INVOLVEMENT LOAD HYPOTHESIS PLUS

Akifumi Yanagisawa,Stuart Webb

doi:10.1017/s0272263121000577

Abstract

AbstractThe present meta-analysis aimed to improve on Involvement Load Hypothesis (ILH) by incorporating it into a broader framework that predicts incidental vocabulary learning. Studies testing the ILH were systematically collected and 42 studies meeting our inclusion criteria were analyzed. The model-selection approach was used to determine the optimal statistical model (i.e., a set of predictor variables) that best predicts learning gains. Following previous findings, we investigated whether the prediction of the ILH improved by (a) examining the influence of each level of individual ILH components (need, search, and evaluation), (b) adopting optimal operationalization of the ILH components and test format grouping, and (c) including other empirically motivated variables. Results showed that the resulting models explained a greater variance in learning gains. Based on the models, we created incidental vocabulary learning formulas. Using these formulas, one can calculate the effectiveness index of activities to predict their relative effectiveness more accurately on incidental vocabulary learning.

Highlights

Laufer and Hulstijn’s (2001) Involvement Load Hypothesis (ILH) was designed to predict the effectiveness of instructional activities1 on incidental vocabulary learning
The results showed that the modified ILH component model was the best model as indicated by its smallest Akaike’s Information Criterion corrected (AICc) value (–149.23 on the immediate posttest and –159.72 on the delayed posttest) followed by the ILH component (–147.27 and –166.01) and the original ILH (–139.81 and –158.79) in that order
We aimed to enhance the prediction of incidental vocabulary learning by meta-analyzing studies examining the ILH

Summary

Introduction

Laufer and Hulstijn’s (2001) Involvement Load Hypothesis (ILH) was designed to predict the effectiveness of instructional activities on incidental vocabulary learning. Several studies revealed that the predictions of the ILH were not always accurate (e.g., Bao, 2015; Folse, 2006; Keating, 2008; Rott, 2012; Zou, 2017) These studies argued that the individual components (need, search, and evaluation) might contribute to learning differently (e.g., Kim, 2008; Laufer & Hulstijn, 2001) and other factors (e.g., frequency, mode of activity, and test format) should be included (e.g., Folse, 2006). Candidate influential factors as well as the ILH components are analyzed by using a model selection approach to obtain a statistical model including a combination of predictor variables that meaningfully contribute to the prediction of learning gains Based on this resulting model, we aim to create formulas to calculate the effectiveness index of activities. Need is strong when a student consults with a dictionary to look up an unknown word because they want to use the word in speech or in writing

Objectives

Methods

Results

Discussion

Conclusion