Comparing the simulated power of discrete goodness-of-fit tests for small sample sizes
Michael Craig Steele, Janet Chaseling, Cameron Paul Hurst
Abstract
Open-access reader
Michael Craig Steele, Janet Chaseling, Cameron Paul Hurst
Abstract
Open-access reader
Although a variety of goodness-of-fit test statistics are used by applied researchers, studies of their power have been limited. This paper investigates the simulated power of six goodness-of-fit test statistics for discrete data for small sample sizes. The null distribution is uniform and the simulated power of each of the test statistics is calculated for a number of alternative distributions including trend, triangular, flat/platykurtic type, sharp/leptokurtic type and bimodal. KEY WORDS: Goodness-of-fit, simulation.
A significance statement is not available in the OpenAlex record.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
Although a variety of goodness-of-fit test statistics are used by applied researchers, studies of their power have been limited. This paper investigates the simulated power of six goodness-of-fit test statistics for discrete data for small sample sizes. The null distribution is uniform and the simulated power of each of the test statistics is calculated for a number of alternative distributions including trend, triangular, flat/platykurtic type, sharp/leptokurtic type and bimodal. KEY WORDS: Goodness-of-fit, simulation.
Key concepts: Goodness of fit, Statistics, Sample size determination, Mathematics, Kurtosis, Statistical power, Null hypothesis, Econometrics