Heteroscedasticity testing after outlier removal

Abstract

Given the e¤ect that outliers can have on regression and speci…cation testing, a vastly used robusti…cation strategy by practitioners consists in: (i) starting the empirical analysis with an outlier detection procedure to deselect atypical data values; then (ii) continuing the analysis with the selected non-outlying observations. The repercussions of such robustifying procedure on the asymptotic properties of subsequent inferential procedures are, however, underexplored. We study the e¤ects of such a strategy on testing for heteroscedasticity. Speci…cally, using weighted and marked empirical processes of residuals theory, we show that the White test implemented after the outlier detection and removal is asymptotically chi-square if the underlying errors are symmetric. In a simulation study, we show that -depending on the type of outliers -the standard White test can be either severely undersized or oversized, as well as have trivial power. The statistic applied after deselecting outliers has good …nite sample properties under symmetry but can su¤er from size distortions under asymmetric errors. Given these results, we devise an empirical modeling strategy to guide practitioners whose preferred approach is to remove outliers from the sample.

Analysis of Ergonomic Factor Measurement for Transjakarta Bus Drivers

“…Parameters of data that had a normal distribution had a P-Value with an alpha value of more than 0.05 while for data that has an abnormal distribution the alpha value was less than 0.05. The results of the assumption test of the activity variable reaching the button from the use of the RULA analysis method could be seen in Figure 5, which displayed the Summary Report for Button Outreach Activities [10]. The results of the variable assumption test to reach the steering wheel from the use of the RULA analysis method can be seen in Figure 5 Table Summary Report for Steering Outreach Activities.…”

Valid Inference Corrected for Outlier Removal

“…In fact, illustration of the problem with detect-and-forget has appeared in some literature. Berenguer-Rico and Wilms (2018) showed how the White test for heteroscedasticity can fail using the detect-andforget approach under asymmetric errors; however, when the errors come from a symmetric distribution, they show how the theory of Berenguer-Rico and Nielsen (2017) can lead to the detect-and-forget approach being valid asymptotically and having good finite-sample performance.…”