Upload
arthur-barber
View
214
Download
1
Embed Size (px)
Citation preview
Debating the Issue of Tutoring
Interactivity: Intuition vs. Experimentation
Tanner JacksonIt’s a MAD MAD MAD MAD Morning
2
Effective Tutoring
• Chi, M. T. H., Siler, S., Jeong, H., Yamauchi, T., & Hausmann, R. G. (2001). Learning from human tutoring. Cognitive Science, 25, 471-533.
– Three possible explanations:
• Tutor’s pedagogical skills
• Student active construction
• Joint effort of the Tutor and Student
3
Effective Tutoring
• Chi’s components of effective tutoring:– If effectiveness is due to tutor’s pedagogical skills
• Dialog dominated by tutor• Frequent/common use of strategies
– Student active construction• Dialog dominated by student• Frequent/common use of self-explanation (construction)
– Joint effort of the two• Distribution of turn-taking in the dialog• Frequent/common use of elicited construction and
scaffolding
4
Interactivity
• Interactivity could include any joint effort (Clark, 1996; Chi et
al, 2001):– Everyday conversation– Mixed initiative dialogue– More than one person taking action– Getting feedback from an outside source– Acknowledging feedback (head nods, gestures, eye gaze)
• Anything that involves some sort of internal AND external participation/communication.– This includes: people, computers programs, cars, ATMs, etc.– This excludes: monologues, lectures, reading, watching
(without doing or acknowledging), etc.
5
Interactivity Hypothesis
1. Communicative claim: Both tutors and students are maximally interactive.• Tutors’ turns are largely communicative, which
elicit content from the students• Students’ turns are responsive to the tutor
2. Learning claim: Interactive construction by the students (elicited by the tutor) should enable more learning than non-interactive construction (self-initiated).• If this second claim is true, then this would help to
pinpoint the advantage of tutoring above and beyond other learning methods
6
Tutoring Interactivity
• For our own purposes, “Tutoring Interactivity” includes:– Dialogue (of any sort)– Turn taking (not necessarily mixed initiative)– At least 2 entities (tutor and tutee)
• Theoretical claim for Tutoring Interactivity– More interactivity => More learning– Students learn better/more/deeper if they learn
through interactive means.
7
Tutoring Interactivity - Pseudo-theory #1
Interactivity allows for adaptive responses
Adaptive responses tailor the tutoring to each student’s appropriate level
Tutoring to each student’s appropriate level should be the most effective method of tutoring
8
Tutoring Interactivity - Pseudo-theory #2
Interactivity allows for joint knowledge construction
Joint knowledge construction allows for all parties to contribute their pieces of knowledge
If everyone contributes their piece of knowledge, then the group can understand
something that the individuals did not initially know.
9
Tutoring Interactivity - Pseudo-theory #3
Interactivity allows for prompting
Prompting allows for construction
If students construct knowledge on their own they will understand it more deeply and retain it
10
Tutoring Interactivity – Pseudo-theory #4
Interactivity is good
Interactivity alone accounts for some benefits of tutoring
Benefits of tutoring are good
11
Tutoring Interactivity
Intermission for:CommentsQuestionsConcerns
12
Experimental Support
• Hypothesis: More interaction causes more learning
• Many experiments support– Tutoring > textbook (Graesser; Lane; Why2 expt. 2)
– Tutoring > nothing (Graesser; Merrill)
– Contingent tutoring > lecturing (Wood; Swanson)
– Canned text remediation > nothing (Katz)
– Tutoring > Canned text remediation (Why2 experiments 4 & 5Lo)
13
Experimental Conflicts
• Also many experiments do not support:– Socratic tutoring = didactic tutoring (Rosé 2001)
– Scaffolding = lecturing or text (Chi; Rosé 2003)
– Tutoring = multiple choice & feedback (Aleven; Reif)
– Tutoring = Canned Text Remediation (Katz; Why2 experiments 1, 3 & 5)
14
Experimental Conflicts
• VanLehn, Graesser, Jackson, Olney (submitted) Why2: Submitted to Cognitive Science Journal & Conference 2005.
– Three possible exceptions (when tutoring interactivity is not better than non-interactive learning methods):
• Sufficient prior knowledge
• Text/Monologue content = tutoring content
• Motivated to self-explain text/monologue content
15
More on Interactivity
• Some vs. No interaction between steps– Wood, Wood & Middleton (1978)
• Assembling a complex block structure• Face-to-face contingent tutoring > demonstrating
– Swanson (1992)• Understanding how lens affect images• Face-to-face contingent tutoring > lecturing
– Merrill, Reiser, Merrill & Landes (1995)• Lisp programming• Tutoring > just flagging incorrect Lisp code
– Coleman (1998)• Photosynthesis• Conversational prompts > no prompts
– Lane & VanLehn (in press)• Pseudo-code design• Tutoring > reading
16
More on Interactivity
• High vs. Low interaction between steps– Rosé, Moore, VanLehn & Albritton (2001)
• Electricity problem solving• Socratic tutoring = didactic tutoring
– Aleven, Koedinger & Popescu (???)• Geometry problem solving with justifications• Dialogue elicitation of justification = menu selections
– Reif & Scott (???)• Face-to-face human physics tutors = a CAI tutor
– Chi et al. (2001)• Deep, incremental reading of a text on blood circulation• Scaffolding = lecturing after each sentence
– Rosé, Bhembe, Siler & Srivastav (2003)• Qualitative physics• Tutoring = reading
17
Tutoring Interactivity
IntellectualPlaytime