Gespeichert in:
| 1. Verfasser: | |
|---|---|
| Format: | Artículo científico |
| Sprache: | en |
| Veröffentlicht: |
Universitat de València
2002
|
| Schlagworte: | |
| Online-Zugang: | https://www.redalyc.org/articulo.oa?id=16923209 |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866812337619468288 |
|---|---|
| author | Bruno D. Zumbo |
| author_facet | Bruno D. Zumbo |
| contents | The Robustness of Validity and Efficiency of the Related Samples t-Test in the Presence of Outliers Bruno D. Zumbo Martha J. Jennings James F. Joula Psicología The performance of the related samples t-test (a one-sample t-test appliedto the difference scores) given data which are essentially normal butcontain outliers is largely unknown. In this Monte Carlo study therobustness of validity and efficiency for both the paired and one-sample ttestsare investigated. The Type I error rate and power of these testsgiven a normal underlying population are compared with the performanceof these tests given a systematic range of outlier contamination in theunderlying population. Sample sizes of 8, 16, 32, 64, and 128 areincluded in the design. Robustness of validity results are explored usingregression models. Robustness of efficiency results are expressed using aproposed fairly stringent criterion for power. The results indicate that thet-test demonstrates fairly stringent robustness of validity for the range ofsymmetric contamination explored. When contamination is asymmetricthe Type I error rate becomes inflated as the proportion of contaminationincreases. If robustness of validity is intact, power is not greatly affectedwhen medium or large effect sizes are examined. This is not necessarilytrue for small effect sizes and the problems are further exacerbated whensample sizes are also small. Finally, a model with practical relevance fordata analysts confronted with outlier contaminated data is developed usinga novel index of contamination. This model is compared with a modelusing skewness and kurtosis values as disributional measures. 2002 artículo científico 0211-2159 https://www.redalyc.org/articulo.oa?id=16923209 en http://www.redalyc.org/revista.oa?id=169 Psicológica application/pdf Universitat de València Psicológica (España) Num.2 Vol.23 |
| format | Artículo científico |
| id | redalyc_16923209 |
| language | en |
| publishDate | 2002 |
| publisher | Universitat de València |
| spellingShingle | The Robustness of Validity and Efficiency of the Related Samples t-Test in the Presence of Outliers Bruno D. Zumbo Psicología The Robustness of Validity and Efficiency of the Related Samples t-Test in the Presence of Outliers Bruno D. Zumbo Martha J. Jennings James F. Joula Psicología The performance of the related samples t-test (a one-sample t-test appliedto the difference scores) given data which are essentially normal butcontain outliers is largely unknown. In this Monte Carlo study therobustness of validity and efficiency for both the paired and one-sample ttestsare investigated. The Type I error rate and power of these testsgiven a normal underlying population are compared with the performanceof these tests given a systematic range of outlier contamination in theunderlying population. Sample sizes of 8, 16, 32, 64, and 128 areincluded in the design. Robustness of validity results are explored usingregression models. Robustness of efficiency results are expressed using aproposed fairly stringent criterion for power. The results indicate that thet-test demonstrates fairly stringent robustness of validity for the range ofsymmetric contamination explored. When contamination is asymmetricthe Type I error rate becomes inflated as the proportion of contaminationincreases. If robustness of validity is intact, power is not greatly affectedwhen medium or large effect sizes are examined. This is not necessarilytrue for small effect sizes and the problems are further exacerbated whensample sizes are also small. Finally, a model with practical relevance fordata analysts confronted with outlier contaminated data is developed usinga novel index of contamination. This model is compared with a modelusing skewness and kurtosis values as disributional measures. 2002 artículo científico 0211-2159 https://www.redalyc.org/articulo.oa?id=16923209 en http://www.redalyc.org/revista.oa?id=169 Psicológica application/pdf Universitat de València Psicológica (España) Num.2 Vol.23 |
| title | The Robustness of Validity and Efficiency of the Related Samples t-Test in the Presence of Outliers |
| topic | Psicología |
| url | https://www.redalyc.org/articulo.oa?id=16923209 |