.”

Echoic prompts

For each participant, one stimulus set was taught using echoic prompts. Echoic prompts were presented vocally with the instruction “Say, (______)” prior to the prompt delivery. The experimenter began each conversation in the same manner as described in baseline. That is, the experimenter presented a question, the participant answered the question, and then small talk was presented by the experimenter. A delay (ranging from 0 to 8 s) was implemented during this phase to provide the participant with an opportunity to ask the experimenter a question on the topic. As in baseline, each session consisted of four trials, or opportunities, to ask questions.

Prompts were delivered using a progressive time delay procedure ranging from 0 to 8 s. Initially, prompts were provided immediately following the experimenter’s small talk or answer to a question (i.e., 0-s delay). The delay between small talk and the opportunity to respond was systematically increased. The initial trial block consisted of a 0-s prompt delay, and this was immediately increased to 2 s in the second trial block. Following the second trial block, the prompt delay increased by 2 s after every four sessions.

Textual prompts

Textual prompts were also preceded by the vocal instruction “Say” and immediately followed by presentation of the index card. Prompts were delivered in the same fashion described for echoic prompts, with the exception that written responses were presented instead of a vocal stimulus. Textual prompts were presented at eye level and approximately arm’s length away from the participant. Incorrect responses resulted in re-presentation of the instruction “Say” and a gestural point to the textual prompt. This process was repeated, when needed, until the participant responded by asking the experimenter the question presented on the card.

Prompt Fading

If mastery criterion was not met following implementation of the 8-s prompt delay in eight sessions, a prompt fading procedure was incorporated into the training. This consisted of a continuous 8-s prompt delay in addition to systematically removing one word from both textual and verbal prompts after every two sessions. Corrective and approving feedback was provided in the same manner described above for the prompt-delay phase.

Novel Stimulus Set Probe

After mastery criterion was demonstrated with both the textual and echoic training conditions, one 4-trial session was conducted with each participant, in which a novel conversation topic (i.e., school) was assessed for response generalization. These probes were conducted in the same manner as baseline, but with a novel set of questions (see Table 1).

Natural Environment Probes

Following novel stimulus set probes, a probe in the context of a social skills lesson was conducted with a peer from each student’s classroom who was not part of the study. This probe consisted of presentation of two 4-trial sessions, one with both of the previously trained stimulus sets. Participants were provided with the following instruction: “We are going to have a conversation lesson today. I want each of you to ask the other four questions about (topic).”

During these probes, the experimenter mediated conversation by delivering verbal and textual prompts to the untrained peer if no responding occurred after 8 s. To avoid prompting the participant, a textual prompt from the training phase was provided to the untrained peer only after 8 s with no response. Feedback was not delivered to the participant at any time during these probes, and correct, incorrect, or no responding was recorded as defined above.

Follow Up

Following all novel stimulus set and natural environment probes, follow-up probes were conducted for 1 to 3 weeks depending on participant availability. These probes were conducted in a manner identical to that described during baseline.

Results

Results for individual and total sessions to criterion during each prompt condition are presented in Figure 1. Overall, two of the three participants (Steve and Ivan) met mastery criterion in fewer sessions when echoic prompts were used; and the textual prompt condition resulted in fewer sessions to mastery criterion for one participant (Wayne).

Figure 1.

Individual and total trials to criterion per prompt type for all participants.

Figure 2 shows the results for correct independent responses during baseline, training, novel stimulus set probes, natural environment probes, and follow-up sessions for Wayne. During baseline, correct responses varied between zero and two in both the textual and echoic prompt conditions. Following implementation of the prompt-delay procedure, correct independent responding increased, and Wayne met mastery criterion in both conditions. Wayne required 8 sessions to meet mastery criterion in the textual prompt condition and 10 sessions to meet mastery criterion in the echoic prompt condition. He provided three correct responses during the novel stimulus set probe and four correct responses in both the textual and echoic stimulus sets during the natural environment probes. Wayne emitted four correct responses for both stimulus sets during the 2- and 3-week follow-up sessions.

Figure 2.

Number of correct independent responses per session for Wayne.

Results for Ivan (see Figure 3) displayed a variable baseline for both textual and echoic stimulus sets ranging from zero to two correct responses. Unlike Wayne, Ivan did not meet mastery criterion during the prompt-delay procedure in either the textual or echoic prompt condition. The prompt fading procedure was implemented for both stimulus sets after 21 training sessions for this participant only. Ivan met mastery criterion following 6 and 10 training sessions for the echoic and textual stimulus sets, respectively. He provided four correct responses during the novel stimulus set probe, and three and two correct responses in textual and echoic stimulus set during the natural environment probes, respectively. Ivan emitted one and zero correct responses (echoic and textual prompt conditions, respectively) during the initial follow-up probe. These responses increased at the 2-week follow-up (two and zero for textual and echoic, respectively) and varied again at the 3-week follow-up probe (zero and two for echoic and textual, respectively).

Figure 3.

Number of correct independent responses per session for Ivan.

Steve’s results (see Figure 4) yielded consistent baseline responding at near zero levels. During the prompt-delay phase, he required 20 sessions to meet mastery criterion in the textual prompt condition and 15 sessions to meet mastery criterion in the echoic prompt condition. He provided three correct responses during the novel stimulus set probe and zero correct responses in both stimulus sets during the first natural environment probe. A second natural environment probe with a peer resulted in two and zero correct responses for the textual and echoic stimulus sets, respectively. Steve emitted three and four correct responses (echoic and textual prompt conditions, respectively) at 1-week follow-up.

Figure 4.

Number of correct independent responses per session for Steve.

Discussion

Overall, two of the three participants learned to ask questions in fewer sessions to mastery criterion when echoic prompts were provided. These results provide further support for the efficacy of echoic prompts to teach vocal responses, including answering questions, to students with ASD (Ingvarsson & Le, 2011), and extend results of prior studies by demonstrating the efficacy of this procedure to teach question-asking behavior within the context of a conversation. Previous studies have evaluated teaching strategies to establish question-asking behavior (e.g., mands for information) that result in access to tangible items (Lechago et al., 2010; Shillingsburg & Valentino, 2011; Sundberg et al., 2002; Williams, Perez-Gonzalez, & Vogt, 2003). Results from this study provide support for the use of a transfer of stimulus control procedure to teach students with ASD to initiate appropriate interactions with peers (Charlop-Christy & Kelso, 2003; Doggett et al., 2013; Koegel et al., 2014; Palmen et al., 2008); which can be considered a mand for information maintained by social reinforcement.

These results are in contrast with findings that indicate textual prompts may be more efficient to teach intraverbal behavior (Finkel & Williams, 2001; Vedora et al., 2009). Previous research has addressed the possibility that the efficacy of a prompt type may be attributed in part to participants’ exposure to a specific prompt rather than the prompt itself (Coon & Miguel, 2012). In the present study, we incorporated a prompt exposure phase prior to training so that all participants had at least some experience with each prompting procedure. However, it is possible that the efficacy of a prompt type is idiosyncratic, in which case, it would be helpful to assess and determine the type of prompting procedure that is best suited for a learner prior to the start of training. Assessments to determine ideal teaching strategies for children with autism are sorely needed to capitalize on the time that is available for therapy. Some prior research has been conducted in this area to evaluate functions of emerging speech (Lerman et al., 2005) and readiness skills prior to implementing conditional discrimination training (Kodak, Clements, & LeBlanc, 2013). Because an individual’s history with a prompt type may affect performance (Coon & Miguel, 2012), incorporating a brief assessment related to prompt effectiveness or prompt preference may provide educators with information that can enhance the use of instructional time.

Results of the present study indicate echoic prompts may be more efficient at producing correct independent responses. However, both prompt types were effective in producing the target responses across all participants. In addition, the difference may not be considered socially significant. That is, Steve required 20 (textual) and 15 (echoic) trials to meet the mastery criterion, Wayne required 8 (textual) and 10 (echoic) trials to meet mastery criterion, and Ivan required 31 (textual) and 27 (echoic) trials to meet the mastery criterion. These results may be attributed in part to the prompt exposure phase conducted prior to baseline.

There are a few reasons why the echoic prompts may have been slightly more effective than textual prompts for two of the three participants. First, it is possible that the participants’ responses more closely resemble that of the echoic prompt, and thus, transfer of stimulus control was more efficient than when textual prompts were used. In future studies, this may be controlled by having participants respond in the same format as the prompt (i.e., writing a response to a textual prompt). It should also be noted that even though participants received equal exposure to each prompt type prior to the start of the study, the distribution of prompt types by teacher, parents, and peers may have differed. These individual differences may have created a stronger history of reinforcement for echoic prompts.

Limitations and Future Directions

The results of this study are preliminary and replications are needed to further evaluate the relative effectiveness of echoic and textual prompts to teach students with ASD to ask questions in the context of a conversation. Some limitations of the present study should be noted for future research. First, all baseline sessions were conducted in a manner whereby participants were provided with 5 s to respond before the termination of a trial or additional instructions by the experimenter were provided. It should be noted that only one stimulus set met the pre-determined mastery criterion prior to 6-s prompt delay. Thus, most correct responses were emitted only after at least a 5-s period of time had elapsed. Future studies should consider increasing the time allotted for participants to provide a response during baseline sessions.

Second, Steve began the 0-s prompt delay with one response in each stimulus set. This may have been a result of the instructions presented at the beginning of each session: “I’m going to ask you some questions, and I want you to ask me four questions about (topic) today.” That is, this set of instructions may have functioned as a discriminative stimulus for the participants to emit appropriate responses that were not a result of the intervention. Future research should incorporate the use of motivating operations (MOs) as has been done with previous studies on teaching mands for information maintained by access to tangible reinforcement (Lechago et al., 2010; Shillingsburg et al., 2014; Sundberg et al., 2002) to eliminate this potential threat to internal validity and demonstrate a more pure form of mands for information within this context.

Third, we did not formally assess whether participants asked questions in context other than in conversation prior to the start of the study. That is, it may be important to identify students who already ask questions that will result in access to tangible reinforcers (i.e., asking where something is or who has something of interest), because this may be a pre-requisite skill for students to ask questions within the context of a conversation. Although this skill was not formally assessed, anecdotal evidence suggests all three participants had this skill within their repertoire. It may also be prudent for future researchers to conduct preference assessments to determine topics of conversation that are of interest to each participant. The opportunity to obtain novel information regarding a topic of interest may function as an MO to ask questions. Social reinforcement should also be assessed across multiple individuals to determine preference for one particular individual’s social reinforcement.

Fourth, future studies should consider increasing the number of questions in each stimulus set or the number of stimulus sets incorporated during training. The addition of a control set of questions (i.e., a fourth stimulus set that did not receive the intervention) may have further strengthened the internal validity if responses to these questions did not increase throughout the study. In addition, although we incorporated small talk and randomized prompt order to control for the possibility that participants were memorizing a script, additional opportunities to emit questions within the context of a conversation (i.e., multiple exemplars) may help to promote generalization. The use of multiple experimenters during training can also help improve generalization to peers.

Last, results from probes for generalization (with a peer and with a novel stimulus set) were variable. Future research should focus on ways to improve and program for generalization. In addition to establishing effective MOs regarding conversational topics and increasing the number of questions and stimulus sets, incorporating peers into the training phase may help promote generalization. Future studies can utilize peers with experimenter guidance throughout the entire training, as was done in the novel environment probe in the present study. Training in a more naturalistic environment may also help promote the use of the skills learned during training in everyday settings.

Footnotes

Acknowledgements

We thank Zachary Vargo and Lauren Stout for help with reliability data.

Authors’ Note

This project constituted the master’s thesis completed by the first author under the supervision of the second author to fulfill graduation requirements for the Applied Behavior Analysis master’s program at Youngstown State University.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Charlop-Christy

M. H.

Kelso

S. E.

(2003). Teaching children with autism conversational speech using a cue card/written script program. Education and Treatment of Children, 26, 108–127.

Coon

J. T.

Miguel

C. F.

(2012). The role of increased exposure to transfer-of-stimulus-control procedures on the acquisition of intraverbal behavior. Journal of Applied Behavior Analysis, 45, 657–666. doi:10.1901/jaba.2012.45-657

Doggett

R. A.

Krasno

A. M.

Koegel

L. K.

Koegel

R. L.

(2013). Acquisition of multiple questions in the context of social conversation in children with autism. Journal of Autism and Developmental Disorders, 43, 2015–2025. doi:10.1007/s10803-012-1749-8

Finkel

A. S.

Williams

R. L.

(2001). A comparison of textual and echoic prompts on the acquisition of intraverbal behavior in a six-year-old boy with autism. The Analysis of Verbal Behavior, 18, 61–70.

Ingvarsson

E. T.

Hollobaugh

(2011). A comparison of prompting tactics to establish intraverbals in children with autism. Journal of Applied Behavior Analysis, 44, 659–664. doi:10.1901/jaba/2011/44-659

Ingvarsson

E. T.

D. D.

(2011). Further evaluation of prompting tactics for establishing intraverbal responding in children with autism. The Analysis of Verbal Behavior, 27, 75–93.

Kodak

Clements

LeBlanc

(2013). A rapid assessment of instructional strategies to teach auditory-visual conditional discriminations to children with autism. Research in Autism Spectrum Disorders, 7, 801–807. doi:10.1016/j.rasd.2013.02.007

Koegel

R. L.

Bradshaw

J. L.

Ashbaugh

Koegel

L. K.

(2014). Improving question-asking initiations in young children with autism using pivotal response treatment. Journal of Autism and Developmental Disorders, 44, 816-827. doi:10.1007/s10803-013-1932-6

Lechago

S. A.

Carr

J. E.

Grow

L. L.

Love

J. R.

Almason

S. M.

(2010). Mands for information generalize across establishing operations. Journal of Applied Behavior Analysis, 43, 381–395. doi:10.1901/jaba.2010.43-381

10.

Lerman

D. C.

Parten

Addison

L. R.

Vorndran

C. M.

Volkert

V. M.

Kodak

(2005). A methodology for assessing the functions of emerging speech in children with developmental disabilities. Journal of Applied Behavior Analysis, 38, 303–316. doi:1901/jaba.2005.106-04

11.

Palmen

Didden

Arts

(2008). Improving question asking in high-functioning adolescents with autism spectrum disorders: Effectiveness of small-group training. Autism, 12, 83–98. doi:10.1177/1362361307085265

12.

Raulston

Carnett

Lang

Tostanoski

Lee

Machalicek

. . . Lancioni

G. E.

(2013). Teaching individuals with autism spectrum disorder to ask questions: A systematic review. Research in Autism Spectrum Disorders, 7, 866–868. doi:10.1007/s10803-0100938-6

13.

Shillingsburg

M. A.

Bowen

C. N.

Valentino

A. L.

(2014). Mands for information using “how” under EO-absent and EO-present conditions. The Analysis of Verbal Behavior, 30, 54–61. doi: 10.1007/s406116-013-002-7

14.

Shillingsburg

M. A.

Valentino

A. L.

(2011). Teaching a child with autism to mand for information using “how.” The Analysis of Verbal Behavior, 27, 179–184.

15.

Sindelar

P. T.

Rosenberg

M. S.

Wilson

R. J.

(1985). An adapted alternating treatments design for instructional research. Education and Treatment of Children, 8, 67–76.

16.

Skinner

B. F.

(1957). Verbal behavior. Acton, MA: Copley.

17.

Sundberg

M. L.

(2008). VB-MAPP: Verbal behavior milestones assessment and placement program. Concord, CA: AVB Press.

18.

Sundberg

M. L.

Loeb

Hale

Eigenheer

(2002). Contriving establishing operations to teach mands for information. The Analysis of Verbal Behavior, 18, 15–29.

19.

Sundberg

M. L.

Michael

(2001). The benefits of Skinner’s analysis of verbal behavior for children with autism. Behavior Modification, 25, 698–724.

20.

Vedora

Meunier

Mackay

(2009). Teaching intraverbal behavior to children with autism: A comparison of textual and echoic prompts. The Analysis of Verbal Behavior, 25, 79–86. doi:10.1901/jaba.2011.44-659

21.

Williams

Perez-Gonzalez

L. A.

Vogt

(2003). The role of specific consequences in the maintenance of three types of questions. Journal of Applied Behavior Analysis, 36, 285–296. doi:10.1901/jaba.2003.36-285