心理科学
心理科學
심이과학
Psychological Science
2010年
6期
1477~1480
,共null页
零假设显著性检验 元分析 假设检验结果的三分性
零假設顯著性檢驗 元分析 假設檢驗結果的三分性
령가설현저성검험 원분석 가설검험결과적삼분성
null hypothesis significance testing; meta-analysis; trichotomy of the results of hypothesis-testing
零假设检验的结果分为三种情况,它把样本空间分为三个部分：备择假设的接受域S1、零假设的接受域S0,以及接受备择假设和零假设都不合适的模糊区域S2,从而β被分为两个部分：P（X∈S2｜H1）和P（X∈S0｜H1）。数据分析表明：P（X∈S2｜H1）远大于P（X∈S0｜H1）,并且元分析所导致β的降低主要是P（X∈S2｜H1）的降低。当样本观察值落在S0中时,统计推断的不明确性会驱使研究者做元分析研究从而能够明确地作出或者接受备择假设或者接受零假设的统计决策。
零假設檢驗的結果分為三種情況,它把樣本空間分為三箇部分：備擇假設的接受域S1、零假設的接受域S0,以及接受備擇假設和零假設都不閤適的模糊區域S2,從而β被分為兩箇部分：P（X∈S2｜H1）和P（X∈S0｜H1）。數據分析錶明：P（X∈S2｜H1）遠大于P（X∈S0｜H1）,併且元分析所導緻β的降低主要是P（X∈S2｜H1）的降低。噹樣本觀察值落在S0中時,統計推斷的不明確性會驅使研究者做元分析研究從而能夠明確地作齣或者接受備擇假設或者接受零假設的統計決策。
령가설검험적결과분위삼충정황,타파양본공간분위삼개부분：비택가설적접수역S1、령가설적접수역S0,이급접수비택가설화령가설도불합괄적모호구역S2,종이β피분위량개부분：P（X∈S2｜H1）화P（X∈S0｜H1）。수거분석표명：P（X∈S2｜H1）원대우P（X∈S0｜H1）,병차원분석소도치β적강저주요시P（X∈S2｜H1）적강저。당양본관찰치락재S0중시,통계추단적불명학성회구사연구자주원분석연구종이능구명학지작출혹자접수비택가설혹자접수령가설적통계결책。
There are three kinds of results of null hypothesis significance testing,which divides the sample space into three parts：accept-area of the alternative hypothesis S1,accept-area of the null hypothesis S0,and the blurry area S2 in which one can not accept the alternative hypothesis or null hypothesis. Accordingly,β is divided into two parts：P（X∈S2-H1） and P（X∈S0-H1）. The analysis of data shows that P（X∈S2-H1） is far much larger than P（X∈S0-H1）,and the decreasing of β resulting from meta-analysis is mainly ascribed to the decreasing of P（X∈S2-H1）. When the viewed value of a sample drops S2,the indeterminacy of statistic inference will drive researchers to do meta-analysis,which would make it more possible that the researchers make a definit decision that is either the alternative hypothesis or to accept the null hypothesis.
