Trends in Spontaneous Language Contact During a One-Semester Study Abroad Program: Data Synthesis and Bootstrapping Approach

Session Submission Type: Paper

Summary

This study estimates average hours per week spent on spontaneous target language reading, listening, writing, and speaking during a one-semester study abroad program. We synthesized the data of seven primary studies (N = 221) and computed the bootstrapped means for each one of four language activities.

Abstract

This study provides estimates on the average hours per week (hours/week) spent on the four types of spontaneous target language (TL) activities (reading, listening, writing, and speaking) in a study abroad (SA) setting. Although such statistics are of great importance to academic discussions, the trend toward small samples thus far has prevented SA research from obtaining precise population-level estimates on this topic. To solve this problem, we employed research synthesis and bootstrapping approaches. Of 192 initially pooled studies (published between 1990 and 2016), the data of seven studies (N = 221) were synthesized based on their similarities in measurement and participants’ profiles: all seven studies used a self-report questionnaire called a language contact profile (LCP) for the data collection and sampled intermediate-level adult language learners who participated in a one-semester SA program. In data synthesis, we computed weighted means and standard deviations for the hours/week spent on spontaneous TL reading, listening, writing, and speaking (N = 221, each). Then, to obtain as precise estimates as possible, we computed the bootstrapped 95% confidence intervals (CIs) for the mean hours/week spent on each one of four TL activities using the weighted statistics calculated above (B = 20,000, each). Note that our sample size satisfies the minimum requirement for non-parametric bootstrapping (N ≥ 50) (Chernick, 2008). The 95% CIs of mean hours/week were [9.44, 11.75] (M = 10.60) for reading, [19.40, 23,40] (M = 21.45) for listening, [9.02, 11.53] (M = 10.36) for writing, and [17.31, 21.04] (M = 19.20) for speaking. Such baseline estimates help facilitate discussions on individual differences and group characteristics; thus, the results of this study will benefit future SA studies, especially ones that use the LCP as a form of measurement and are undertaken with small sample sizes.

Strand

• Assessment and Evaluation (ASE)

Presenters

• Hitoshi Mikami, Chubu University

• Toru Kinoshita, Nagoya University