Quantitative Research Methods in Translation and Interpreting Studies by Christopher D. Mellinger & Thomas A. Hanson

Author:Christopher D. Mellinger & Thomas A. Hanson
Language: eng
Format: epub
ISBN: 9781317299226
Publisher: Taylor & Francis
Published: 2017-04-11T04:00:00+00:00

Repeated measures ANOVA

As we have described so far in this chapter, the one-way ANOVA model can be conceived as an extension of the t-test. In both cases, a measurement is made on different groups (two groups for the t-test and three or more groups for the one-way ANOVA). The groups are independent from each other, with data collected from different people or different corpora, for example. However, in some cases it is preferable to collect data from the same participants multiple times. This experimental design is called a repeated measures ANOVA (rANOVA) and can also be referred to as a within-subjects design.23 A one-way ANOVA is an extension of the independent samples t-test, and rANOVA is the analogous extension of the paired t-test.

There are two primary benefits to rANOVA. First, because participants in a study are measured multiple times, their individual variation can be eliminated as a source of variation between the groups. Sometimes this is described by saying that each individual serves as his or her own control. From a statistical perspective, the error term will be reduced and the power of statistical tests will be increased. Second, the experimental design of rANOVA requires fewer participants, so it is particularly attractive when participant pools are limited in size, difficult to access, or expensive to obtain. This situation is common to T&I research; finding volunteers who meet study criteria can be difficult, time-consuming, or costly. Therefore, it is important to maximize data collection and inference.24 Repeated measures ANOVA contributes to those goals by using the same participants in multiple rounds of data collection.

The design of a repeated measures ANOVA can be recognized from the fact that each participant is measured on the dependent variable three or more times. There are two main ways to implement repeated measurements. First, the participants could be observed at multiple points in time. For example, in Fraser’s (1999) study of lexical processing strategies, eight participants were assessed at four different points in time. Measurements included their rate of use of three lexical processing strategies (consult, ignore, and infer), as well as vocabulary retention. The independent variable in this study is time, and we will refer to this kind of model as a time-based rANOVA.

A second model that utilizes rANOVA involves measuring a response variable for participants in three or more different treatment conditions. Folse (2006) employs this experimental design by having students practice vocabulary exercises in three different conditions. The first learning strategy was a fill-in-the-blank exercise; the second consisted of three fill-in-the-blank exercises; and the third strategy required original sentence writing. Each participant completed a vocabulary quiz following each of the three instructional methods. The levels of the independent factor in this type of model are the different conditions under which the measurements are made (the three learning strategies in this case), and we will refer to this as a treatment-based rANOVA.

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.