This study was designed to address next matter: Just what are adolescent users’ responses in order to questions posed by a psychological fitness chatbot? Far more specifically, the intention of this study would be to check adolescents’ tastes (internet explorer, psychological valence and you will probability of responding) regarding the formulation out of questions that will be presented because of the a good psychological state chatbot. Considering earlier in the day search indicating that folks could possibly get like emojis and you can friendly styles for the psychological state chatbots, the questions made available to professionals differed considering the tone (friendly otherwise authoritative) as well as the presence out-of GIFs (establish or absent). Inquiries plus differed inside type of (yes/no, multiple impulse choices, or open-ended). These circumstances were selected created each other to your previous lookup towards mental fitness chatbots [ 24 , 36 ] plus the proven fact that he could be effortlessly malleable affairs that will increase associate experiences. We hypothesized one to adolescents do show a choice having inquiries including 
Recruitment
Since the the reason for this study would be to evaluate user tastes to possess mental health chatbot interaction certainly one of people kids, 19 kids aged 14 to help you 17 many years was basically recruited throughout the general people through leaflets and you can Fb ads. Participants were informed towards study aims and you will voluntary contribution, and each fellow member obtained settlement away from a whole worth of You $.
Construction and Processes
It inside the-research study try did having fun with a two?2?step three contained in this-victims factorial construction; the standards was exposure out-of GIF (introduce compared to missing), concern build (amicable versus authoritative), and you may matter type (open-concluded against sure/zero compared to multiple effect selection). Seven main questions was created ( Multimedia Appendix step one ), for every single addressing a different theme established around general better-getting, together with spirits, fret administration, and you will peer stress. For every single matter are modified based on different combinations of every basis, producing 12 distinctions for each and every of 8 fundamental statements and you will hence creating a total of 96 questions. The issue each and every question is actually managed along the different distinctions to deal with with the effect of motif on the users’ responses. When you compare two amounts of that fresh factor (such as, GIF expose against GIF missing), a similar concern was utilized for both criteria. All the questions and you can GIFs had been setup and you may pretested because of the four positives who had been experienced in chatbot creativity. Concurrently, a couple of teens was in fact requested to incorporate viewpoints to your advised questions in advance of evaluation, commenting on the readability and you can knowledge of the questions. Sixteen GIFs had been analyzed therefore the latest 7 (you to definitely per chief question) was in fact selected of the an expert committee. Test inquiries are offered for the Data step 1 and you can dos .
Immediately following professionals got see and you can finalized the new agree function, a report secretary told me the analysis rationale and you can offered participants brief spoken instructions. Participants was advised to imagine your issues made available to them had been posed by a good chatbot whose goal is to help you speak to users about their general really-being. Intricate information looked on the pc display at the start of the research. Participants were motivated to bring the some time and to inquire of concerns as needed to be certain it understood the job. All members had been including requested doing a go bullet before beginning the analysis. Study collection began once participants shown an obvious knowledge of the new activity. Each of the 96 inquiries try shown sequentially towards the a computer monitor to have an amount of 8 moments. Following the for each and every speech, people was indeed immediately redirected so you’re able to a short questionnaire displayed via Qualtrics (USA) and you may requested to point the likelihood of answering issue that they had merely see, in addition to their perceived affective reaction to issue. Your order of one’s chatbot issues is randomized for each and every participant. To end fellow member weakness, an initial dos-second video are starred after every band of thirty-two questions having a total of one or two films holiday breaks. At the conclusion of the research, i built-up demographic study courtesy various other on line survey exhibited via Qualtrics. Everyday debriefing are conducted after studies range, and participant views is solicited and you may listed. Studies range survived ranging from 60 and ninety minutes for every single new member. An illustration of the analysis process are found inside Shape step three . This research acquired integrity recognition from the Browse Ethics Board out-of HEC Montreal.