11.07.2015 Views

Abstracts - Earli

Abstracts - Earli

Abstracts - Earli

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

experiment. First-year students in a Chinese language program at an American university weretrained to identify the tones of 435 learned characters and 69 novel characters across fourteenlessons in the first and second semester. Three learning conditions were designed to support tonelearning by presenting: (1) visual pitch contours of the tones, together with Pinyin spelling of thespoken syllable (contour + pinyin condition); (2) digits that traditionally represent the tones,together with Pinyin spelling of the spoken syllable (digit + pinyin condition); (3) visual pitchcontours without Pinyin spelling (contour only condition). Analyses of student activity logs(learning curves) showed a significant effect of learning condition. The contour + pinyin and digit+ pinyin conditions produced significantly faster learning. Furthermore, improvement from a pretestto post-test was largest for contour + pinyin condition. These findings support the value ofredundant multi-modal information sources for supporting the learning of complex linguisticforms that include both segmental and tonal features. Differences between novel and familiarforms further suggest a two-level learning process, one that extracts tone as a general feature toapply to any perceived syllable and one that must acquire each syllable as a unique segment+tonerepresentation.The role of the self in self-explanationKurt VanLehn, University of Pittsburgh, USARobert Hausmann, University of Pittsburgh, USAScotty Craig, University of Pittsburgh, USADoes deep learning depend on an individual generating an instructional explanation or will similarlearning result from processing equivalent information, while controlling for "active" learning? Toaddress this question, two in vivo classroom studies were conducted in the PSLC physicsLearnLab. The first study contrasted the generative learning activity of self-explaining with thenon-generative, learning activity of paraphrasing instructional explanations. During a classroomlaboratory period on electrodynamics, students were asked to talk aloud as they alternated betweensolving problems and studying examples. Thus, the data for this study came from problems solvedduring the training session as well as assigned homework problems. The results suggest thatstudents who self-explained had higher scores on the training problems than students who studiedinstructional explanations. Moreover, self-explainers displayed evidence of accelerated futurelearning of a related, yet new domain (i.e., magnetism). The second in vivo study contrasted thelevel of guidance giving during initial learning of material. Students were presented informationon magnetism with deep-level, rhetorical questions for full guidance, pauses for self-explanationwhere learners build their own guidance, or pauses for reflection where learners were not givenencouragement to self-explain. The data from this study also came from LearnLab homeworkproblems. The results suggest that, while there were no differences between the three conditions interms of their homework performance, those who were provided guidance via deep-level questionsexhibited faster solution times than those who were given no explicit guidance. The results fromthese studies suggest two conclusions. First, prompting students to self-explain or answer deepquestions while studying problems in an authentic classroom environment can result in robustlearning. Second, it is important to motivate students, not to engage merely in active learningprocesses, but to also be generative.– 798 –

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!