Mental rotations during vocabulary training
Robust Vocabulary Learning and Sentence Processing in French
Natasha Tokowicz and Tamar Degani
PI: Natasha Tokowicz
Others who have contributed 160 hours or more: Graduate Students: Tamar Degani
Study Start Date: 8/1/07
Study End Date: 6/30/07
LearnLab Site and Courses: Pitt/CMU French On Line and Classroom
Number of Students: Spring2006: 13 started the tutor, but only 7 completed the three days. Fall2006: 37 students received a password and 32 began the tutor. Of these, 27 completed all three days, 5 completed 2 days.
Total Participant Hours for the study: above plus additional lab hours (approximately 100).
Data in the Data Shop - after completion of analysis--data were not collected in Data Shop format.
This study focuses on testing a vocabulary training method with the goal of achieving robust learning (particularly in terms of longer-term retention). We teach students French vocabulary in association to pictures in unusual orientations (condition of interest) and in association to pictures in normal orientations, English words, and to a combination of both English words and pictures in unusual orientations (one presentation of each). In past research, learning vocabulary in association to pictures in unusual orientations was superior to learning vocabulary in association to pictures in usual orientations or to English words, when tested using recall of the new-language vocabulary. The probable benefit to learners is enhanced memory for the instructed vocabulary. Data collection has completed. Participants included students in French on-line and off-line courses at Pitt and CMU. We also collected data from a sample of Introductory Psychology students at Pitt; several returned for a 5-month retention test and those data are in the process of being analyzed.
Procedure: Students participated in three sessions approximately one week apart. Session 1 consisted of two cycles of training and a recognition test in which first an English word or a picture was presented, followed by a French word; students pushed a button to indicate whether the first item represented the meaning of the French word. Session 2 consisted of two cycles of training and a recall test in which an English word was presented and the students typed in the French translation. Session 3 was a test-only session in which students completed the recognition test followed by the recall test. At the end of Session 3, a language history questionnaire was completed.
3. Research question
How is robust learning affected by vocabulary instruction methods that discourage activation of dominant-language words, and encourage activation of the appropriate knowledge component (here, meaning)?
4. Dependent variables
We tested learning immediately following training on Sessions 1 and 2, and alone on Session 3.
The recognition test includeed accuracy as the primary measure of robust learning. We operationalized accuracy using d-prime, which is a measure of sensitivity that takes response bias into account because second language learners are often biased to respond "yes" on yes/no decision tasks. The recall test also included accuracy as the primary measure of robust learning, operationalized as percent accuracy.
For both tests, reaction times on correct trials form the secondary measure of learning. Because responses are typed, these times include response formation as well as typing, therefore they are a somewhat less informative measure than response accuracy.
5. Independent variables
Training condition is a within-participants factor with 4 levels: English word, picture in normal orientation; picture in unusual orientation; English word and picture in unusual orientation.
Second language vocabulary for picturable objects that are learned by associating second language words to pictures in unusual orientations (e.g., upside down) will be learned better than vocabulary learned by associating second language words to pictures in normal orientations (e.g., right side up) or to dominant language words.
Learning second language vocabulary in association to both pictures in unusual orientations and to dominant language words may assist learning more than the unusual orientation condition alone. In past research, the benefit for training in association to pictures in unusual orientations was limited to familiar pictures. Therefore, familiarization with the picture materials (accomplished via a word training trial) may help the learner take advantage of a subsequent training trial with the picture.
Associating new second language vocabulary to pictures in unusual orientations reduces the likelihood that the dominant-language word will become active. Because robust learning of second language vocabulary requires that a strong connection be formed between the new word and its meaning, reducing the activation of over-practiced associations between that meaning and the dominant-language word should ensure a stronger connection for the second language word. In other words, only the appropriate knowledge components will become active in the unusual orientation condition. Such learning should form a strong association that endures despite testing conditions--even when tested by showing a dominant-language word (considered to be one of the most difficult tests of second language vocabulary knowledge), words learned in association to pictures in unusual orientations should have an advantage.
Results: The preliminary results of 14 lab participants who used the tutor are as follows: A higher d’ for words learned in the word and unusual picture condition on the recognition test in session 1, suggest that words presented with both the English translation and a picture in an unusual orientation were learned better. There were no other significant differences in the recognition test on of session 1, or on session 2. Performance of participants, as measured by accuracy on the recall test, was the same in all 4 conditions.
In the future we intend to further analyze the data we have collected from the students in the online and regular French classes, to examine whether robust learning of the words is found with these learners. Furthermore, we intend to examine robust learning by looking at the recognition and recall accuracy approximately 3 months after learning, with of a subset of the participants who used our tutor.
9. Annotated bibliography