Interpreting null-hypothesis significance tests: p, effect size, and power

被引：8

作者：

Bourque, Jimmy ^{[1
]}

Blais, Jean-Guy ^{[2
]}

Larose, Francois ^{[3
]}

机构：

[1] Univ Moncton, Moncton, NB, Canada

[2] Univ Montreal, Montreal, PQ, Canada

[3] Univ Sherbrooke, Sherbrooke, PQ, Canada

来源：

REVUE DES SCIENCES DE L EDUCATION | 2009年 / 35卷 / 01期

关键词：

hypothesis testing; statistical significance; effect size; power; inference;

D O I：

10.7202/029931ar

中图分类号：

G40 [教育学];

学科分类号：

040101 ; 120403 ;

摘要：

This paper aims to explain the rationale of hypothesis testing in research. Many studies have shown some confusion as to the significance of hypothesis testing results. This confusion could partially come from discrepancies between the approaches of Fisher, Neyman and Pearson as well as Bayes. The article clarifies the information provided by the significance coefficient, effect size and power, and proposes various software to perform power and effect size analysis.

引用

页码：211 / 226

页数：16

共 51 条

[1]

American Psychological Association, 2010, PUBL MAN AM PSYCH AS

[2]

[Anonymous], 1998, BEHAV BRAIN SCI, V21

[3] Understanding statistical power in the context of applied research [J].