Test Statistics for Unknown Means of Autocorrelated and Skewed Populations

• Activity Number: 494 - Simulation-Based Approaches
• Type: Contributed
• Date/Time: Thursday, August 6, 2020 : 10:00 AM to 2:00 PM
• Sponsor: Section on Statistical Computing
• Author(s): Kyungduk Ko*
• Companies: Boise State University
• Keywords: Autocorrelation; Robustness; Skewness; Test of hypothesis

Abstract

We propose test statistics for an unknown population mean, which are robust to not only autocorrelation but also skewness. An existing statistic widely used for i.i.d. cases is adjusted for skewness through expansions of approximating quantiles with cumulants and then modified those skewness-adjusted statistics for autocorrelation. The procedure is presented in detail, and the proposed test statistics are assessed via Monte Carlo simulation and an application to real data.