JP Journal of Fundamental and Applied Statistics 4(1/2), pp. 1-22.
ISSN/ISBN: Not available at this time. DOI: Not available at this time.
Abstract: In this paper, we discuss several methods used to examine the goodness-of-fit of a given dataset to the so-called Benford’s law. While the use of distance measures itself suffer from theoretical founded critical values, the procedures used in literature so far do not provide a more clear statement. Apart from using graphical methods (like histogram, etc.), which are not part of the discussion, the often used Distortion Factor Model is extended to the Mantissa-Distortion-Factor (MDF) and a Benford-specific test procedure is introduced. This test is based the property of invariance of scale and base, which is an underlying feature of the First-Digit-Law. The so called Transformation-Invariance-Test (TIT) is derived as a closed form test- procedure. The practical power of the TIT is shown using a widely used macroeconomic dataset.
Bibtex:
@article {,
AUTHOR = {Peter N. Posch},
TITLE = {Benford Or Not-Benford? How To Test For The First-Digit-Law},
JOURNAL = {JP Journal of Fundamental and Applied Statistics},
YEAR = {2013},
VOLUME = {4},
NUMBER = {1-2},
PAGES = {1--22},
DOI = {},
URL = {},
}
Reference Type: Journal Article
Subject Area(s): Statistics