4
Results were similar for number of Error-Free phrases written (table 4), with the English majors showing greater gains, as indicated by pre and post-test effect sizes and the results of a t-test comparing gains (t = 3.96, df = 38, p = .0003, two-tails), but the Health Science students were more efficient.
Table 4: Number of error-free phrases written
| Pre-test | Post-test | gain | efficiency | |
| English Majors | 9.8 (10.4) | 28.2 (14.4) | 18.4 (12.4) | .15 (18.4/126) |
| Health Science Majors | 6.2 (5.1) | 13.0 (6.5) | 6.8 (5.6) | .38 (6.8/18) |
| Effect Size | 0.46 | 1.45 |
Table 5 presents improvements in accuracy as the ratio of error-free phrases to total phrases written. Again, the English majors made better gains, but were less efficient.
Table 5: Percentage of error-free phrases
| Pre-test | Post-test | gain | efficiency | |
| English Majors | 47% (9.8/20.9) | 55% (28.2/51.6) | 8% | .06 (8/126) |
| Health Science Majors | 35% (6.2/17.6) | 40% (13/32.9) | 5% | .28 (5/18) |
DISCUSSION
This study is not a pure comparison of comprehensible-input versus focus on form instruction; in the story-telling/self-selected reading condition, there was some focus on form in class, as vocabulary was emphasized during the story-telling and students were tested on vocabulary. Neither of the post-tests, however, specifically emphasized vocabulary knowledge. In addition, the claim that the Self-Access Pair Learning approach is heavily focus on form is based on Ross's analysis (Ross, 1992). There was no detailed analysis of the content of classroom instruction for the Self-Access Pair Learning group in this study. Nevertheless, it is highly likely that the instruction in the story-telling/self-selected reading class contained far more comprehensible input, and far less form-focused instruction than the Self-Access Pair Learning classes.
It is tempting to conclude from this data that comprehension-based instruction is much more efficient than traditional instruction, a finding that agrees with previous efficiency analyses (Mason, 2004; Mason and Krashen, 2004): Students who experienced one hour per week of comprehension-based instruction progressed at several times the rate as those who had one hour of comprehension-based instruction and six hours of traditional instruction.
Comparing one hour per week of comprehension-based instruction to several hours of similar instruction could, however, produce the same results. In other words, it may be the case that after a certain point, instruction of any kind is less efficient.
The differences in efficiency are so large, however, that it is doubtful that the diminishing returns seen here is due entirely to instructional fatigue. The extra form-focused instruction may not have been worth it.
4