Reminders, reflections, and relationships: insights from the design of a chatbot for college advising

Published date04 April 2023
Date04 April 2023
Subject MatterLibrary & information science,Librarianship/library management,Library & information services
AuthorHa Nguyen,John Lopez,Bruce Homer,Alisha Ali,June Ahn
Reminders, reections, and
relationships: insights from the
design of a chatbot for
college advising
Ha Nguyen
Instructional Technology and Learning Sciences,
Utah State University, Logan, Utah, USA
John Lopez
School of Education, University of California Irvine, Irvine, California, USA
Bruce Homer
Learning, Development, and Instruction, The Graduate Center,
City University of New York, New York, New York, USA
Alisha Ali
Department of Applied Psychology, New York University,
New York, New York, USA, and
June Ahn
School of Education, University of California Irvine, Irvine, California, USA
Purpose In the USA, 2240% of youth who have been accepted to college do not enroll. Researchers call this
phenomenon summer melt, which disproportionately affects students from disadvantaged backgrounds. A major
challenge is providing enough mentorship with the limited number of available college counselors. The purpose of
this study is to present a case study of a design and user study of a chatbot (Lilo), designed to provide college
advising interactions.
Design/methodology/approach This study adopted four primary data sources to capture aspects of user
experience: daily diary entries; in-depth, semi-structured interviews; user logs of interactions with the chatbot; and
daily user surveys. User study was conducted with nine participants who represent a range of college experiences.
Findings Participants illuminatedthe types of interactions designs that would be particularly impactful
for chatbots for college advising including setting reminders, brokering social connections and prompting
deeper introspectionthat build efcacy and identity toward college-going.
Originality/value As a growing body of human-computer interaction research delves into the design of
chatbots for different social interactions, this study illuminates key design needs for continued work in this
domain. The study explores the implications for a specic domain to improve college enrollment: providing
college advising to youth.
Keywords Chatbot, Interaction design, College, Mentoring, User study,
Humancomputer interaction
Paper type Research paper
This work was supported by the Institute of Education Sciences, award number R305H180051.
Received21 October 2022
Revised5 March 2023
6 March2023
Accepted7 March 2023
Informationand Learning
Vol.124 No. 3/4, 2023
pp. 128-146
© Emerald Publishing Limited
DOI 10.1108/ILS-10-2022-0116
The current issue and full text archive of this journal is available on Emerald Insight at:
Conversational chatbots have shown promisein promoting social good through interaction
designs that provide just-in-time information, connecting users to others with shared
experiences or guidingbehavior (Følstad et al.,2018). Examples in the health domain include
Babylon, which leads users through easyaccess to medical triage for breast cancer (Bibault
et al., 2019); Woebot, whichprovides mental health awareness and support (Fitzpatricket al.,
2017); and Vivibot, which delivers positive psychology interventions for young adults
recovering from cancer (Greer et al.,2019). Researchers have also turned to chatbots for
training and education to help students improve writing and listening skills (Kim, 2018),
scaffold the learning of content while watching online video lectures (Winkler et al.,2020),
intervene academically (Lim et al., 2021) and provide informational resources to students
who are transitioning to college or seeking academic advicein college (Daswani et al., 2020;
Ho et al.,2018;Page and Gehlbach, 2017).
A growing number of humancomputer interaction (HCI) studies have explored the
design implications for chatbots (or conversational agentsmore broadly) across a variety of
domains (Fitzpatricket al.,2017;Lee et al.,2020;Xiao et al.,2020;Xiao et al.,2020;Zhouet al.,
2020). However, a clear insight from the research literature is that designing chatbots to
support social interactions is a nontrivial task. For different domains, designers must
consider unique social dynamics in the development of a given chatbot (Bridge et al.,2021;
Winkler et al.,2020). In the following paper, we explore the implications for a specic
domain: providingcollege advising to youth who are deciding to enroll in highereducation.
We came to explore this design space through partnership with community nonprot
organizations in the Northeastern United States that helped low-income youth from ethnic-
minority communities to successfully enroll in college or university. Anywhere from 20 to
80% of youth in our partner high schools would decide not to enroll in college, even after
they gained successful admission. This pattern is called summer melt, because the dropout
happens in the summer months between high school graduation and subsequent fall
enrollment when there is a lack of available college counselors and mentorshipinteractions
(Castleman and Page, 2014;Xiao et al., 2020). To address this key bottleneck, our research
team explored the potential for designing chatbots to help advise students through this
This domain presentsa unique problem space. Youth who are decidingwhether to attend
college are reaching out to many people (e.g. peers and advisors) and seeking both
information and social-emotionalsupport to develop a sense of self as a college-going person
(e.g. an identity). Despite this multifaceted situation, chatbots for college advising have
generally relied solely on information-providing paradigms such as sending reminders for
key deadlines (Nurshatayeva et al., 2021;Page and Gehlbach, 2017). For students whose
college-bound identities are not concretely shaped yet and who may not necessarily know
what information to lookfor (Ober et al.,2020), combining information with added social and
emotional support might provide an opportunity for social connectedness and identity
formation (Fitzpatricket al.,2017). This conjecture motivated the following work.
In this case study, we designed, developed and piloted a chatbot Lilo to provide college
advising in the summer months, priorto enrolling in college. We present results from a four-
week, exploratory user study where we sought insight into how designed features of Lilo
were taken up by college-going youth, their interactions with Lilo and the ways in which
college advising chatbotscould better support the needs of youth in this domain. We present
how we leveraged HCI research to inform our design and the lessons learned in our user-
study around chatbots in college advising, a specic domain that has potential societal
benets to provide resources and advising at scale. We build rich descriptions of how our
Chatbot for

To continue reading

Request your trial

VLEX uses login cookies to provide you with a better browsing experience. If you click on 'Accept' or continue browsing this site we consider that you accept our cookie policy. ACCEPT