Statistical Report Reform in Second Language Research: A Case Of Experimental Designs

Abstract

This survey aims to review statisical report procedures in the experimental studies appearing in ten SLA and Applied Linguistic journals from 2011 to 2017. We specify our study on how the authors report and interprete their power analyses, effect sizes, and confidence intervals. Results reveal that of 217 articles, the authors reported effect sizes (70%), apriori power and posthoc power consecutively (1.8% and 6.9%), and confidence intervals (18.4%). Additionally, it shows that the authors interprete those statistical terms counted 5.5%, 27.2%, and 6%, respectively. The call for statistical report reform recommended and endorsed by scholars, researchers, and editors is inevitably echoed to shed more light on the trustworthiness and practicality of the data presented.

Keywords

• statistical reforms
• power analyses
• effect size
• confidence interval
• second language research