| 摘要 | 第1-11页 |
| ABSTRACT | 第11-24页 |
| 1 背景介绍 | 第24-44页 |
| ·Meta-analysis | 第24-27页 |
| ·Brief review about origin and development of meta-analysis | 第24-25页 |
| ·Fix effect and random effect of meta-analytic model | 第25-27页 |
| ·r-tpye effect size family | 第27-39页 |
| ·Converting reported statistical information into r-type effect size metric | 第29-31页 |
| ·Meta-analysis of r | 第31-32页 |
| ·r_(eq) effect size,point-biserial correlation(r_(pb)),r_(pb)~2,d | 第32-39页 |
| ·Research direction | 第39-44页 |
| 2 研究目的 | 第44-46页 |
| ·Motivation for current study | 第44页 |
| ·Research question | 第44-46页 |
| 3 方法 | 第46-54页 |
| ·Design factors | 第48-50页 |
| ·Number of study(k):4 levels | 第49页 |
| ·Average sample size in primary study(n=n1=n2):4 levels | 第49页 |
| ·Group variance in primary study(var):2 levels | 第49页 |
| ·Statistical dependence and independence (corr):3 levels | 第49页 |
| ·Normality and non-normality (pop):2 levels | 第49-50页 |
| ·Population effect size (p):4 levels | 第50页 |
| ·Inferential test (Test):2 levels | 第50页 |
| ·Heterogeneity in effect size (H):2 levels | 第50页 |
| ·Publication bias (PB):2 levels | 第50页 |
| ·Programming | 第50-51页 |
| ·Evaluation criteria | 第51-53页 |
| ·Bias | 第51-52页 |
| ·Mean square error (MSE) | 第52页 |
| ·Standardized bias (Stdb) | 第52-53页 |
| ·Statistical Method | 第53-54页 |
| ·Logistic regression analysis | 第53-54页 |
| 4 结果 | 第54-154页 |
| ·Aggregated results | 第54-60页 |
| ·Number of study (k) | 第54-55页 |
| ·Average sample size (n) | 第55-56页 |
| ·Homogeneity and heterogeneity in effect size | 第56-57页 |
| ·With and without publication bias | 第57-58页 |
| ·Population effect size | 第58-60页 |
| ·Condition classification when ρ=0.1 | 第60-148页 |
| ·Under equal variances (1:1), independent (corr=0.0), normally distributed data and observed r_(eq) derived from t test | 第60-64页 |
| ·Under equal variances (1:1), independent (corr=0.0), normally distributed data and observed r_(eq) derived from Wilcoxon test | 第64-68页 |
| ·Under unequal variances (1:10), independent (corr=0.0), normally distributed data and observed r_(eq) derived from t test | 第68-71页 |
| ·Under unequal variances (1:10), independent (corr=0.0), normally distributed data and observed r_(eq) derived from Wilcoxon test | 第71-75页 |
| ·Under equal variances (1:1), non-independent (corr=0.2), normal data and observed r_(eq) derived from t test | 第75-78页 |
| ·Under equal variances (1:1), non-independent (corr=0.2), normal dataand observed r_(eq) derived from Wilcoxon test | 第78-82页 |
| ·Under equal variances (1:1), non-independent (corr=0.7), normal dataand observed r_(eq) derived from t test | 第82-86页 |
| ·Under equal variances (1:1), non-independent (corr=0.7), normal dataand observed r_(eq) derived from Wilcoxon test | 第86-90页 |
| ·Under equal variances (1:1), independent (corr=0.0), non-normal dataand observed r_(eq) derived from t test | 第90-93页 |
| ·Under equal variances (1:1), independent (corr=0.0), non-normaldata and observed r_(eq) derived from Wilcoxon test | 第93-96页 |
| ·Under unequal variance (1:10), non-independent (corr=0.2), normaldata and observed r_(eq) derived from t test | 第96-100页 |
| ·Under unequal variance (1:10), non-independent (corr=0.2), normaldata and observed r_(eq) derived from Wilcoxon test | 第100-104页 |
| ·Under unequal variances (1:10), non-independent (corr=0.7),normal data and observed r_(eq) derived from t test | 第104-108页 |
| ·Under unequal variances (1:10), non-independent (corr=0.7),normal data and observed r_(eq) derived from Wilcoxon test | 第108-112页 |
| ·Under unequal variances (1:10), independent (corr=0.0), non-normal data and observed r_(eq) derived from t test | 第112-115页 |
| ·Under unequal variances (1:10), independent (corr=0.0), non-normal data and observed r_(eq) derived from Wilcoxon test | 第115-119页 |
| ·Under equal variances (1:1), non-independent (corr=0.2), non-normally distributed data and observed r_(eq) derived from t test | 第119-123页 |
| ·Under equal variances (1:1), non-independent (corr=0.2), non-normally distributed data and observed r_(eq) derived from Wilcoxon test | 第123-127页 |
| ·Under equal variances (1:1), non-independent (corr=0.7), non-normal data and observed r_(eq) derived from t test | 第127-131页 |
| ·Under equal variances (1:1), non-independent (corr=0.7), non-normal data and observed r_(eq) derived from Wilcoxon test | 第131-134页 |
| ·Under enequal variances (1:10), non-independent (corr=0.2), non-normal data and observed r_(eq) derived from t test | 第134-137页 |
| ·Under unequal variances (1:10), non-independent (corr=0.2), non-normal data and observed r_(eq) derived from Wilcoxon test | 第137-141页 |
| ·Under unequal variances (1:10), non-independent (corr=0.7), non-normal data and observed r_(eq) derived from t test | 第141-145页 |
| ·Under unequal variances (1:10), non-independent (corr=0.7), non-normal data and observed r_(eq) derived from Wilcoxon test | 第145-148页 |
| ·Effect of study factors contributing to the likelihood of Stdb50 for each meta-analytic estimator | 第148-149页 |
| ·HSr estimator | 第148页 |
| ·HOz and HOr estimator | 第148-149页 |
| ·Condition classification when ρ=0.3 | 第149-153页 |
| ·Example | 第153-154页 |
| 5 讨论 | 第154-163页 |
| ·Overview | 第154-158页 |
| ·Methodological assumptions and inferential tests, number of study and average sample size | 第155-157页 |
| ·The effect of Publication bias and heterogeneity in effect size | 第157-158页 |
| ·Recommendation for the use of r_(eq) in meta-analysis | 第158-159页 |
| ·Emphasis on the appropriate use and adequate report of effect size | 第159-161页 |
| ·Strength, limitations, and future directions | 第161-163页 |
| 6 结论 | 第163-164页 |
| 参考文献 | 第164-170页 |
| 附录 | 第170-181页 |
| 攻读学位期间成果 | 第181-182页 |
| 致谢 | 第182-184页 |
| 统计学证明 | 第184页 |
