Saved in:
| Main Author: | |
|---|---|
| Format: | Artículo científico |
| Language: | en |
| Published: |
Universitat de València
2003
|
| Subjects: | |
| Online Access: | https://www.redalyc.org/articulo.oa?id=16924109 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Table of Contents:
- Bias in Estimation and Hypothesis Testing of Correlation Donald W. Zimmerman Richard H. Williams Bruno D. Zumbo Psicología This study examined bias in the sample correlation coefficient, r, and itscorrection by unbiased estimators. Computer simulations revealed that theexpected value of correlation coefficients in samples from a normalpopulation is slightly less than the population correlation, ρ, and that thebias is almost eliminated by an estimator suggested by R.A. Fisher and ismore completely eliminated by a related estimator recommended by Olkinand Pratt. Transformation of initial scores to ranks and calculation of theSpearman rank correlation, rS, produces somewhat greater bias. Type I errorprobabilities of significance tests of zero correlation based on the Student tstatistic and exact tests based on critical values of rS obtained frompermutations remain fairly close to the significance level for normal andseveral non-normal distributions. However, significance tests of non-zerovalues of correlation based on the r to Z transformation are grossly distortedfor distributions that violate bivariate normality. Also, significance tests ofnon-zero values of rS based on the r to Z transformation are distorted evenfor normal distributions. 2003 artículo científico 0211-2159 https://www.redalyc.org/articulo.oa?id=16924109 en http://www.redalyc.org/revista.oa?id=169 Psicológica application/pdf Universitat de València Psicológica (España) Num.1 Vol.24