Re-assessing type and amount of exposure in form-focused instruction: Symposium

A large body of research on form-focused instruction (FFI) has been carried out over the past decades; many of these studies have been synthesized in several meta-analyses and research reviews (e.g. Norris & Ortega, 2000; De Graaff & Housen, 2009; Spada & Tomita, 2010). These indicate quite convincingly that explicit types of instruction are more effective than implicit types of instruction, judging from the substantially larger effect sizes that are generally observed for explicit treatments. However, several researchers have cautioned that the effects of explicit instruction may have been overestimated (e.g. Doughty, 2003). Research designs may have favored explicit types of instruction for several reasons; three of these reasons are discussed in this colloquium. First, Meta-analyses have not taken differences in amount of exposure to the target structure into account. This may have favored explicit types of instruction, as these may often be more intensive in terms of exposure: each and every item in explicit exercises tends to expose learners to the target structure, while in implicit types of instruction exposure may be more incidental and less intensive. By means of a meta-reanalysis of Spada and Tomita (2010), Sible Andringa investigated whether explicit instruction may be effective because explicit information as such offers an advantage to L2 learners or because of inherent extra exposure in explicit instruction settings. Second, instructional treatments in FFI research have tended to focus on specific language structures and on short-term outcomes; only very few longitudinal FFI studies have been conducted. Short treatments may well work against implicit types of instruction, as implicit learning processes are often claimed to require more time on task in order to be effective. Two studies in this colloquium present data in which explicit and implicit types of instruction are compared in longitudinal designs. Leslie Piggott’s study one examined differences in L2 English writing proficiency after a one-year communicative program. In one version of the program students received explicit grammar instruction and practice; in the other version they did not, receiving additional communicative tasks instead. Writing assignments were rated and analyzed, focusing on accuracy and complexity. Audrey Rousse-Malpat’s longitudinal study investigated differences in L2 French oral and writing proficiency after a three-year communicative program. In one version of the program students received explicit grammar instruction and practice; in the other version they followed a completely implicit method. Free productive oral and written assignments were scored holistically. A third reason why FFI research may have been biased against implicit treatments is its overreliance on more explicit measures of linguistic ability, which may favor explicit types of instruction. Wim Gombert’s study is a validation of a new measure for tracking oral language proficiency using the Student Oral Proficiency Assessment in intermediate L2 French high-school students. It discusses the reliability and validity of the test for the assessment of fluency, lexical adequacy, grammatical accuracy and listening comprehension. Marjolijn Verspoor will conclude and wrap up the methodology and findings from the other studies, and propose some recommendations for future research as well as for classroom practice. The papers in this colloquium emphasize different challenges that FFI research faces for it to be maximally relevant to educational practice. It is proposed that longitudinal studies in naturalistic FL teaching settings are required that include measurements of L2 proficiency in non-focused productive tasks. In a concluding discussion we will address issues such as: How can implicit and explicit treatments be optimally balanced in language teaching and how can this balance be estimated? What are the advantages of focused and unfocused measures of progress in long term treatment designs? Should exposure differences be controlled for and how can this be achieved?