August 11, 2000 (Preliminary Report)

5.  Duration of the Experiment, Data Collection and Types of Feedback
The experiment was conducted for about three months between May and July, 2000. During this period, the subjects in the Control Group submitted seven letter-writing assignments (Tests 1 through 7) as part of the requirements of the correspondence course mentioned above. These letters were checked by two native-speaking staff members of the course under the supervision of this author and, after being added a full (i.e., non-selective) and explicit error correction in accordance with the usual procedure of the course, were sent back to respective subjects by surface mail. These letters were copied and filed for records before they were sent back.

The Experiment Group was also given the same letter-writing assignments as those given to the Control Group. The subjects in this group, however, were specifically instructed to make full use of the BLC Concordancer in writing the assignments except Test 1. They were also instructed to read an online introductory guide to the BLC Concordancer available on the Concordancer Website to get acquainted with the Concordancer before they start writing the assignments (See Appendix 4: A printout of the top page of BLC Concordancer Website). All the seven assignments were given via the Internet on the "Data-Driven Writing Project Homepage" (See Appendix 1: A printout of the DDW Project 2000 Homepage). The subjects were asked to submit their answers by email on a weekly basis, or one assignment per week, directly to this author, using the online JavaScript form mail provided on the DDW 2000 HP for each of the seven assignments.

All the assignments submitted by the subjects in the Experiment Group were checked by this author and returned by email after being added a selective feedback with no explicit error correction (as opposed to the full and explicit error correction given to the Control Group). In doing so, a particular attempt was made to encourage the subjects to refer to the BLC Concordancer in finding the right "solutions" to the problems in question, as shown in the following sample of instructor's feedback: