The Effect of a Mediation-Blocking Task on the Acquisition of Instructive Feedback Targets

Abstract

The inclusion of instructive feedback in discrete-trial training has been shown to increase the efficiency of learning. However, the behavioral mechanism underlying the effectiveness of this procedure has not yet been determined. Researchers have suggested that learners covertly self-echo the feedback, which mediates later responding. The present study sought to understand the role of self-echoics in the acquisition of untaught targets. Participants were directly taught to tact pictures, then given instructive feedback after the praise statement. The 3 experimental conditions were (a) a typical instructive feedback procedure; (b) a vocal mediation-blocking procedure, in which the participants were asked to engage in a competing vocal response immediately after the instructive feedback; and (c) a motor-distraction procedure, in which the participants were asked to engage in a motor response immediately after the instructive feedback. The inclusion of the vocal mediation-blocking task had little effect on the participants’ ability to learn the instructive feedback targets for all 3 participants.

Instructive feedback (IF) is a procedure that involves presenting additional, nontarget stimuli (i.e., a response is not reinforced or prompted) during the antecedent or consequent condition during discrete-trial training (DTT; Wolery, Werts, Holcombe, Billings, & Vassilaros, 1993). For example, Doyle Gast, Wolery, and Ault (1990) taught four secondary-age students with developmental disabilities to identify local and federal government buildings and over-the-counter medications by including incidental information (e.g., Bayer is an aspirin) in the praise statements for correct, directly taught targets.

When compared to DTT alone, DTT with IF has been shown to be more efficient because it allows more targets to be learned in the same amount of teaching time (Nottingham, Vladescu, Kodak, & Kisamore, 2017). Nottingham et al. (2017) evaluated the efficiency of DTT with and without IF using four conditions with between zero and four IF stimuli in the antecedent or consequent conditions. Both participants required less time in session per target in the conditions that included multiple IF stimuli, regardless of whether the targets were taught or presented as IF. Additionally, conditions in which IF stimuli were presented resulted in the same amount of time spent in teaching, on average, with up to three times the amount of information acquired. Not only did more acquisition occur in the same amount of time, but new target behaviors were acquired without being directly taught, which increased the efficiency of the procedure. Given the need for efficient programming for children with autism and other disabilities (Wolery et al., 1991), procedures such as IF warrant further study.

Although IF can be effective in the acquisition of untaught targets, the behavioral mechanism that allows this to occur is unclear. Some researchers have proposed that learners may be covertly echoing the IF (Carroll & Kodak, 2015; Loughrey, Betz, Majdalany, & Nicholson, 2014; Nottingham et al., 2017; Vladescu & Kodak, 2013). It is thought that learners are able to covertly echo due to a generalized imitative repertoire (Carroll & Kodak, 2015; Nottingham et al., 2017; Reichow & Wolery, 2011; Vladescu & Kodak, 2013), specifically a generalized echoic repertoire. That is, echoic behavior is maintained by a history of reinforcement for responding (i.e., making sounds) that matches sounds in the environment (Schlinger, 1995). This hypothesis came from the observation that learners echoed the experimenter’s IF on 85%–100% of opportunities during research sessions (Nottingham et al., 2017; Vladescu & Kodak, 2013). Even when learners did not overtly echo an IF statement, they may have done so covertly.

In Verbal Behavior, Skinner stated, “We pick up a large part of our verbal repertoire by echoing the behavior of others under circumstances which eventually control the behavior non-echoically” (Skinner, 1957, p. 62). Horne and Lowe (1996) elaborated on the idea that echoics mediate later responding. As infants, humans first learn listener behavior. Then we learn to overtly echo; next we learn to tact, and then to covertly echo (Horne & Lowe, 1996). All of these responses are maintained by a variety of consequences. Learning to echo is a critical milestone in childhood development, not only because it shortcuts the lengthy process of shaping that would otherwise be required for every utterance in one’s repertoire (Skinner, 1957, p. 56), but also because it allows the listener to transform a verbal stimulus (i.e., the antecedent verbal stimulus that was heard) into a verbal response (i.e., the echoic response). Through a history of listener, echoic, and tact training in early childhood, a child may hear a caregiver tact a nonverbal stimulus, orient to the nonverbal stimulus, echo the verbal response made by the caregiver, and then later be able to tact a similar stimulus in the future. When children learn to respond to stimuli in a bidirectional manner, the higher order operant known as naming (Horne & Lowe, 1996) or bidirectional naming (Miguel, 2016) is said to have developed. For example, a toddler may hear his mother say, “That noisy bird is looking for a girlfriend.” A few days later, the toddler says, “That bird is looking for a girlfriend,” when he hears a bird singing. Naming is not restricted to listener-tact relations, though. Intraverbal bidirectional naming (Miguel, 2016) occurs when a child hears a sequence of verbal utterances that go together, such as “I love you to the moon and back.” Later, the mother begins to say, “I love you,” and the child finishes by saying, “to the moon and back.”

This echoic response is referred to as “mediation” and is thought to aid in remembering information at a later time (Paulesu, Frith, & Frackowiak, 1993). That is, a covert echoic is made upon hearing new information. Similar antecedent stimuli may then evoke the same topography of response in the future. Schlinger (2008) asserts that humans will echo, either overtly or covertly, when they learn new information. Although it is difficult to get reliable data on the occurrence of covert echoing in remembering and learning tasks, there is evidence that blocking overt rehearsal can lead to a decrease in learning (Gutierrez, 2006; Keller, Cowan, & Saults, 1995; Parsons, Taylor, & Joyce, 1981).

Gutierrez (2006) successfully used a mediation-blocking task, singing the song “The Wheels on the Bus,” to prevent covert and overt rehearsal from occurring during a sequencing task with adults. Compared to trials in which participants engaged in a mediation response (e.g., overt or covert echoics), the trials with the mediation-blocking task resulted in a decrease in accurate responding. This suggests mediation is necessary to retain information. It is unclear whether overt mediation-blocking responses can preclude covert mediation from occurring, but it is unlikely that covert and overt speech can occur at the same time (Palmer, 2009).

If covert echoing is the mechanism that accounts for how IF targets can be acquired without direct teaching, a mediation-blocking task should disrupt learning the IF relations. In other words, by having participants engage in vocal behavior immediately after hearing the IF, they would be prevented from covertly echoing. It is possible, however, that any distracting task could disrupt learning the IF targets, even if the learner could still engage in covert echoic behavior. The purpose of the proposed investigation is to compare the effects of a vocal mediation-blocking task to the effects of a motor-distraction task on the acquisition of IF targets to test the hypothesis that covert echoing facilitates indirect learning.

Method

Participants, Setting, and Materials

Three children, aged 3 to 6 years, participated in this study. All were receiving 30 hr per week of early intensive behavioral intervention (EIBI) at a university-based clinic. Mason was 3 years, 3 months old at the commencement of the study and had been receiving EIBI services for 13 months. He had a dual diagnosis of autism spectrum disorder (ASD) and avoidance/restrictive food intake disorder. He scored 81.5 (out of a possible 170 points) on the Verbal Behavior Milestones Assessment and Placement Program (VB-MAPP; Sundberg, 2008), demonstrating skills in Level 2 for both the tact and intraverbal sections.

The second participant, Zach, was 3 years, 3 months old and had been receiving EIBI for 5 months. He had a dual diagnosis of ASD and avoidant/restrictive food intake disorder. He scored in Level 2 and Level 3 in the tact and intraverbal sections, respectively, of the VB-MAPP, with a total score of 101.5.

The final participant, Austin, was 6 years old at the commencement of the project and had been receiving EIBI intermittently for almost 2 years. He had multiple diagnoses, including ASD; unspecified disruptive, impulse-control, and conduct disorder; and stereotypic movement disorder with self-injury. He scored in Level 3 in both the tact and intraverbal sections of the VB-MAPP, with a total score of 142.5.

Sessions took place in a small treatment room at the university-based autism treatment clinic that participants regularly attended. Materials included data sheets, pens, a timer, a video camera, and colored cards designated for each condition. The teaching materials were pictures of cartoon characters or scenes corresponding to each primary target. Each primary target (i.e., a tact target to be directly taught) was paired with one related IF target (i.e., an intraverbal target to be probed). The difficulty of the targets was equated across conditions by grouping targets with similar features and numbers of syllables in the spoken responses into triads and then randomly assigning each target in the triad to one of the three experimental conditions using a random number generator. Four primary target/IF pairs were assigned to each condition. See Table 1 for a summary of the specific targets for each participant.

Table 1.

Taught Targets and IF Statements per Condition

	IF	IF With Vocal Mediation-Blocking Task	IF With Motor-Distraction Task
Mason
Tact Responses	Spiderman is climbing a wall. Aurora is lying down. Timon is crying small tears. Abu is eating some bread.	Ariel is reading a book. Batman is punching a man. Jessie is yelling loudly. Peter Pan is sword fighting.	Stitch is playing a small guitar. Superman is flying fast. Patrick is licking ice cream. SpongeBob is blowing bubbles.
IF Statementsa	Spiderman is from (New York). Aurora is (sleeping). Timon only (eats bugs). Abu is friends with (Aladdin).	Ariel is a (mermaid). Batman’s name is (Bruce Wayne). Jessie likes to (yodel). Peter Pan lives in (Neverland).	Stitch likes music by (Elvis). Superman is from (Krypton). Patrick is a (sea star). SpongeBob’s house is a (pineapple).
Zach
Tact Responses	Marlin Bruce Mater Donatello	Jessie Stitch The Flash Wonder Woman	Pumba Thor The Hulk Cinderella
IF Statementsa	Marlin is a (clown fish). Bruce is friends with (Nemo). Mater tows other (cars). Donatello lives in (the sewer).	Jessie loves to (yodel). Stitch listens to (Elvis). The Flash is very (fast). Wonder Woman’s name is (Diana).	Pumba only eats (bugs). Thor fights with a (hammer). The Hulk has super (strength). Cinderella wears a (glass slipper).
Austin
Tact Responses	Mulan is climbing a pole. Abu is eating bread. Aurora is taking a nap. Charizard is blowing out fire.	Roadrunner is running fast. Peter Pan is sword fighting. Stitch is strumming a guitar. Pocahontas is paddling slow.	Squirtle is spitting water. Dopey is holding a pick. Timon is hula dancing. Wonder Woman is crossing her arms.
IF Statementsa	Mulan dresses like (a boy). Abu is friends with (Aladdin). Aurora likes (roses). Charizard fights other (animals).	Roadrunner runs from (coyotes). Peter Pan lives in (Neverland). Stitch listens to (Elvis). Pocahontas has a (pet raccoon).	Squirtle is a (Pokémon). Dopey works in a (diamond mine). Timon likes to (eat bugs). Wonder Woman’s name is (Diana).

^aThe words in parentheses are the portion of the statement that was left for the participants to fill in during the IF probes.

Preexperimental Interviews and Assessments

The lead author conducted interviews with the participants, caregivers, and case managers to identify potential teaching targets and reinforcers. All potential targets were then probed to ensure the participants did not respond correctly to either the primary (i.e., tacts) or IF (i.e., intraverbal) targets. The case managers overseeing the participants’ EIBI programs reported that praise functioned as a reinforcer for all participants, and informal observations of EIBI sessions verified these reports.

Next, the lead author conducted an assessment to determine which tasks would be presented as distraction tasks during the experimental conditions. The purpose of this assessment was to select a vocal task that would compete with echoic responses and a motor task that would not compete with echoic responses for each participant. For the vocal mediation-blocking task, the participant was prompted to engage in an extended vocal task (e.g., sing a song, recite the alphabet). The researcher periodically presented echoic trials (e.g., “Say ‘dog.’”) while the learner was engaged in the vocal task. A task was selected to be the vocal mediation-blocking task for each participant if he engaged in the task for at least 10 s and did not respond to the majority of echoic trials. To assess motor-distraction tasks, the therapist presented echoic trials periodically while the participant engaged with a motor task (e.g., doing a puzzle). If the participant engaged in the task for at least 10 s and responded to the majority of echoic trials during the assessment, that motor task was selected for use in the study. See Table 2 for a summary of the specific vocal and motor tasks selected for use during the training conditions for each participant.

Table 2.

Mediation-Blocking and Motor-Distraction Tasks per Participant

	Vocal Mediation-Blocking Tasks	Motor-Distraction Tasks
Mason	Singing the alphabet Singing “Twinkle, Twinkle, Little Star” Singing “Bingo”	Playing a bongo drum Playing with a rain stick Playing with blocks
Zach	Singing the alphabet Singing “Twinkle, Twinkle, Little Star” Counting out loud	Playing with stuffed animals Playing with cars Playing with Play-Doh
Austin	Singing the alphabet Counting out loud Spelling written words	Putting together a puzzle Playing with cars

Experimental Design

An adapted alternating-treatments design (Sindelar, Rosenberg, & Wilson, 1985) embedded in a nonconcurrent multiple probe across participants (Tawney & Gast, 1984) was used to determine the effect of vocal mediation-blocking and motor-distraction tasks on the acquisition of IF targets. The teaching package for the primary targets was evaluated in a nonconcurrent multiple-baseline design (Harvey, May, & Kennedy, 2004) in which the number of baseline sessions was staggered for each participant. Due to participant availability and time constraints, a concurrent multiple-baseline design was not possible. Although this is considered a weaker research design (Gast & Ledford, 2014), the acquisition of the primary targets was not a central interest of this study. The untrained relations were evaluated in a probe design because previous researchers have recommended including the minimal amount of probe sessions and teaching conditions to reduce exposure to the IF and better capture whether the learner is acquiring the IF targets incidentally (Horner & Baer, 1978; Werts et al., 2003). The experimental question of interest in this study was the comparison of the three conditions: IF, IF with a vocal mediation-blocking task, and IF with a motor-distraction task, as described in the following sections. The adapted alternating-treatments design is useful for comparing treatments when the different instructional target sets are equally difficult to acquire (Sindelar et al., 1985). The conditions were presented in a rapidly alternating fashion, with the order predetermined through a random number generator.

Dependent Variables and Data Collection

Data were collected by graduate students who demonstrated 95% accuracy on a procedural integrity checklist. The two primary dependent variables were the percentage of correct responses on the taught target trials and the percentage of correct responses on the untaught IF target probe trials. Data were also collected on the occurrence of overt echoing of the IF statement and the occurrence of performing the vocal mediation-blocking or motor-distraction task during each trial. All measures were calculated by dividing the number of correct responses or occurrences by the number of opportunities to respond and multiplying by 100.

Interobserver Agreement (IOA) and Treatment Integrity

A second independent observer scored data on at least 33% of sessions across all participants and conditions. The first author compared data from the two observers, scoring agreements, disagreements, and the total number of trials for each session. IOA was calculated by dividing the number of agreements by the total number of trials and multiplying by 100. IOA data were collected in 47% of Mason’s sessions, resulting in 98.3% agreement (range 83.30%–100%). IOA was conducted in 41% of Zach’s sessions. The average score was 96.80% agreement (range 83.30%–100%). For Austin, IOA was conducted in 45% of sessions, with an average of 97.40% agreement (range 83.3%–100%).

Data were also collected on the accuracy of the instructor’s implementation of the experimental procedures using a treatment integrity checklist (Table 3), either at the time of the session or from a video recording of the session. The lead author calculated treatment integrity by dividing the number of correctly implemented components of the checklist by the total number of components and multiplying by 100. Treatment integrity data were collected for 47% of Mason’s sessions, resulting in an average of 98.40% (range 90%–100%) correctly implemented steps. For Zach’s sessions, treatment integrity data were collected for 42% of sessions, with an average score of 98.80% fidelity (range 87.50%–100%). For Austin’s sessions, treatment integrity data were collected for 45% of sessions, with a mean of 98.80% fidelity (range 90.90%–100%).

Table 3.

Treatment Integrity Checklist Example

General Procedures	Condition 1		Condition 2		Condition 3
Did the therapist have the correct materials ready for the session (tact targets, data sheets, distractor materials, etc.)?	Yes	No	Yes	No	Yes	No
Did the therapist run an MSWO before running the first condition?	Yes	No	Yes	No	Yes	No
Did the therapist indicate the condition (color card/statement) before bringing out the materials for each condition and clean up the materials at the end?	Yes	No	Yes	No	Yes	No
Did the therapist start the condition timer when giving the first instruction and stop the condition timer when delivering the reinforcer following the 12th trial?	Yes	No	Yes	No	Yes	No
Did the therapist give the correct instructions, as stated on the data sheet?	Yes	No	Yes	No	Yes	No
Did the therapist provide the correct IF, as stated on the data sheet?	Yes	No	Yes	No	Yes	No
Did the therapist collect data immediately after each trial?	Yes	No	Yes	No	Yes	No
Did the therapist provide a 2-min break between each condition and a 10-min break between sessions?	Yes	No	Yes	No	Yes	No
Total per block:
Baseline and Probe Procedures	Condition 1		Condition 2		Condition 3
Did the therapist wait 5 s for response initiation and 5 s after initiation for response completion?	Yes	No	Yes	No	Yes	No
Did the therapist provide an edible and behavior-specific praise on a VI 30-s schedule for appropriate sitting, eye contact, and so on?	Yes	No	Yes	No	Yes	No
Did the therapist refrain from prompting, error correction, and differential reinforcement?	Yes	No	Yes	No	Yes	No
Total per block:
Teaching Procedures	Condition 1		Condition 2		Condition 3
Did the therapist prompt and fade prompts as specified by the protocol?	Yes	No	Yes	No	Yes	No
Did the therapist provide the appropriate consequence as specified by the protocol, for correct, prompted, incorrect, and no response?	Yes	No	Yes	No	Yes	No
If the participant initiated a response within 3 s of the instruction, did the therapist wait 5 s for response completion?	Yes	No	Yes	No	Yes	No
Did the therapist implement the ITT or present the correct distractor task after the IF statement, as stated on the data sheet, for 10 s?	Yes	No	Yes	No	Yes	No
Did the therapist record if the participant repeated any part of the IF statement and if the participant engaged in the vocal or motor task?	Yes	No	Yes	No	Yes	No
Total per block:
Problem Behavior	Condition 1		Condition 2		Condition 3
Did the therapist block and redirect any problem behavior?	Yes	No	Yes	No	Yes	No
Did the therapist continue teaching?	Yes	No	Yes	No	Yes	No
Did the therapist record the frequency of problem behavior throughout the session?	Yes	No	Yes	No	Yes	No
Total per block:

MSWO = multiple-stimulus without-replacement preference assessment.

Procedure

Sessions were conducted 3 to 5 days per week. Each condition consisted of 12 total trials, comprising three trials for each of the four primary/IF pairs assigned to that condition. These trials were presented in a semirandom order, with the same primary/IF target pair presented no more than twice in a row. Across all conditions, teaching sessions were implemented until participants demonstrated mastery at or above 80% correct, independent responses per session for three consecutive sessions.

Baseline Primary Targets

At the onset of each session, the therapist presented a color card corresponding to the condition in effect. Next, trials began. In Zach’s sessions, the instructor presented a picture card, asked “Who is it?” and waited 5 s for a response. A correct response was scored if Zach gave a one-word response corresponding to the cartoon character in the picture (e.g., “Marlin.”). In Mason’s and Austin’s sessions, the instructor presented a picture card, and asked “What is he/she doing?” or “What’s happening?” and waited 5 s for a response. A correct response was scored if the participant tacted with a noun, verb, and an adjective or adverb (e.g., “Spiderman is climbing the wall.”). No programmed consequences were delivered for correct or incorrect responses. The instructor praised appropriate attending behaviors on a variable interval (VI) 30-s schedule.

IF Probes

IF probe sessions were conducted in the same manner as baseline of primary targets described previously. However, instead of presenting trials of the primary targets (i.e., tacts), the instructor presented trials of each IF statement as an intraverbal fill-in-the-blank statement (e.g., “A dog says _____.”). During baseline, the IF probe sessions occurred after the first and last baseline session for each condition with a total of two probe sessions per condition. IF probes also occurred at fixed points in between sessions of the tact-teaching conditions. The probes occurred after the first session of each condition in teaching and again once the learner met mastery criterion for the tact targets for one session (i.e., one session at 80% correct, independent responding) and each subsequent session at 80%. These occurred after the first session rather than after meeting the full mastery criterion of three sessions because previous research determined IF was acquired at the same rate as taught targets (Werts, Caldwell, & Wolery, 2003).

IF

At the onset of these sessions, the instructor presented the color card designated for this condition and the vocal statement “I’m going to ask you some questions.” Trials were implemented as described for the baseline condition with three exceptions. First, an echoic prompt (e.g., “Dog.”) was delivered immediately (i.e., 0-s delay) after the visual and verbal antecedents were presented for the primary target (i.e., tact trial) for three sessions. Prompts were faded in subsequent sessions using a constant time delay (Wolery, Ault, & Doyle, 1992) of 3 s. Second, the instructor immediately praised correct responses. Third, the instructor presented the IF statement (e.g., “A dog says ‘woof.’”) following the praise. If the learner did not respond or responded incorrectly to the primary tact trial, an error correction procedure was employed, followed by the IF statement. The error correction procedure was as follows: Following an incorrect response, the materials were removed for 1 s. The trial was then re-presented (e.g., “What is he doing?” while displaying the picture) along with an immediate prompt for the correct response (e.g., “Stitch is playing guitar.”). After the participant made the correct response, the instructor gave neutral feedback (e.g., “That’s better.”). No programmed consequences were delivered if the learner overtly echoed the IF statement; however, it was recorded on the data sheet. After the presentation of the IF statement, a 10-s inter-trial interval began, which was equivalent to the same amount of time in which participants engaged in the vocal or motor tasks in the next two conditions.

The IF condition was designed to evaluate whether the participants could learn from IF. If participants were unable to learn the IF targets in either of the blocking conditions, we would not know whether distraction was causing the disruption in learning or whether these particular participants simply could not learn from IF.

IF + Vocal Mediation-Blocking Task

At the onset of these sessions, the instructor presented the color card designated for the condition along with the vocal statement “I’m going to ask you some questions, and then you are going to say/sing something.” The instructor then placed colored cards that corresponded to the prearranged vocal tasks selected from the preexperimental assessment in front of the participant and had the participant choose which vocal task to engage in for the session. This condition was conducted in the same manner as the IF condition with one notable exception. Immediately following the IF statement, the instructor told the participant to engage in the vocal mediation-blocking task (e.g., “Sing the ABC song.”). After the learner engaged in the vocal task for 10 s, the instructor praised the response and then moved on to the next trial.

IF + Motor-Distraction Task

At the onset of these sessions, the instructor presented the color card designated for the condition along with the vocal statement “I’m going to ask you some questions, and then you are going to play with something.” The instructor then placed the materials for the predetermined motor-distraction tasks in front of the learner. Whichever item the participant chose was used as the distractor task throughout the session. This condition was identical to the IF + vocal mediation-blocking task condition, with one exception. Instead of presenting a demand to engage in a vocal mediation-blocking task, the instructor told the participant to engage in a motor-distraction task while handing the necessary materials to the participant (e.g., “Here, do this puzzle,” while handing a puzzle to participant).

The purpose of this condition was to determine whether any distracting task would impede the acquisition of IF statements. The tasks in this condition were selected if they did not compete with echoic behavior—that is, the participants previously demonstrated they could engage in echoic responses while completing the motor task. If acquisition of IF statements occurred in this condition, but not in the IF + verbal mediation-blocking task condition, it would lend support to the hypothesis that covert echoing was meditating subsequent intraverbal behavior.

Results

Figure 1 depicts the primary dependent measures for all three participants. For each participant, the upper and lower panels show the percentage of correct, independent responses per session for the taught tact targets (upper panel) and the intraverbal fill-in-the-blank IF probes (lower panel). Both panels depict these data across the three experimental conditions: IF, IF + vocal mediation-blocking task, and IF + motor-distraction task. Mason’s data are displayed in the top two panels, Zach’s in the middle two panels, and Austin’s in the bottom two panels.

Fig. 1.

Percentage of correct, independent responses to the taught targets and IF probes across three participants

Mason

Taught Tact Targets

For the taught tact targets, Mason emitted 0% correct responding during baseline and the first two sessions of teaching (i.e., prompted at 0-s delay) for all three experimental conditions. Once the prompt was faded to 3 s, Mason began emitting correct responses to the taught tact targets across all conditions. He reached mastery criterion for the taught tact targets in five sessions for the IF and IF + vocal mediation-blocking task conditions and in six sessions for the IF + motor-distraction task condition.

IF Probe Targets

During IF probes, Mason’s responding stayed at 0% correct during baseline and after the first two teaching sessions with 0-s delay. Following the change to the 3-s delay for the tact targets, Mason’s responding during the IF probes initially increased in the IF and IF + motor-distraction tasks conditions, then quickly returned to and stabilized at 0% correct across all conditions. After all tact targets had been mastered, Mason continued at 0% correct responding to the IF probes. Mason’s case manager reported that he sometimes demonstrated noncompliance with intraverbal tasks during his EIBI sessions. Therefore, we changed the response form to a match to sample, as suggested by his clinical team. In this procedure, each of the IF statements was printed on a piece of paper with the target response left as a blank (e.g., if the IF statement was “Abu is friends with Aladdin,” the paper said “Abu is friends with _____.”). The four target intraverbal responses were written on separate pieces of paper and placed on the table. The experimenter read the IF statement to Mason and asked him to match the statement to the target response on the table. Upon implementing this procedure, Mason’s responding immediately reached mastery criterion across all conditions. We then reintroduced the original procedure using vocal antecedents and responses only. Mason demonstrated mastery of all fill-in-the-blank IF probe targets.

Zach

Taught Tact Targets

Zach emitted 0% correct responses in the taught tact target trials throughout six sessions of baseline and two sessions of teaching at a 0-s time delay. When the time delay was increased to 3 s, Zach began emitting correct responses to taught tact targets across all conditions. He reached mastery criterion for the taught tact targets in 8 sessions for the IF condition, 9 sessions for the IF + motor-distraction task condition, and 10 sessions for the IF + vocal mediation-blocking task condition.

IF Probe Targets

During IF probes, Zach’s responding stayed at 0% correct during baseline. After the first session of teaching the tact targets at the 0-s delay, he responded correctly to 1 of the 12 probe trials (8.33%) for the targets in the IF + vocal mediation-blocking task condition. Following the change to the 3-s delay for the taught tact targets, Zach’s responding during the IF probes began increasing, then decreased slightly and stabilized at 50% correct across all conditions. In addition to a decrease in accurate responding, an increase in problem behaviors (e.g., negative vocalizations, slapping the cards) was observed during these sessions. Assessment data collected during his regular EIBI sessions indicated that Zach’s problem behavior was maintained by positive reinforcement in the form of attention from adults. Because there were no programmed consequences for probe trials, Zach’s clinical team hypothesized that the lack of attention following responding was functioning as an establishing operation, evoking problem behavior and possibly interfering with responding to the probe trials. Thus, in Session 48, neutral feedback (e.g., saying “OK” or “thanks” after each response) was introduced contingent on any correct or incorrect response to probe trials in an attempt to abolish the reinforcing value of attention for problem behavior and potentially increase compliance with the task. Although a reversal of the neutral feedback was not conducted, problem behavior did decrease upon its introduction, which apparently resulted in an increase in mastery of targets in the IF condition after three sessions and an increase to 75% correct responding for targets in the other two conditions. Next, we implemented the teaching procedures for the untaught IF probe targets, resulting in mastery of all remaining IF probe targets across the IF + vocal mediation-blocking and IF + motor-distraction task conditions.

Austin

Taught Tact Targets

Austin emitted 0% correct responding on the taught tact targets throughout nine baseline sessions and two teaching sessions with a 0-s delay. When the time delay was increased to 3 s, Austin began emitting correct responses to the taught tact targets across all conditions. He reached mastery criterion for the taught tact targets in six sessions for the IF condition and in seven sessions for the IF + vocal mediation-blocking task and IF + motor-distraction task conditions.

IF Probe Targets

During IF probes, Austin’s responding stayed at 0% correct during baseline and after the first session of teaching the tact targets at the 0-s delay. Following the change to the 3-s delay for the tact targets, Austin’s responding to the IF intraverbal probes steadily increased. He learned all the IF probe targets without direct teaching, although acquisition of the IF probe targets took one session longer than for the taught targets in the IF and IF + vocal mediation-blocking task conditions.

Secondary Measures

Overt Echoing

Figure 2 shows the percentage of trials, across conditions, in which each participant repeated at least some part of the IF statement. Mason echoed the IF statement on 49.16% of trials in the IF condition, 8.33% of trials in the IF + vocal mediation-blocking task condition, and 37.12% of trials in the motor-distraction task condition. Zach echoed the IF statement on 2.38% of trials in the IF condition, 1.11% in the IF + vocal mediation-blocking task condition, and 2.98% in the IF + motor-distraction task condition. Austin echoed the IF statement on 1.04% of trials in the IF condition and on 1.19% of the trials in both the IF + vocal mediation-blocking task and IF + motor-distraction task conditions.

Fig. 2.

Average percentage across trials of repeating the IF stimulus for the IF, verbal mediation-blocking task, and motor-distraction task conditions for all participants

Compliance with the Distraction Task

Figure 3 shows the percentage of trials during which the participants engaged in the vocal mediation-blocking task or motor-distraction task for the entire 10 s. Mason correctly performed the vocal mediation-blocking task for 83.33% of trials and the motor-distraction task for 90.90% of trials. Zach correctly performed the vocal mediation-blocking task for 62.78% of trials and the motor-distraction task for 86.90% of trials. Austin correctly performed the vocal mediation-blocking task for 84.5% of trials and the motor-distraction task for 92.85% of trials.

Fig. 3.

Average percentage across trials of correct responding to the verbal mediation-blocking and motor-distraction tasks for all participants

Discussion

The purpose of this study was to compare the effects of a vocal mediation-blocking task with the effects of a motor-distraction task on the acquisition of IF targets to evaluate the hypothesis that overt or covert echoing is the mechanism of action underlying the acquisition of IF targets. Although participants engaged in the vocal mediation-blocking task during the majority of opportunities, both participants (i.e., Mason and Austin) who acquired IF targets without direct teaching did so across all conditions, including the mediation-blocking task condition. Additionally, all three participants acquired the IF responses at the same rate across conditions. If the vocal mediation-blocking task was successful at preventing echoing, this suggests covert or overt echoing may not be necessary to acquire IF targets.

These results contrast with Gutierrez’s (2006) findings, in which participants were largely unable to complete a sequencing task without the aid of verbal mediation strategies. There are some differences between the current study and Gutierrez’s study that could account for this contrast. In Gutierrez’s study, six adult participants were asked to arrange a sequence of pictures that corresponded to spoken Mandarin words. The experimenter said four words in Mandarin (e.g., pen, cup, water, fork), and after a delay, the participants were asked to arrange the corresponding pictures in the order the Mandarin words were spoken.

The different tasks (i.e., sequencing vs. intraverbaling) could account for the discrepancy in results between Gutierrez’s (2006) and the current study. Gutierrez specifically taught the participants to engage in an echoic/tact response as a mediation strategy. Tacts have different sources of antecedent control than do intraverbals. Intraverbal fill-ins in particular are likely under very strong stimulus control of the “running start” (Skinner, 1957, p. 72) of the antecedent verbal stimuli and thus may not require verbal mediation. Moreover, the participants in this study heard the entire intraverbal chain dozens of times during the intervention phase, which likely served to strengthen the relation between the antecedent and the response during IF probes. Future research can examine the effects of mediation blocking on different kinds of verbal operants taught in the context of IF.

It is also possible that the complexity of the response impacted the need for a mediating strategy in the present study. The participants only needed to fill in the very last part of the IF statement using a word or phrase that was one to three syllables. Although mediation is often helpful for complex sequences of behaviors, such as remember long lists of numbers or multiple steps in a chain, these mediation tactics might not be needed for simpler tasks (Michael, Palmer, & Sundberg, 2011). Future researchers should examine the impact of these conditions on more difficult or complex tasks.

Another possible explanation for the differing results from the two studies relates to the prior history with the stimuli and responses. The participants in Gutierrez’s (2006) study likely had little if any prior exposure to the Mandarin words. In contrast, it is likely the participants in this study had heard some or all of the particular words in the IF statements at some earlier point in time, even though they were not able to emit the particular sequence of utterances at the onset of the study. Familiarity with the words may have facilitated acquisition, as has been shown in previous research (Fields, Arntzen, Nartey, & Eilifsen, 2012; Holth & Arntzen, 1998).

The Effectiveness of the Vocal Mediation-Blocking Task on Suppressing Echoics

The limitation of greatest concern in the present study is that we cannot know for sure that the vocal mediation-blocking task prevented the occurrence of covert echoing. The preexperimental assessment attempted to address this concern by selecting a vocal task that the participants could not simultaneously engage in while responding to echoic trials. However, it is possible the participants were unable to engage in both behaviors at the same time during the preassessment but learned to do so through continuous exposure and practice effects during the course of the study.

Timing

Another explanation for how mediational echoics could have occurred despite the vocal blocking task lies in the timing of the presentation of the vocal mediation-blocking task. Although researchers attempted to present the distraction task immediately after the completion of the IF statement, there was still a latency of approximately 1 s between the occurrence of the IF statement and the onset of the blocking response. It is possible the participants were able to covertly echo the statement so quickly that it occurred before the initiation of the vocal mediation-blocking task (e.g., they echoed it almost simultaneously to the instructor saying it). Future studies could assess participants’ latency to overt echoic responding to better ensure the mediation-blocking task is actually blocking covert echoics.

Compliance with the Vocal Mediation-Blocking Task

It is noteworthy that the participants did not engage in the vocal mediation-blocking task for the entire 10-s interval in 100% of trials. Mason engaged in the response in 83.33% trials of the IF + vocal mediation-blocking condition, Zach in 62.78%, and Austin in 84.5%. It is possible they engaged in covert echoing on the trials in which they did not engage in the vocal mediation-blocking task. However, previous research has shown that participants acquire the IF targets at the same rate as the taught targets, suggesting that the number of opportunities to respond is a key variable (Clough, Meyer, & Miguel, 2016). Because the participants in this study did engage in the vocal mediation-blocking task in the majority of trials, this would decrease the number of opportunities participants had to echo. Given that, one would expect slower acquisition of the IF targets in the vocal mediation-blocking condition as compared to the other conditions; however, this did not occur. The present study can be expanded and improved upon by including preexperimental training sessions to increase participants’ compliance with requests to both the vocal mediation-blocking tasks and the motor-distraction tasks. This would encourage more fluent responding to the task (i.e., engaging in the task for the entire 10 s, decreasing the latency to beginning the task).

Occurrence of Overt Echoics

In addition, it is difficult to determine if the vocal mediation-blocking task was successful because the participants did not reliably engage in overt echoing of the IF statements, even when it was possible (i.e., during the IF and IF + motor-distraction task conditions). This is in contrast with previous studies (Nottingham et al., 2017; Vladescu & Kodak, 2013) in which participants engaged in a high rate of unprompted overt echoing. In this study, Zach and Austin engaged in little if any overt echoing; Mason repeated the statement in fewer than 50% of opportunities. Future studies may consider prompting participants to echo the IF statement in the IF and IF + motor-distraction task conditions to ensure the response occurs. This would be particularly important if participants did not acquire the IF targets in the IF condition, which did not occur in this study.

Only one participant, Mason, engaged in overt echoing during sessions. Interestingly, he echoed the IF statement more during the IF and the IF + motor-distraction tasks than he did in the IF + vocal mediation-blocking condition (IF: 49.16%; IF + vocal mediation-blocking task: 8.33%; IF + motor-distraction task: 37.12%). These data seem to suggest that the vocal mediation-blocking task may have suppressed echoing for Mason, although we cannot know if this would generalize to covert echoing. It seems unlikely that the lack of echoing the IF statement in the mediation-blocking task condition had any effect on acquisition because Mason was able to correctly match IF statements in the match-to-sample arrangement without additional teaching in all of the conditions, including the IF targets from the mediation-blocking task condition. If vocal mediation was required, acquisition during the match-to-sample procedure in the mediation-blocking task condition would have occurred at a slower rate than the other two conditions, if it occurred at all.

Although it was difficult to determine whether covert echoics were occurring, two participants (Austin and Mason) were observed engaging in an additional problem-solving behavior during the IF probe sessions. On one to three occasions, when presented with the intraverbal fill-in-the-blank statement (e.g., “Abu is friends with _____.”), Mason and Austin used the directly taught tact target to facilitate their responding. They overtly said the tact target (e.g., “Abu is eating bread.”) and then answered the fill-in-the-blank statement (e.g., “Abu is friends with Aladdin.”). This suggests the participants learned an intraverbal chain (Skinner, 1957), which may have assisted in their ability to correctly respond to the intraverbal fill-in-the-blank statement.

The participants in the present study were able to demonstrate more advanced verbal skills, such as the problem solving described previously. In a recent review of studies on IF, Albarran and Sandbank (2019) reported on 226 participants across 48 journal articles and 6 dissertations. Of the participants included in these studies, 64% (range 0%–100%) gained IF targets. Participants included young children, older adults, adolescents, and adults; the youngest was 3 years old. All but 10 participants had a diagnosis of an intellectual or developmental disability (e.g., ASD, Down syndrome, attention deficit hyperactivity disorder). When cognitive ability was reported, the mean IQ score was 61 (range 30–130; Albarran & Sandbank, 2019). It would be interesting to examine whether there is any interaction between age or cognitive ability with the ability to learn targets incidentally. However, it would be most important to determine which prerequisite skills facilitate this kind of learning. If a generalized echoic repertoire is necessary to learn IF targets, one would expect early learners whose echoic repertoires were only beginning to be formed would have difficulty learning IF targets, whereas those with a generalized echoic repertoire would learn them with ease.

Conclusion

Although the current study did not support the hypothesis that covert echoing is required for the acquisition of IF statements, it is still possible a generalized delayed imitative repertoire (Baer & Sherman, 1964) could be the behavioral mechanism that explains the acquisition of IF statements. It is possible the verbal relation of hearing the IF statement and correctly recalling the information later is part of a larger generalized response class. The participants had an extensive history of repeating the therapist’s model, both within the context of this study with the primary targets and in their regular EIBI sessions. The contingencies for the IF targets may have been indiscriminate from those of the primary targets (Carroll & Kodak, 2015). Even without mediation (i.e., immediately echoing or self-echoing upon initially hearing a vocal statement), it is possible to hear a statement at one point in time and then vocally repeat that statement at a later point in time. This generalized repertoire could be developed by intermittent social reinforcement for appropriately recalling information across various contexts, such as something that happened earlier in the day or the name of a new acquaintance. Consider an example in which you meet a new coworker first thing in the morning. She introduces herself by name, and you return an introduction, but never repeat her name. Later, at lunch, someone asks you what her name is. For an individual with a generalized delayed imitative repertoire, this question could function as a discriminative stimulus (S^D) to finally potentiate a response that had been bubbling near the surface but may never yet have been emitted (Palmer, 2006, p. 13).

Some individuals with a generalized echoic repertoire may still need to engage in additional problem-solving behaviors to answer this question. They may scroll through the alphabet, softly saying names starting with different letters until prompting themselves to a name with the right starting letter, or rhyming sounds, and eventually, they go from whispering and testing incorrect names to shouting the correct name out loud. In this case, the response of saying the coworker’s name is emitted partly as an intraverbal, in response to the question, and partly as a self-echoic, in response to the problem-solving strategy. This is an example of joint stimulus control (Lowenkron, 1998; Palmer, 2006).

Across both examples of recalling the new coworker’s name, however, the controlling role of having heard the name in the first place cannot be discounted simply because it was temporally distant. If we observe behavior from a molar perspective, we can observe responses as a pattern of activities that occur over time Baum, 2005. From this perspective, the stimuli that immediately precede a response are not the only controlling factors for determining why specific responses occur. For instance, in the example of recalling the new coworker’s name, you may have used the overt problem-solving strategy in addition to having previously heard your new coworker’s name. In our study, a participant may have engaged in a self-echoic or covert mediational response at the time of hearing the intraverbal fill-in-the-blank statement during the IF probes. Responding to the verbal antecedent could have been a chain of behaviors that involved hearing the intraverbal S^D, engaging in covert echoic behavior, and overtly speaking the correct answer. Rather than viewing this mediational responding as a separate set of behaviors, it could be viewed simply as part of the larger behavior of responding to the verbal S^D.

Some researchers have suggested the mediational, self-echoic explanation for response acquisition is not parsimonious. Hayes and Barnes-Holmes (2004) suggested relying on an explanation such as a mediation response acquisition is unnecessary and could even be considered a mentalism, an explanation that relies on a mental, unobservable dimension to explain behavior. However, disregarding a mediating behavior, should it be determined as necessary, would preclude practitioners from incorporating critical components of language instruction for learners with ASD and other language delays. To date, there are still many unanswered questions and conflicting research findings. Future research, such as that described herein, is needed to further elucidate the role of overt and possibly covert mediational responses in the acquisition of untrained human operants in general but also specifically during language training with children with autism.

Author Note

Amelia Dressel, School of Behavior Analysis, Florida Institute of Technology, and the Scott Center for Autism Melbourne, Florida; Katie Nicholson, School of Behavior Analysis, Florida Institute of Technology, and the Scott Center for Autism Melbourne, Florida; Kristin M. Albert, School of Behavior Analysis, Florida Institute of Technology, and the Scott Center for Autism Melbourne, Florida; Victoria M. Ryan, School of Behavior Analysis, Florida Institute of Technology, and the Scott Center for Autism Melbourne, Florida.

Amelia Dressel is now at the Knoxville Center for Autism, Knoxville, Tennessee.

This research was conducted by the first author, under the supervision of the second author, at Florida Institute of Technology in partial fulfillment of the thesis requirements for the master of science degree in applied behavior analysis.

Compliance with Ethical Standards

This paper has not been previously published and has not been and will not be submitted elsewhere during the review process.

Conflict of Interest

The authors declare that they have no conflict of interest.

Ethical Approval

All procedures performed in this research were approved by the Institutional Review Board at the Florida Institute of Technology and were in accordance with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.

Informed Consent

Informed consent was obtained from all individual participants included in the study.