Broad comparisons between methods of teaching in all academic areas are criticized due to the fact that so many of the results indicate no significant
differences. Surveys have confirmed that few studies find significant variations among groups. There are at least two plausible explanations for this state of affairs. First, experimental design is not sophisticated enough at present to control all the variables involved. Until educational research reaches a higher level of development, conservative interpretations of trends may have to suffice as indicators of superiority or inferiority of different teaching methods. Additional support for favoring differences can be gained by replicating the original study. Second, in some problem areas, such as sense modality preference, no significant differences should be expected. If some students prefer to learn with their ears and some prefer to learn with their eyes, a random sample should distribute these two preference groups equally into the experimental and control classes. Other factors being equal, the effects of this preference for learning equally distributed in both groups should counterbalance the effects of the other. A plausible and more productive alternative would be to place all those students preferring visual learning in a given section and teach them via visual materials. All those students favoring auditory learning should be placed in another section and taught by procedures emphasizing aural learning. For example, Chastain (1969) found that results, although tentative, of a statistical analysis indicated that on the basis of selected predictors it would be possible to select those students who would be more likely to succeed in an audio-lingual class and those who would be more likely to do better in a cognitive class. Of course, in such a situation there might still be no difference in achievement between groups, but one would suspect that overall achievement would be much higher. (For a description of studies on audio-lingual teaching, see chapter 5, pages 124-26. For a description of studies on cognitive teaching, see chapter 6, pages 157-59.)
In short, methodological comparisons occupy an important, although not unique, position in research in second-language education. Such studies are needed for several reasons:
1. Teachers in the classroom can identify more readily with the procedures and results that relate directly to the classroom situation.
2. Comparisons of different classroom methods in their descriptive forms can be irsed to survey emphases in the classroom.
3. Methodological comparisons can temper the exaggerated claims of over- zealous supporters of innovative materials and procedures and provide an objective perspective to a subjective task.
4. Although the problems of language teaching can be approached by examining selected variables in isolation, the researcher can also begin with broad methodological studies and vary selected factors in subsequent tests. Approaching the questions confronting second-language education from both angles seems preferable to the exclusive use of one or the other.