On Misuses of the Kolmogorov–Smirnov Test for One-Sample Goodness-of-Fit

Anthony Zeimbekakis, Elizabeth D. Schifano and Jun Yan

The American Statistician, 2024, vol. 78, issue 4, 481-487

Abstract: The Kolmogorov–Smirnov (KS) test is widely employed to assess the goodness-of-fit of a hypothesized continuous distribution to a sample. Despite its popularity, the test is frequently misused in the literature and practice. While originally intended for independent, continuous data with precisely specified hypothesized distributions, it is erroneously applied to scenarios with dependent, discrete, or rounded data, with hypothesized distributions requiring estimated parameters. For example, it has been “discovered” multiple times that the test is too conservative when the hypothesized distribution has parameters that need to be estimated. We demonstrate misuses of the one-sample KS test in three scenarios through simulation studies: (a) the hypothesized distribution has unspecified parameters; (b) the data are serially dependent; and (c) a combination of the first two scenarios. For each scenario, we provide remedies for practical applications using appropriate bootstrap approaches. The whole demonstration can be used as hands-on education materials on both goodness-of-fit tests and bootstrap.

Date: 2024
References: Add references at CitEc
Citations:

Downloads: (external link)
http://hdl.handle.net/10.1080/00031305.2024.2356095 (text/html)
Access to full text is restricted to subscribers.

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:taf:amstat:v:78:y:2024:i:4:p:481-487

Ordering information: This journal article can be ordered from
http://www.tandfonline.com/pricing/journal/UTAS20

DOI: 10.1080/00031305.2024.2356095

Access Statistics for this article

The American Statistician is currently edited by Eric Sampson

More articles in The American Statistician from Taylor & Francis Journals
Bibliographic data for series maintained by Chris Longhurst ().