Method
Participant
The participant selected for this study was a 10-year-old boy diagnosed with ASD. He was receiving intensive behavioral intervention services at the time of the study. The participant met the following inclusion criteria to participate. First, he possessed a sufficient verbal repertoire for the study. He demonstrated this by independently supplying his name, age, and address when requested. Second, the participant was able to follow multi-step instructions. Third, he demonstrated imitation skills, including gross motor actions (e.g., raising hand to notify adult), fine motor actions (e.g., manipulating game controller to stop game), and echoing phrases (e.g., “I need help”). Lastly, the participant’s caregivers indicated he engaged in frequent video game use (i.e., at least 7 hours per week). The participant had not received formal online-safety skill instruction or training prior to the start of the study.
Setting and Materials
All probes and BST sessions were conducted at the center-based clinic where the participant was receiving behavior-analytic services. The materials used during all probe and BST sessions included: a) two PC computers using Microsoft operating systems, b) the video game Plants vs Zombies: Battle for Neighborville that was accessed with a digital subscription to the online gaming platform Origin (Electronic Arts), c) Easy-SMX wireless 2.4g gaming controller with controller-charging cable, d) Microsoft Power Point, e) a cable modem with internet connection (10 Mbps or higher download speed required to support online gaming), f) personal hotspot accessible from a Sprint mobile device.
Dependent Variables and Data Collection
Target Behaviors
The participant was taught the following online-safety responses: a) abstain from providing personal information (i.e., address, name, or current location), b) say “no” to IRPI, c) pause/leave game within 30 s, and d) report the event to an adult within 30 s (e.g., “I need help”). Safety responses were scored on a four-point rating scale, similar to those employed in previous abduction prevention studies (e.g., Gunby et al., 2010; Johnson et al., 2005, 2006). For each observation, the participant was given a score of 0-4, with a point counted for each safety response given. When the participant provided personal information to the confederate a 0 was scored automatically.
Performance during BST (i.e., rehearsal phase) and IRPI probes were scored as follows: 0 = gave personal information; 1 = abstained from giving personal information but did not perform any other safety responses; 2 = abstained from providing information and performed only one additional safety response (i.e., said “no”, stopped the game, or reported to an adult within 30 s); 3 = abstained from providing information and performed two additional safety responses; 4 = performed all four safety responses (abstained from providing information, said “no”, stopped the game, and reported to an adult within 30 s).
Observer and Interobserver Agreement
The first author served as the principal observer of whether the participant reported the IRPI within 30 s, abstained from providing the requested personal information, said “no”, paused the game, and requested help from an adult. Another trained observer, located in the therapy room, scored safety responses performed by the participant for 33% of the trials during the study. The two scores were compared for each trial, and an agreement occurred when both scores matched exactly. Point-by-point interobserver agreement was 100%.
Treatment Integrity
Fidelity data were also collected. This included the completion of all steps in the BST procedure and was collected by the first author and a second trained observer, located in the therapy room, for 60% of trials using a six-question checklist (see Appendix). Point-by-point interobserver agreement was 100%.
Procedure
IRPI Probes
Online-safety responses were assessed using IRPI probes conducted before and after BST sessions (see Table 1). During IRPI probes, the subject was unaware of the observation and was not informed of the assessment. During IRPI probes, two PC computers, with an internet connection, were concurrently logged into the online the gaming platform Origin. IRPI probes were conducted in the multiplayer environment of the game Plants vs Zombies: Battle for Neighborville. A second trainer was also present online in the multiplayer environment during IRPI probes to act as a confederate. Confederates were novel adults situated in a separate physical location from the first trainer and the participant.
At the beginning of the assessment the first trainer instructed the participant that it was “game time” (i.e., 30 min free access was provided to play an online video game). The video game used in the study (i.e., Plants vs Zombies: Battle for Neighborville) allowed for both in-game voice-chat and textual messaging between players. All communications made between the participant and the confederate during IRPI probes were made vocally using the microphone of the PC computer.
Next, the second trainer, acting as a confederate, contacted the participant within the video game (in-situ) after 5 min. The second trainer maintained the conversation for another 2.5 min (i.e., 7.5 min into “game time”) at which point the second trainer initiated an IRPI (e.g., What is your real name?). Participant-safety responses were scored according to the four-point rating scale detailed above. If the participant complied with the IRPI (i.e., provided the confederate the requested information) during the probe, the second trainer made an excuse (e.g., “Hey, I am sorry but I need to leave”) and the assessment was terminated immediately to avoid potentially reinforcing the future fulfillment of IRPIs. If the participant independently left the game before the IRPI occurred, the trial was scored as a failed trial.
Baseline
The participant received three IRPI probes during baseline and his responses were observed. No feedback was given to the participant regarding his performance.
Behavioral Skills Training
During BST, safety responses to IRPIs were taught to the participant using verbal instruction, modeling, rehearsal, praise, and corrective feedback. These responses consisted of abstaining from providing the requested information, saying “no” in response to the IRPI, stopping the game, and reporting the incident to an adult within 30 s.
BST sessions occurred only once per week and were concluded after 30 min or when a safety score of 4 was earned during roleplay assessments. BST sessions were conducted until the performance criterion was reached. The performance criterion was a score of 4, without prompting, for two consecutive sessions on different days, similar to the criterion used by Bergstrom et al. (2014). During the first BST session five trials were conducted and the session was terminated after 30 min. The second BST training session also lasted 30 min and four trials were conducted. Finally, the last BST session was concluded following a single trial after the participant reached the performance criterion.
The training sessions started by briefly reviewing with the participant a Microsoft PowerPoint presentation. The presentations identified three different types of IRPIs, clarified to the participant that it is dangerous to provide personal information to strangers, and discussed appropriate safety responses to use with IRPIs presented online. Next, the participant was required to correctly state the safety responses to the first trainer before advancing to the modeling phase. The participant was prompted until he repeated all the safety responses correctly.
After the safety responses were stated, the training continued to the modeling phase. A second trainer, in a separate location, messaged the first trainer during gameplay, similar to IRPI probes. The first trainer modeled the correct safety responses to use with the IRPI. The online aliases (i.e., screen name) associated with the strangers during the modeling phase differed from those used by the confederate during IRPI probes.
Finally, the participant roleplayed the online safety responses during the rehearsal phase. The participant was told, “Hey let’s practice during game time”. IRPIs were made to the participant, with the trainer present, similar to IRPI probes. Verbal praise was provided for the successful completion of each step of the four-step response (i.e., abstaining from giving information, saying “no”, stopping game, and reporting to an adult within 30 s). If the participant received a score lower than 4, corrective feedback was provided for each missed step of the four-step response.
Posttest
Following the completion of the BST phase of the study, multiple IRPI probes were performed, the same as previously described during baseline.
In-situ Training
During the posttest assessment, in-situ training (IST) was included. During IST, the first trainer interrupted “game time” when the confederate presented an IRPI and the participant did not complete the four-step safety response sequence correctly. The first trainer provided the participant praise for safety responses that were performed correctly and corrective feedback for each missed step of the sequence, similar to the role play assessments conducted during BST.