In statistics, the Lepage test is an exact distribution-free test (
nonparametric test) for jointly monitoring the location (
central tendency) and scale (
variability) in two-sample treatment versus control comparisons. It is a
rank test for the
two-sample location-scale problem. The Lepage test statistic is the squared
Euclidean distance of the standardized
Wilcoxon rank-sum test for location and the standardized
Ansari–Bradley test for scale. The Lepage test was first introduced by
Yves Lepage in 1971 in a paper in Biometrika.[1] A large number of Lepage-type tests exists in statistical literature for simultaneously testing location and scale shifts in case-control studies. The details may be found in the book: Nonparametric statistical tests: A computational approach.[2] Wolfgang Kössler[3] in 2006 also introduced various Lepage type tests using some alternative score functions optimal for various distributions. Amitava Mukherjee and Marco Marozzi introduced a class of percentile modified versions of the Lepage test.[4] An alternative to the Lepage-type tests is known as the
Cucconi test proposed by Odoardo Cucconi in 1968.[5]
Conducting the Lepage test with R
Practitioners can apply the Lepage test using the pLepage function of the contributory package NSM3,[6] built under
R software. Andreas Schulz and Markus Neuhäuser also provided detailed R code for computation of test statistic and p-value of the Lepage test[7] for the users.
Application in statistical process monitoring
In recent years, the Lepage statistic is a widely used statistical process for monitoring and quality control. In 2012, Amitava Mukherjee and Subhabrata Chakraborti introduced a distribution-free
Shewhart-type Phase-II monitoring scheme[8] (
control chart) for simultaneously monitoring of location and scale parameter of a process using a test sample of fixed size, when a reference sample of sufficiently large size is available from an in-control population. Later in 2015, the same statisticians along with Shovan Chowdhury, proposed a distribution-free
CUSUM-type Phase-II monitoring scheme[9] based on the Lepage statistic. In 2017, Mukherjee further designed an EWMA-type distribution-free Phase-II monitoring scheme[10] for joint monitoring of location and scale. In the same year, Mukherjee, with Marco Marozzi, known for promoting the Cucconi test, came together to design the Circular-Grid Lepage chart – a new type of joint monitoring scheme.[11]
Multisample version of the Lepage test
In 2005, František Rublìk introduced the multisample version of the original two-sample Lepage test.[12]
^Kössler, W. (Wolfgang) (2006). Asymptotic power and efficiency of lepage-type tests for the treatment of combined location-scale alternatives. Humboldt-Universität zu Berlin.
doi:
10.18452/2462.
hdl:
18452/3114.
OCLC243600853.
^Mukherjee, Amitava; Marozzi, Marco (2019-08-01). "A class of percentile modified Lepage-type tests". Metrika. 82 (6): 657–689.
doi:
10.1007/s00184-018-0700-1.
ISSN1435-926X.
^Cucconi, Odoardo (1968). "Un Nuovo Test non Parametrico per Il Confronto Fra Due Gruppi di Valori Campionari". Giornale Degli Economisti e Annali di Economia. 27 (3/4): 225–248.
JSTOR23241361.
^Mukherjee, A.; Chakraborti, S. (2011-09-26). "A Distribution-free Control Chart for the Joint Monitoring of Location and Scale". Quality and Reliability Engineering International. 28 (3): 335–352.
doi:
10.1002/qre.1249.
ISSN0748-8017.
^Mukherjee, Amitava (2017-02-18). "Distribution-free phase-II exponentially weighted moving average schemes for joint monitoring of location and scale based on subgroup samples". The International Journal of Advanced Manufacturing Technology. 92 (1–4): 101–116.
doi:
10.1007/s00170-016-9977-2.
ISSN0268-3768.
^Mukherjee, Amitava; Marozzi, Marco (2016-05-17). "Distribution-free Lepage Type Circular-grid Charts for Joint Monitoring of Location and Scale Parameters of a Process". Quality and Reliability Engineering International. 33 (2): 241–274.
doi:
10.1002/qre.2002.
ISSN0748-8017.
^Rublík, František (2005). "The multisample version of the Lepage test". Kybernetika. 41 (6): [713]–733.
hdl:
10338.dmlcz/135688.
ISSN0023-5954.