CHAPTER 2. Communicative Language Teaching 33 Recent FFI studies, which seem to be more and more designed to accommodate some of the research aws mentioned previously, show a di erent tendency in outcomes, as witnessed by three more recent meta-analyses. Goo, Granema, Yilmaz and Novella (2015) analyzed 30 studies published between 1990 and 2006. Although their analysis, which addressed the concerns raised above, still con rmed an advantage of explicit over implicit instruction, they found a greater e ect size for implicit instruction than reported by both Norris and Ortega (2000) and by Spada and Tomita (2010). Andringa and Schultz (2018) re-analyzed the studies in the Spada and Tomita (2010) meta-analysis and found that, when controlling for the amount of exposure, explicit and implicit instruction were equally e ective. ey suggested that the e ectiveness of explicit instruction might reside in more intensive exposure (Andringa & Schultz, 2018). Finally, Kang et al. (2019) meta-analyzed 54 empirical studies conducted between 1980 and 2015 and found that explicit and implicit instruction were equally e ective in immediate post-tests, but implicit instruction was more e ective in delayed post-tests. Kang et al. (2019) suggest that the larger relative number of studies using free response measures might have contributed to this result, supporting Doughty’s (2003) concern that testing instruments in these past studies have favored explicit treatments. CLASSROOM RESEARCH AND SCIENTIFIC CONTROL e fact that SLA research is usually based on short interventions and uses speci c testing instruments rather than free response data may be related to the demands imposed on scienti c research. Scienti c research is considered reliable if there is maximal control over variables and ndings can be generalized (Polit & Beck, 2010). A good study can be repeated with di erent participants, using the same experimental design and producing the same outcomes (Lamal, 1990). Consequently, most SLA studies conducted to support teaching practices are short-term laboratory studies in which a single intervention targeting one speci c linguistic feature is typically tested (o en with university students as participants), and experimental conditions are fully controlled, as explained above. However, the question remains if these ndings are relevant for the multi-faceted and complex teaching programs as a whole, also evidenced by the earlier nding that delayed retention tests have been found to favor implicit over explicit language teaching in terms of long-term e ects (Kang et al. (2019). It follows that classroom research, especially in the case of longitudinal approaches, may be considered incompatible with scienti c values. According to Waters (2012), it “seems likely that classroom-level teaching methods, rather than undergoing some kind of theory-driven ‘second coming’, will continue to be based rather less on the ndings of SLA studies than on enduring situational realities” (p.448). Yet, what happens in a classroom is too complex to allow laboratory studies to inform L2 teachers on the instructional e ectiveness of an integrated foreign language teaching program.
RkJQdWJsaXNoZXIy MTk4NDMw