Increasing Social Communication by Teaching Texting to Autistic Children

Participants

The participants were four autistic children, between the ages of 8.7 and 10.9 years old, who were all diagnosed according to The Diagnostic and Statistical Manual of Mental Disorders (DS APA, American Psychiatric Association, 2013). All participants attended after-school social skill groups. Requirements of each participant were that they needed to be literate and demonstrate the basic motor skills needed to use a mobile phone. Pretests to assess these skills are discussed below.

The participants were grouped into two dyads based on age, interests, social skill group affiliation, and desire to communicate with each other. The dyad pairings, ages of the participants, genders, ethnicities, expressive verbal age (EVT-III), receptive language age (PPVT-IV), and autism severity (CARS-2) are presented in Table 1.

Table 1 Participant characteristics Dyad 1

Bennett was 10 years and 9 months old at the start of the study (see Table 1). In addition to his autism spectrum disorder (ASD) diagnosis, Bennett also had an ADHD diagnosis. Bennett had difficulty maintaining a back and forth conversation. He usually engaged in monologues about inappropriate topics. Bennett and his texting partner Milo often partnered together in social skill sessions and shared similar interests. Milo, an 8-year, 7-month-old boy, primarily discussed preferred topics and communicated more with therapists as opposed to peers.

Dyad 2

When engaging in conversations, Anna (10.2 years) stood too close to her peers, discussed inappropriate topics, focused primarily on her perseverative interests, and had difficulty in engaging in a back and forth conversations. Veronica (10.8 years) showed some rigidity in terms of conversational topics and activities, and had difficulty responding appropriately to social cues. They had both participated in activities with each other inside the social skill group and had indicated an interest in being friends. When they interacted during activities in the social group, they had difficulty maintaining a back and forth conversation.

Procedure Materials

For the study, one iPhone® version 5 and three Samsung Galaxy phones belonging to the participants’ parents or the participants themselves were used. Additional materials included four guidebooks. The guidebooks focused on content and included a different example of a back and forth texting conversation for two people. Materials also included iPads® used to record the texting training sessions and computers used by the children and the experimenter when conducting sessions over videoconferencing software due to Covid-19 lockdown which occurred during this study. Lastly, the texting intervention was implemented using an application called TextFree®.

Baseline

Baseline sessions, intervention training sessions, generalization texting partner probes, and FaceTime® probes were conducted in two different lounge settings (1.5 m by 3 m) at the social skill program that the participants attended weekly. The participants were exposed to these rooms prior to the study. Each participant was assigned to one of the two rooms through the baseline and training sessions. The first lounge style room contained a couch, a table, and a chair. The second room contained three chairs, a circular table, and a bookshelf. Partway through baseline, due to Covid-19 lockdown, the setting was moved to the children’s respective homes. The timing of the change in settings from the social skill program to the home environment corresponded with the restrictions put in place by the government.

Design

A non-concurrent multiple baseline design across dyads was used to examine the effect of the texting intervention upon appropriate text exchanges between participants. In addition, response generalization probes across verbal conversations over FaceTime® and stimulus generalization probes across parents and siblings were assessed. Probes for stimulus generalization across texting partners and FaceTime® occurred during baseline, post-intervention, and at maintenance.

Pre-baseline Assessments

Prior to taking part in the study, the parents of each participant, as well as the participants themselves, were asked if they wanted to establish a texting relationship with a friend. Next, the participants were asked to read two different sentences presented on a phone in text format to assess their reading ability (i.e., “I like to play with dinosaurs and talk to my friends about them,” “I also have a sister named Chloe who sometimes plays with me”). In addition, the participants were asked to write two different sentences that were presented verbally (i.e., “I like playing sports outside,” “My favorite sport is soccer”).

Baseline Probes

During baseline, the dyads were created based on child preference, and each participant was initially separated and seated in two different lounge-style rooms. They were then each given their phone that had been turned on and given the instruction “send a text to your friend _______.” Their peer’s phone number was already pre-programed into the phone and could be accessed by typing in their friend’s name and taping the number when it appeared below the name. Dyads were given 10 min per each baseline session to simulate how much texting would be done during intervention.

Text Conversation Intervention

The intervention involved teaching the participants to have a back and forth conversation through text. Once again, the dyads were separated and were taught this skill using two sample conversations (see Fig. 1). In each of the sample conversations, two fictional persons created by the experimenter were conversing: Conversation A was between Brad and Kim, and conversation B was between Claire and Luis. The conversations were presented using multiple pictures of a smart phone with one to two novel lines of text presented on the screen in each picture. These sample conversations were centered around two different main topics: Conversation A was about soccer, and conversation B was about movies. In addition, these conversations both began and ended with some form of a greeting (i.e., hi, hello, see you later, bye). Both conversations had a central topic, and both of the conversers in each scripted conversation asked and responded to questions. The two sample conversation books were alternated across the participants. That is, the two participants in each dyad never had the same conversation manual during a single session (session one: Veronica had book A, and Anna had book B; session two: Veronica had book B, and Anna had book A). This was done in an effort to encourage more varied speech.

Fig. 1figure 1

Example of textual conversation between two friends

During intervention, both participants were handed one of the conversation guidebooks and were told to read it. The participants were given 5 min at the start of each 10-min intervention session to read the entire conversation. After reading the sample conversation, the experimenter discussed the characteristics of the conversations with the participants (“In the conversation the friends asked each other questions, responded to their friend’s questions and talked about things that they both liked”). The participants were then instructed to have a text conversation with their peer, similar to the one in the guidebook. The experimenter also provided assistance as needed during the session such as if they had any questions for their friend and reminding them to read what their peer wrote before responding. After 10 min of conversing, if the participants were still texting, the experimenter told them that it was time to end their conversation. Criterion was met when the participants maintained an appropriate back and forth conversation to 100% accuracy on two separate sessions. Then fading was implemented.

Fading

Fading consisted of removing the guidebooks and no longer providing prompting. The participants were then given 10 min to text with each other per session. The fading criterion was set at 80% or above of correct conversational texting on two consecutive sessions.

Booster Sessions

If a participant’s percentage of appropriate texting regressed back to baseline levels during fading after a single session, two booster sessions of intervention were presented. Then, fading was again implemented across two more sessions.

Independent Weekly Texts

After fading, the dyads were instructed to have a conversation with their friend at least once a week independently. Specifically, the experimenter was no longer present while the children texted, but rather examined the conversation after the fact using the permanent product (i.e., pictures of the conversations). The experimenter instructed the parents to help the dyads identify a day each week when they would have a texting conversation. A screenshot of the texting conversation was taken each week by the parent, and the texting conversation was examined in terms of occurrence, appropriate beginning and end to the conversation, appropriate language, the length of the conversation, staying on topic, and both asking and responding to peer questions.

Generalization Texting Partner Probes

These generalization probes were implemented during baseline, following intervention fading, and at 1-month follow-up. Generalization probes involved the same conditions as in baseline except each participant texted another communication partner that they were likely to text in the future.

FaceTime® Probes

These probes also had the same conditions as in baseline, but instead of texting, the participant made a FaceTime® call to their friend. Because FaceTime® was not taught in this study, the steps to access FaceTime® were done by the experimenter or parent. Specifically, this probe accessed the ancillary effects texting with a peer has on the verbal conversation skills via the smart phone.

Follow-up

Follow-up data were collected 1 month following treatment. The conditions during follow-up were identical to baseline.

Measures Dependent Measures

Texting content included the use of greetings at the beginning and end of the conversation, appropriate language, staying on topic, appropriate length of texts, and both asking and responding to peer questions. The text conversations were scored by examining whether the conversation as a whole was contextually appropriate. A summary of the operational definitions for the dependent variables and examples is presented in Table 2.

Table 2 Dependent variables Novel Content

The operational definition of novel content was that the topic was not used in the same conversation and not used in more than two other conversations. The operational conversation was that the content had not been used in a prior conversation.

Ancillary Measure

In order to assess for potential collateral effects (Ledbetter-Cho et al., 2017), we evaluated how the impact of learning to text with a peer subsequently affected the content of verbal back and forth conversations between peers when communicating over FaceTime® (see Table 3 for operational definitions and examples). The conversational content was scored similarly to that of the dependent variables.

Table 3 Ancillary variable Data Analyses Interrater Reliability and Procedural Fidelity

The primary research observer and one secondary observer were trained on how to score the texting conversations for content and how to score the ancillary measures. To assist with scoring, checklists were provided that contained the observational definitions. The secondary observer reviewed the permanent products of 33% of the texting conversations across conditions, along with 33% of the videotapes of baseline, intervention training, generalization texting partner probes, FaceTime® probes, and follow-up sessions for each participant. The primary and secondary observers then compared scores to each other to determine interobserver agreement. If scorers disagreed, they re-watched the videotapes and re-examined the permanent product of that session to resolve discrepancies. Interrater reliability was high across both participants and phases of the study, ranging from 88 to 100%. A summary of interrater reliability across participants can be seen below in Table 4.

Table 4 Interrater reliability

Additionally, two observers who did not participate in the texting intervention assessed procedural integrity in 33% of the sessions across conditions and participants. This was done using the videotapes of the baseline, intervention training, generalization texting partner probes, FaceTime® probes, and follow-up sessions for each participant. The observers received training on how procedures were implemented and were each given a check sheet to use when scoring the presence and absence of each step in the procedure across sessions and participants. Procedural fidelity for all participants ranged from 94 to 100% on average. Mean procedural fidelity for Bennett was 95%, Milo = 95%, Anna = 94%, Veronica = 94%, and Levi = 100%. The only errors involved the experimenter not ending a few of the sessions at exactly ten minutes.

留言 (0)

沒有登入
gif