Interval estimation in the presence of an outlier
Outliers are often ubiquitous in surveys that involve linear measurements. Knowing that the presence of such extreme points can grossly distort statistical analyses, most researchers are often tempted to conveniently eliminate them from the data set without much careful consideration. In this study,...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Sunway University
2011
|
| Subjects: | |
| Online Access: | http://eprints.sunway.edu.my/386/ http://eprints.sunway.edu.my/386/1/SAJ_8_2011_26-37.pdf |
| Summary: | Outliers are often ubiquitous in surveys that involve linear measurements. Knowing that the presence of such extreme points can grossly distort statistical analyses, most researchers are often tempted to conveniently eliminate them from the data set without much careful consideration. In this study, we investigate the performance of confidence intervals for the population mean under the various probabilities of outlier being caused by uncorrectable human errors. The sample under study is randomly generated and subscribed to a normal distribution, and it contains only one outlier at one of the two extreme ends. |
|---|