Difference between revisions of "McLaren - The Assistance Dilemma And Discovery Learning"
(→Background and Significance)
|Line 30:||Line 30:|
===Background and Significance===
===Background and Significance===
goal is to the to --and to to of the the .
the a , a the -.
to better on the .,.
Revision as of 19:13, 20 November 2009
- 1 The Assistance Dilemma and Discovery Learning
The Assistance Dilemma and Discovery Learning
Bruce M. McLaren
PI: Bruce M. McLaren, Carnegie Mellon University, Pittsburgh
Others who have contributed 160 hours or more:
- Alex Borek, University of Karlsruhe, Germany, research, programming, statistical analysis
- Dave Yaron, Carnegie Mellon University, Chemistry domain expertise, Support of classroom study
- Mike Karabinos, Carnegie Mellon University, Chemistry domain expertise, Support of classroom study
How much help helps in discovery learning? This question is one instance of the assistance dilemma, an important issue in the learning sciences and educational technology research. To explore this question, we conducted a study involving 87 college students solving problems in a virtual chemistry laboratory (VLab), testing three points along an assistance continuum: (1) a minimal assistance, inquiry-learning approach, in which students used the VLab with no hints and minimal feedback; (2) a mid-level assistance, tutored approach, in which students received intelligent tutoring hints and feedback while using the VLab (i.e., help given on request and feedback on incorrect steps); and (3) a high assistance, direct-instruction approach, in which students were coaxed to follow a specific set of steps in the VLab. Although there was no difference in learning results between conditions on near transfer posttest questions, students in the tutored condition did significantly better on conceptual posttest questions than students in the other two conditions. Furthermore, the more advanced students in the tutored condition, those who performed better on a pretest, did significantly better on the conceptual posttest than their counterparts in the other two conditions. Thus, it appears that students in the tutored condition had just the right amount of assistance, and that the better students in that condition used their superior metacognitive skills and/or motivation to decide when to use the available assistance to their best advantage.
How much help helps in discovery learning?
Background and Significance
A key goal of educational technology research is to find the right level of support to imbue in computer-based educational systems. The so-called assistance dilemma is central to this goal: “How should learning environments balance assistance giving and withholding to achieve optimal student learning?” (Koedinger & Aleven, 2007). Assistance giving allows students to move forward when they are struggling and truly need help, yet can rob them of the motivation to learn on their own. On the other hand, assistance withholding encourages students to think and learn for themselves, yet can cause frustration when they are unsure of what to do next.
Although the “assistance dilemma” is a relatively new term, it describes a central issue in the learning sciences that has been debated for some time. The extreme position of assistance giving is usually called direct-instruction or guided learning. Supporters of this position (e.g. [2,3,4]) argue that higher assistance (direct instruction and/or tutoring of basic skills) leads to better learning results because it provides information that students cannot create on their own. Supporters of the opposing position (e.g. [5,6,7,8]) advocate a much lower assistance approach (i.e.,assistance withholding), often called discovery or inquiry learning.
The study compared three conditions in which students used different versions of the VLab to solve problems in thermo chemistry:
- (Condition 1) the Inquiry-learning Condition, in which students worked with a version of VLab with no hints and minimal feedback,
- (Condition 2) the Tutored Condition, in which students could request hints and received feedback only when they were severely off track, and
- (Condition 3) the Direct-instruction Condition, in which students were directed to follow a prescribed problem-solving path.
Our plan is to include the following robust learning dependent variables in our studies.
- Normal post-test: Students will take an immediate post-test, right after completing work with the stoichiometry tutor
- Transfer: Conceptual, transfer questions will be included in the post-tests
- Long-term retention: Students will take a second post-test, including conceptual, transfer questions, 7 days after the initial post-test
As mentioned above, a lab study with over 100 subjects was run in early 2009 at the University of California with the above conditions. College students learned to solve chemistry stoichiometry problems with the stoichiometry tutor through hints and feedback, either polite or direct, as described above. There was a pattern in which students with low prior knowledge of chemistry performed better on subsequent problem-solving tests if they learned from the polite tutor rather than the direct tutor (d = .73 on an immediate test, d = .46 on a delayed test), whereas students with high prior knowledge showed the reverse trend (d = -.49 for an immediate test; d = -.13 for a delayed test). On the other hand, the high school study, also run in early 2009 with over 100 subjects, produced different results. In particular, the high school students did not show a pattern in which students with low prior knowledge of chemistry performed better on subsequent tests. We are still analyzing the audio feature of the study, i.e., the comparison of audio to text hints and messages, but preliminary results indicate that adding audio hurt the performance of high knowledge learners and helped low knowledge learners on the delayed test.
This study is part of the Computational Modeling and Data Mining thrust.
Our explanation for the specific findings from our experiment are soon forthcoming. We are currently preparing a paper for the journal of educational psychology that will provide such an explanation.
Connections to Other PSLC Studies
- This study has a clear connection to the McLaren et al study , in that both studies explore the effect of personalized, polite hints and feedback. In fact, it was through McLaren's original studies, built on earlier work on e-Learning principles by Mayer, that Mayer and McLaren decided to join forces.
- McLaren, B.M., DeLeeuw, K.E., & Mayer, R.E. (submitted). A Politeness Effect in Learning with Web-Based Intelligent Tutors. Submitted to the Journal of Human Computer Studies.
- Brown, P., & Levinson, S. C. (1987). Politeness: Some universals in language usage. New York: Cambridge University Press.
- Mayer, R. E. (2005). Principles of multimedia learning based on social cues: Personalization, voice, and image principles. In R. E. Mayer (Ed.), The Cambridge handbook of multimedia learning (pp. 201-212). New York: Cambridge University Press.
- McLaren, B. M., Lim, S., Yaron, D., and Koedinger, K. R. (2007). Can a Polite Intelligent Tutoring System Lead to Improved Learning Outside of the Lab? In the Proceedings of the 13th International Conference on Artificial Intelligence in Education (AIED-07), pp 331-338. [pdf file]
- Nass, C., & Brave, S. (2005). Wired for speech: How voice activates and advances the human-computer relationship. Cambridge, MA: MIT Press.
- Reeves, B., and Nass, C. (1996). The media equation. New York: Cambridge University Press.
- Wang, N., Johnson, W. L., Mayer, R. E., Rizzo, P., Shaw, E., & Collins, H. (2008). The politeness effect: Pedagogical agents and learning outcomes. International Journal of Human-Computer Studies, 66, 98-112.