The Effect of Listening Instruction on the Development of L2 Learners’ Listening Competence: a Meta-Analysis

ABSTRACT

This study synthesized the research on listening instruction (LI) and its effectiveness on L2 learners’ listening competence development based on studies published from 1980 to 2019. After a literature search, well-designed studies that involved not only experimental or quasi-experimental and control groups but also pretests and posttests were selected for the analysis. A total of 39 samples from 32 primary studies with 2,797 participants were included. A random-effect model was employed to assess the differences in the treatment effect, which might be moderated by mediator variables: treatment length, treatment method, implementation types, and measurement types. The result showed a small effect size (Cohen’s d = 0.25) for listening instruction. Meta-regression for treatment length showed that the longer the treatment period is, the better the effectiveness is. A significant difference was found for implementation types, but not so for the treatment methods and measurement types. Pedagogical implications for LI are discussed, and directions for future studies are provided.

Acknowledgments

This study was financially supported by a research grant from the Ministry of Science and Technology, Taiwan (Grant #107-2410-H-266-001-). I would like to express our thanks to statistics consultant, Jack Chen, for helping with the data analysis, to Jerry Mills for proofreading the manuscript, and to my research team members for their various assistance.

Disclosure Statement

No potential conflict of interest was reported by the author(s).

Ethical Approval Number

Not applicable.

Notes

¹ Publication bias is a common term in a meta-analysis, and is “a type systematic error when synthesizing evidence that cannot represent the underlying truth” (Shi & Lin, 2019, p. 1). To be specific. it is an estimate for the difference between published and unpublished results. It occurs due to the directions and the dissemination of research results. Studies with statistically significant or positive results tend to be more likely to be published than studies with nonsignificant or negative results. In a meta-analysis, unpublished results are normally unavailable because it is difficult to obtain papers that have not been published or disseminated (see Song, Hooper & Loke, 2013 for a comprehensive discussion).

² Trim and fill method: When publication bias occurs, the funnel plot is asymmetric. Trim and fill is a method to make the funnel plot more symmetric. The method is to remove the studies that cause a funnel plot’s asymmetry, and then fill imputed missing studies in the funnel plot based on the bias-corrected overall estimate (see Shi & Lin, 2019 for the method).)

³ Formal instruction in this study refers to teachers teach listening based on a pre-designed curriculum or commercially-published textbooks.

⁴ A primary study: A primary study is a study that is included in a meta-analysis; however, a primary study may contain more than one sample study; therefore, in the text, one may read 32 primary studies with 39 sample studies.

⁵ Relative weight: Relative weight is an estimate of the proportional contribution that each individual study makes to the pooled analysis.