期刊
BEHAVIOR RESEARCH METHODS
卷 39, 期 4, 页码 973-978出版社
PSYCHONOMIC SOC INC
DOI: 10.3758/BF03192992
关键词
-
资金
- MRC [MC_U105580446] Funding Source: UKRI
- Medical Research Council [MC_U105580446] Funding Source: researchfish
- Medical Research Council [MC_U105580446] Funding Source: Medline
In most experiments that involve between-subjects or between-items factorial designs, the items and/or the participants in the various experimental groups differ on one or more variables, but need to be matched on all other factors that can affect the outcome measure. Matching large groups of items or participants on multiple dimensions is a difficult and time-consuming task, yet failure to match conditions will lead to suboptimal experiments. We describe a computer program, Match: that automates this process by selecting the best-matching items from larger sets of candidate items. In most cases, the program produces near-optimal solutions in a matter of minutes and selects matches that are typically superior to those obtained using hand matching or other semi-automated processes. We report the results of a case study in which Match was used to generate matched sets of experimental items (words varying in length and frequency) for a published study on language processing. The program was able to come up with better-matching item sets than those hand-selected by the authors of the original study, and in a fraction of the time originally taken up with stimulus matching.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据