#### DMCA

## On the post hoc power in testing mean differences (2005)

Venue: | Journal of Educational and Behavioral Statistics |

Citations: | 6 - 0 self |

### Citations

10859 | Statistical power analysis for the behavioral sciences (2nd ed - Cohen - 1988 |

4960 | An Introduction to the Bootstrap. - Efron, Tibshirani - 1993 |

1599 |
Alternative ways of assessing model fit
- Browne, Cudeck
- 1993
(Show Context)
Citation Context ... Consequently, the 90% confidence limits in [yL, yu] are numbers satisfying F(to IyL)=0.95 and F(to lyu)z= 0.05. (5) Equation 5 has been used by several authors to obtain a confidence interval (e.g., =-=Browne & Cudeck, 1993-=-; MacCallum, Browne, & Sugawara, 1996; Steiger & Fouladi, 1997), which is valid as long as F(t ly) is monotonically decreasing in y. Applying Equation 5 with F = F.a, the appendix provides details lea... |

626 | SPSS survival manual: a step by step guide to data analysis using - Pallant - 2001 |

459 |
Power analysis and determination of sample size for covariance structure modeling
- MacCallum, Browne, et al.
- 1996
(Show Context)
Citation Context ...confidence limits in [yL, yu] are numbers satisfying F(to IyL)=0.95 and F(to lyu)z= 0.05. (5) Equation 5 has been used by several authors to obtain a confidence interval (e.g., Browne & Cudeck, 1993; =-=MacCallum, Browne, & Sugawara, 1996-=-; Steiger & Fouladi, 1997), which is valid as long as F(t ly) is monotonically decreasing in y. Applying Equation 5 with F = F.a, the appendix provides details leading to YL = l- (ZI-a + Z09 5 -rn) an... |

92 | Intermediate Statistics: A Modern Approach - Stevens - 1990 |

78 | The abuse of power: the pervasive fallacy of power calculations for data analysis
- HOENIG, HEISEY
- 2001
(Show Context)
Citation Context ...ll One can go one step further using the estimated effect size to construct a power estimate for the current study. This estimator is called the post hoc power (Gillett, 1994a) or the observed power (=-=Hoenig & Heisey, 2001-=-). The question is how much information the observed power provides us about the power underlying the current study. Hoenig and Heisey (2001) cite 19 journals across a variety of disciplines where rec... |

50 |
Noncentrality interval estimation and the evaluation of statistical methods
- Steiger, Fouladi
- 1997
(Show Context)
Citation Context ...bers satisfying F(to IyL)=0.95 and F(to lyu)z= 0.05. (5) Equation 5 has been used by several authors to obtain a confidence interval (e.g., Browne & Cudeck, 1993; MacCallum, Browne, & Sugawara, 1996; =-=Steiger & Fouladi, 1997-=-), which is valid as long as F(t ly) is monotonically decreasing in y. Applying Equation 5 with F = F.a, the appendix provides details leading to YL = l- (ZI-a + Z09 5 -rn) and Yu = 1-4)(Z-a -ZO95 -W)... |

31 | Effect sizes and p values: What should be reported and what should be replicated - Greenwald, Gonzalez, et al. - 1996 |

18 | Reporting of statistical inference - Finch, Cumming, et al. - 2001 |

7 | Making Sense of Statistics in Psychology - Everitt - 1996 |

4 | Computing confidence bounds for power and sample size of the general linear univariate model. Am Stat - Taylor, Muller - 1995 |

3 | Confidence intervals for power with special reference to medical trials - Dudewicz |

3 |
Fortran 77 with numerical methods for engineers and scientists
- Etter
- 1992
(Show Context)
Citation Context ... + az)]2 can be approximated by 1 fJ ['D(h + at)2e- kdt (A5) with an error of less than 10-8. The integral in Equation A5 can be evaluated by the well-known Simpson rule of numerical integration (see =-=Etter, 1992-=-). We used m = 50,000 points in the Simpson rule when calculating the numbers for plots in Figures 1 and 2. Double the number of points gave the same value of the integral to the 8th decimal place. No... |

3 |
Post hoc power analysis
- Gillett
- 1994
(Show Context)
Citation Context ...mproved the article. 141 Yuan and Maxwell One can go one step further using the estimated effect size to construct a power estimate for the current study. This estimator is called the post hoc power (=-=Gillett, 1994-=-a) or the observed power (Hoenig & Heisey, 2001). The question is how much information the observed power provides us about the power underlying the current study. Hoenig and Heisey (2001) cite 19 jou... |

3 | Using p values to estimate the probability of a statistically significant replication - Posavac - 2002 |

1 |
The average power criterion for sample size estimation
- Gillett
- 1994
(Show Context)
Citation Context ...mproved the article. 141 Yuan and Maxwell One can go one step further using the estimated effect size to construct a power estimate for the current study. This estimator is called the post hoc power (=-=Gillett, 1994-=-a) or the observed power (Hoenig & Heisey, 2001). The question is how much information the observed power provides us about the power underlying the current study. Hoenig and Heisey (2001) cite 19 jou... |

1 | Effect sizes andp values: What should be reported and what should be replicated - Greenwald, Gonzalez, et al. - 1996 |

1 |
On determining replication probabilities: Comments on Posavac
- Macdonald
- 2003
(Show Context)
Citation Context ... observed power is intended to estimate. In fact, Posavac (2002) suggests supplementing such indices as confidence intervals with the probability of a statistically significant exact replication (see =-=Macdonald, 2003-=-; Posavac, 2003, for further discussion of this suggestion). Thus, examining the accuracy of observed power is important not only because some journals and authors have recommended that observed power... |

1 | Hoc Power in Testing Mean Differences Posavac - Post |

1 | Intermediate statistics: A modem approach (2nd - Stevens - 1999 |