learning analytics serious games cognitive disabilities
TRANSCRIPT
GLAID: Designing a Game Learning Analytics Model to Analyze the Learning Process in Users with Cognitive
DisabilitiesBaltasar Fernández-Manjón
Ana R. Cano, Álvaro J. García-Tejedor
Grupo e-UCM: [email protected] @BaltaFM
SGames Conference, Porto, 16/06/2016http://www.slideshare.net/BaltasarFernandezManjon/
LA & GLA 101• Learning Analytics: Improving education based on Data Analysis
7 Data driven7 Evidence-Based Education
• Game Learning Analytics application of LA to Serious Games7 Interaction data in a Serious Game is collected and analyzed for improving the
learning process supported by the game7 Educational game not as “black boxes”
7But LA & GLA is not “informagic”7 We need to relate data with what happens in the game and with the
educational design!
The GLA Problem• Ok, we are collecting ALL the interaction data in a video game but…
IT IS A HUGE AMOUNT OF DATA!
Now what?• What are the relevant observables?• How do I analyze the data collected?• How do I translate it into useful
information about the learning process?
And the problem gets bigger……If the user has an intelectual condition or disability (e.g. Down Syndrome)
User Features:• Interaction with the game (motor skills)• Ordering thoughts and language in a “logical” layout• Listening and taking turns in conversations• Communication in an interactive sense• Relating objects and actions to spoken or written words
H2020 Beaconing project• BEACONING stands for ‘Breaking Educational Barriers with Contextualised,
Pervasive and Gameful Learning’ • Started in january 2016, 15 partners, 9 countries, 6M
• Global goal is learning ‘anytime anywhere’• Exploitation of technologies for contextual pervasive games and use of gamification techniques• Problem based approach to learning• Enriching the Gaming Learning Analytics data model with
the contextual, geolocalized and accessibility information
• Large pilots in real settings with content providers• Formal and informal learning across virtual and physical spaces
• GLA is a key element in the games and pilots evaluation• Using RAGE infrastructure and extending it for these
new requirements and applications
Our approach: The GLAID Model
Present
Individualized Learning Analysis
Collective Learning Analysis
Predictive Learning Analysis
….
Group 1
Group 2
Group 3
Game Sessions
Lear
ning
Pro
gres
s
d1.a d1.n
d2.a
d3.a
d2.n
d3.n
*d = Data collected during a game session
GLAID (Game Learning Analytics for Intellectual Disabilities) Model
Analytics Framework
User 1
User 2
User n
User 1User n
User 3 User 2
User 5User 4
User 1
Data Handling
Designer Perspective Educator Perspective
User cognitive restrictions
Formal Requirements
Game & Learning Design
Group of Observables
Group of Observables
Descriptive Analytics
Clustering Analytics
Predictive/PrescriptiveAnalytics
First Step: From the User Restrictions to a Game Design • Challenges: 1) Transform the user characteristics into
formal requirements 2) Develop a learning game design
adequate for users with intelectual disabilities (such as Down Syndrome, mild cognitive impairments, ASD Autism Spectrum Disorders,…)
3) Select a group of observables/variables that measure the learning outcome of the user for future assessment Present
Individualized Learning Analysis
Collective Learning Analysis
Predictive Learning Analysis
….
Group 1
Group 2
Group 3
Game Sessions
Lear
ning
Pro
gres
s
d1.a d1.n
d2.a
d3.a
d2.n
d3.n
*d = Data collected during a game session
GLAID (Game Learning Analytics for Intellectual Disabilities) Model
Analytics Framework
User 1
User 2
User n
User 1User n
User 3 User 2
User 5User 4
User 1
Data Handling
Designer Perspective Educator Perspective
User cognitive restrictions
Formal Requirements
Game & Learning Design
Group of Observables
Group of Observables
Descriptive Analytics
Clustering Analytics
Predictive/PrescriptiveAnalytics
1st Level Analysis: Individualized Learning Analysis• Goal: Describe and analyze historical
learning data from the student’s perspective• Outcome: Gives an overview of the user’s
learning behaviour through several game sessions• Observables collected individually
• Timestamps• Level changes• Achievements vs. Fails• User interactions (number of clicks, heatmaps,
time between clicks,…)
Individualized Learning Analysis
….
d1.a d1.n
d2.a
d3.a
d2.n
d3.n
*d = Data collected during a game session
User 1
User 2
User n
9
2nd Level Analysis: Collective Learning Analysis • Goal: Identify causes of trends and learning
outcomes for a group of users segmented by disability or cognitive skills• Outcome: Learning patterns• Observables collected collectively• Timestamps• Level changes• Achievements vs. Fails• User interactions (number of clicks, heatmaps,
time between clicks,…)
Collective Learning Analysis
Group 1
Group 2
Group 3User 1User n
User 3 User 2
User 5User 4
3rd Level Analysis: Predictive Learning Analysis• Goal: Analyze current and historical data to make
predictions about future learning outcomes• Outcome: assure the effectiveness of a game as a
learning tool for a user with an specific disability• Observables colected individually and
collectively• Timestamps• Level changes• Achievements vs. Fails• User interactions (number of clicks, heatmaps, time
between clicks,…)
Predictive Learning Analysis
Game Sessions
Lear
ning
Pro
gres
s
User 1
Data Handling: stakeholders• 2 Data handling perspectives:
Game Designer’s Perspective• Collect and analyze all the states that
the user can reach in a game session• Are the mechanics of the game
appropiate for the user?
Educator’s Perspective• Learning experience of each user• Are the users learning or struggling
with the game?
12
Collecting data with xAPI
• We can collect the relevant data in a standard format using xAPI• We are working in a xAPI serious games profile with ADL• This will simplify the analysis and visualization of data (e.g. dashboards)
xAPI
Case study: Downtown• Serious Game designed and develop
to teach young people with Down Syndrome to move around the city using the subway• Status: Designed and developed.
Analysis pending
• Type of game: Serious Game• Audience: People between 15 and 30 y/o with
Down syndrome• Platform: PC and Android (work in progress)
14
Case Study: Downtown• From user requirements to a game
design and its observables• Standards: W3C cognitive
accessibility, accessibility guidelines
Case Study: From user requirements to a game designUser
Requirement Game Restrictions Game Design & Mechanics Observable
Limited intellectual autonomy
The game should be able to guide the user during the learning session through interactive help, pop-up tips or other mechanics
There will be a "help" button permanently in the screen where the user can ask for help at anytime during the game session
Clicks in the Help buttons during a game session
If the user doesn't perform any interaction for more than 2 minutes, a pop-up aid will appear providing guide, tips and advices
Total inactivity time
Inactivity time after pop-up help appears
The phone will act as a help button. If the user needs tips or advices, he can call the police asking for clues to complete the ongoing task
Case Study: From user requirements to a game designUser Requirement Game Restrictions Game Design Observable
Difficulty in the process of abstractions, conceptualization, generalization and learning transfer
The game should explain any action to do, even the easiest, without assuming that the user already know how to complete it
Tutorials: The description about how to achieve the goals in the game will be performed as a video explanation before the task starts
Time consumed in completing the task
Previous research prove that visual explanations help to understand the assignments better than hearing or reading.
Savidis, Grammenos and Stephanidis "Developing inclusive e-learning and e-entertainment“. 2007
ExampleCase Study: Applying GLAID to the game• Observable: Clicks in the “Help” button during a game session
Session #1 Session #nUser #1 3 clicks 1 clickGroup of users #1 5 clicks (avg) 4 clicks (avg)
GLAIDIndividualized Learning Analysis Collective Learning Analysis
•Game designer’s persp: The user improved in the use of the game through sessions•Educator’s persp: The learning experience of
the user seems to improve through sessions (measure with other observables)
•Game designer’s persp: Users with XX disability slightly improved in the use of the game through sessions. May reconsider game design and mechanics for certain tasks•Educator’s persp: The learning experience of
the user slightly improved through sessions. May reconsider the learning experience
User#1 Assessment:•The user is able to use the game as a learning tool better than other users•His intellectual autonomy seems to be above the average for his type of disability•His learning experience seems to be improving through game sessions
18
Just another BEACONING initiative …
Thanks!Questions?Mail: [email protected]: @BaltaFM GScholar: https://scholar.google.es/citations?user=eNJxjcwAAAAJ&hl=en&oi=ao ResearchGate: www.researchgate.net/profile/Baltasar_Fernandez-ManjonSlideshare: http://www.slideshare.net/BaltasarFernandezManjon