-
Time management is known to be a challenge, both for students in higher education,
as well of course as for a wide range of professionals
(cp. Claessens et al., 2007)
. By
time management, we here understand all activities and behaviours that surround
people’s allocation of time to their own activities
(in line with Claessens et al., 2007)
.
For students in higher education, it has been found that time management, such as
reporting short-term and long-term planning, and attitudes towards time as in Britton
& Tesser (1991), positively impacts academic performance, with especially clear
evidence for the positive impact of short-term planning
(Claessens et al., 2007)
.
Further positive impacts of time management are on psychosomatic wellbeing (ibid).
In this paper we develop a vision of intelligent mentoring technology based on the
principle of conversational reflection guidance. This vision is founded most narrowly
on the background of conversational agents for time management, as have been
studied with researchers and developers; and on the wider background of
computermediated reflection prompts and reflection guidance.
Background
Conversational Agents for Time Management
Conversational agents have been shown to be capable mediators of learning in
educational settings, and for different types of learning
(e.g., Adamson et al., 2014,
Graesser et al., 2001; Graesser et al., 2005; Ruan et al., 2019)
, but have only rarely
been used as partners in reflection; and typically not in repeated and long-term
interactions. There are scarce exceptions: With respect to being partners in reflection,
Kocielnik, Avrahami et al., 2018 have studied as proof of concept an conversational
agent for reflecting on time management, albeit with a focus on comparing written to
speech-based interaction, in a field study within a research lab. Wöls (2020) has
studied as proof of concept a conversational agent for reflecting on time management in a
field study with software developers, with a focus on exploring next steps of making
conversational agents effective in long-term, operative time management. Maybe not
surprisingly, integration with study participants’ (digital) work environment such as
existing calendars was identified as both technically challenging, and relevant in order
to increase usability and value of the conversational agent. With respect to long-term
interaction, in addition to the above two studies, Lee et al. (2019) have carried out a
two-week study with a conversational agent as support for learning self-compassion.
Long-term studies in the context of existing studies with conversational agents means
approximately two weeks. Throughout these studies, user acceptance of
conversational technology was high; and the intended positive effects (meaningful reflective
conversations about time management, learning about being self-compassionate) were
observed.
2.2
Computer-mediated Reflection Prompts and Conversational Reflection
Guidance
Technology for reflection can support both reflection-in-action, and
reflection-onaction; and prompts or other means of structure and support can facilitate reflection
within computational tools
(cp. Fessl et al., 2017)
. Researched computer tools for
reflection ranges from artistic interventions
(e.g., Lindley et al., 2009’s tool
ShoddyPop, discussed in the context of reflection by Baumer, 2015)
; to displaying relevant
data with no to little guidance as in literature on self-tracking
(e.g., Li et al., 2011)
; to
reflection prompts with variations in adaptivity and domain-independence
(e.g.,
Renner et al. 2016 - no in-built adaptivity, domain-independent; Fessl et al., 2017; McCall
et al., 1990 - adaptive, context-specific; Kocielnik, Avrahami et al., 2018; Kocielnik,
Xiao et al., 2018 - adaptive, domain-dependent)
. Substantial research in
computermediated reflection focuses on reflection prompts
(e.g., Shin et al., 2018; Renner et
al., 2016)
rather than on adaptivity or long-term interaction. Most narrowly relevant
for the present discussion, Pammer & Bratic (2013) and Pammer et al. (2015) have
studied activity log data with simple reflection prompts as basis for reflection; and
interpreted the overall positive results, that one next step in order to increase the value
of the reflection environment would be to go beyond data towards actionable insights
(Pammer & Bratic, 2013)
, and to explicate plans and commit towards a reflection
partner own plans for changing time management
(Pammer et al., 2015)
.
Consistently, throughout these studies, support for reflection has been effective –
although of course the mechanisms and approach to guiding reflection was extremely
different in all these studies.
Furthermore, more broadly speaking, reflection is an effective strategy to learn in a
self-directed manner from past experience (cp. Boud et al., 1989); albeit one that
requires sufficient knowledge about the domain one is reflecting
(Kirschner et al.,
2006)
. Especially in the case of time management, which is a horizontal activity for
students in comparison to the single subjects they study, such knowledge and
guidance would typically be given by mentors.
3
Discussion and Vision
In summary, we see that
• time management is relevant for students
• there is a plethora of tools for operative time management (e.g., calendars, task list
tools)
• there is extremely scarce work on conversational agents that support reflecting on
own time management – especially in relationship to the plethora of tools for
operative time management;
(in parallel by the way to the scarce evidence beyond
planning and goal-setting on concrete time management practices, cp. Claessens et
al. 2007)
• there is a broad body of literature both on reflection prompts, and a range of
complementary computer-mediated environments that structure and guide reflection
We therefore see that students have the tools available for time management, but not
necessarily the knowledge to reflect on their own about their time management –
and hence to fill the tools with reasonable plans; hence mentoring in terms of time
management is expected to be interesting to students. Given that knowledge
workers in principle are known to also be challenged by time management
(Claessens et
al., 2007)
, even if the concrete time management challenges are maybe different,
we can expect that such mentoring would also be interesting to a broader range of
target users.
In own past work
(in particular Pammer et al., 2015)
, we have seen that while
reflection in a self-directed manner worked, self-awareness of plans for change, and
experimentation with different time management setting changed in a study with a
coach, but didn’t change in a study with only computational support for reflection.
This would of course simply argue for establishing a broad mentoring structure at
university. This doesn’t scale. As a step of intermediate quality (effective, maybe
worse than a good human mentor, but better than no mentoring at all, and hence
not only effective but also efficient), there would be intelligent mentoring
technology. Such technology would, beyond what is currently explored in existing
literature and commercial products, need to
• Have knowledge both on how to lead reflective conversations, and on time
management, and at the same time develop an understanding of each user in
terms of strengths and weaknesses with respect to both reflection and time
management. So far, I argue that intelligent mentoring technology is
somehow similar to intelligent tutoring systems – however, the domain we are
talking about here (time management) is substantially less structured and less
evidence-based, than typical domains of intelligent mentoring systems –
hence, open-ended reflective dialogues as means for support.
• Be conversational in order to act as a reflection partner towards whom to
make explicit own reasoning in reflection; and towards whom to commit
with respect to plans. It is of course uncertain, whether commitment to a
software agent will be as high as towards a human reflection partner, but as
study participants were found to develop something akin to attachment to an
agent in two weeks
(Lee et al., 2019)
, it seems reasonable to expect that
commitment will still be achieved.
• Integrate with the existing digital environment for time management of
users, in order to be able to refer to operative time management in reflective
conversations.
Finally, beyond short-term planning as seen as positively impacting academic
performance in time management literature (Claessens et al., 2017), goal achievement
literature in addition points to the positive effects of setting specific, ambitious
goals that are under the control of the user
(Locke & Latham, 2002)
. However,
people are in general reluctant to refine higher-level goals, and typically tend to
start by refining lower-level goals
(Cropanzano et al., 1995)
. In parallel, there are
indications about the positive impact of reflective goal-setting (Travers et al.,
2015). Hence, I argue that intelligent mentoring technology for students’ time
management should not only consider operative time management, in the sense of
putting up daily, weekly, or semester-plans up for reflection; but should also
consider higher-level goals as objects of reflection.