Handwritten Text Recognition for manuscripts and early printed texts
StoryMap: Using Social Modeling and Self-Modeling to Support Physical Activity Among Families of Low-SES Backgrounds
1. 2022.03.04
StoryMap: Using Social Modeling and Self-Modeling to Support
Physical Activity Among Families of Low-SES Backgrounds
Hermann Saksono et al.
CVPR ’21
Proceedings of the 2021 CHI Conference on Human Factors in Computing Systems
3. 3
Introduction
• Obesity burdens households of low-socioeconomic status (SES)
• Physical activity (PA) behavior is linked to social factors (e.g., family
support)
• Two modes of modeling in Social Cognitive Theory (SCT)
• Self-modeling: reflecting on one’s best performance in the past
• Social modeling: observing similar people’s performances
• Health behavior induced by two attitudinal factors
• Self-efficacy: Beliefs about our ability in completing a task
• Outcome expectations: Beliefs in what happens if we do the task
How to facilitate self-modeling and social modeling?
How to positively impact PA attitudes?
4. 4
Related Work
Social Cognitive Theory (SCT)
• Human behavior
• Human cognition + humans’ interactions with environment
Self-modeling
• Remembering memories of the past help to gain insights about PA
• E.g., Echo, MAHI, Pensieve, and Visualized Self
Social modeling
• Learning other’s behavior help to pick reasonable goals, gain a
feeling of social support and get motivated
Physical Activity Technologies
• Metaphor visualizations
• Goal-setting tools
• Reflection support
• Social comparison
5. 5
Method
StoryMap Prototype Design
• Extend Storywell (open-sourced Android App) by adding the
community story map
• Caregivers record and share PA stories
User Scenario
6. 6
User Scenario
Read a PA-themed storybook
Complete one-day step count goal to read the next
chapter
Record answers to a set of reflection
questions
Share a community story
Log users’ emotions to use for
the interview
7. 7
User Scenario
See other families’ community stories on the story map
Tap to listen to other stories
Spin the magic needle to
randomly get one of the exit
balloons to unlock the story
When miss their goal,
8. 8
Design Rationale
• Develop 8 storybooks through workshops with health an HCI researchers
• A community organization of children’s literacy reviews and provides
feedback
Self-modeling
Example of three-part reflection questions
• Can you tell a time when you tried a new physical activity that was hard at
first?
• How did that activity become easier when you tried it again?
• Being active does not happen overnight. Starting with small activities and
repeating them often can make physical activity easier to do.
Reflection questions Community prompts
9. 9
Design Rationale
Community prompts
• Similar but in merged manner
• Can you describe a time when you enjoyed being active with your
family?
• What are some nice things that happened that you would like
others to know about?
Social modeling
• Community story map
• Listen to other families’ success stories
• Stories on a neighborhood map develop a sense of
connectedness
10. 10
Design Rationale
Social modeling
• More impactful when the model is similar to users
• Display metadata about the caregiver to examine how to support
social modeling by similarity
• Proxy indicators include:
• Step count
• Approximate age
• Number of children
• Neighborhood information
11. 11
Data Collection and Analysis
Data Collection
1) Fill out online survey about the caregiver’s 7-day PA and
perception of their children’s PA
2) Wear the bands daily and check StoryMap at least once a week
3) Choose fitness challenges in one week
4) Have 3 interviews
• Data-Driven Retrospective Interviewing (DDRI) method
• Use combination of emotion and interaction log
• Experiences in using the reflection, story sharing and story
map
1) Open coding
2) Use initial concepts
3) Group the concepts
4) Link categories and sub-categories
Analysis
12. 12
Findings
• 16 caregivers and 16 children participated
• PA experiences evoke exercise memory from the past
• Caregiver relive the emotions felt in the experience
• Enhancing ability perceptions (n=3)
• Inducing the desire to repeat the activities (n=8)
• Intents induced by reflections can be impermanent (n=3)
Reflection as self-modeling
Reflections identify abilities and emotional outcomes
“It depends on what time you read the book [in the app]. Then they ask that question, you feel that way,
yes. That is so true. That after you say, “Okay you’re going to go do this.” Then the emotion and your
feelings then change, that you didn’t even complete the task.” (P8)
13. 13
Findings
• Feeling of uneasiness toward future actions (n=3)
• Discrepancy between aspirations and beliefs about their
ability
• Not experience discomfort if the caregiver is self-efficacious
• Identifying discrepancies leads to engage in discrepancy
reduction
Reflections identify discrepancies
14. 14
Findings
• Listening to other stories can enhance caregivers’ PA beliefs
(n=10)
• Listening to verbal stories are more impactful
• Families exchange several kinds of information
• Task information – courses of action (n=9)
• Emotion information (n=3)
• Normal behavioral information –learning that internal standards
align with the social standards (n=7)
• Adequacy information – caregivers’ step count data(n=6)
Storytelling as Social Modeling
Data and stories affect health beliefs
Just hearing the stories, you were just like, “Oh, okay. So, I’m not the only one.” [. . .] When you sit back and
you’re like, “Okay, I’m not the only one doing this. I’m not the only one trying to find ways to keep my kids
busy and happy.” It’s good. It works out. (P12)
15. 15
Findings
• Metadata helps caregivers predict how much they can replicate
the models’ behavior (n=8)
• Number of children (n=10)
• Greater number of children usually inhibited PA
• Neighborhood (n=6)
• Inhibit PA with lack of places to be active or safe
• Neighborhood information can predict PA by signaling the
availability of resources
Metadata enhances data and stories
16. 16
Findings
• Limit families’ use of StoryMap and ability to be active (n=12)
• Access to the backyard
• Straining caregiving
• Work responsibilities
Barriers
Underlying barriers
Temporary barriers
• Social distancing during COVID-19 pandemic
17. 17
Discussion
Social Modeling Using Data and Stories
• Share data and stories
• Enhance individuals’ beliefs about their ability
• Reinforce beliefs about the joy of performing the behavior
• Comparisons are helpful to give a sense of adequacy
• Being similar includes facing similar obstacle
• Help users to predict whether they can replicate the model’s
behavior 1
2
1
2
3
3
18. 18
Discussion
Reflection as Self-Modeling
• Remember positive experience to experience positive emotion
• Reinforce the outcome expectations of the task
• Reengage cognitively on the courses of actions to complete the task
• Intention can be transient
• Prompting reflections when it is most feasible to do the activity
• Providing goal setting tools after the reflection
1
2
1
2
3
3
Solve
19. 19
Conclusion
• Invite reflections on success stories and invite users to share key
information
• Support the exchange of stories and behavioral data aimed at
enhancing self-efficacy and outcome expectations
• Support users in gaining a sense of similarity with the model by
allowing the exchange of information that helps users determine
whether they share sufficient characteristics with the model
• Invite reflections that lead to action by asking users to reflect on
when it is most feasible to engage in behavior change
• Amplify community voices using stories