The famous chi-squared goodness-of-fit test was discovered by Karl Pearson in 1900. If the partition of a sample space is such that observations are grouped over r disjoined intervals Δ i , and denoting ν i observed frequencies and np i (θ) expected that correspond to a multinomial scheme, the Pearson’s sum is written
where V(θ) is a vector with components v i (θ) = (ν i − np i (θ))(np i (θ)) − 1 ∕ 2, i = 1, …, r. If the number of observations n → ∞, the statistic (1) for a simple null hypothesis, specifying the true value of θ, will follow chi-squared probability distribution with r − 1 degrees of freedom.
Until 1934, Pearson believed that the limit distribution of his chi-squared statistic would be the same if unknown parameters of the null hypothesis were replaced by estimates based on a sample (Stigler (2008), p....
This is a preview of subscription content, log in via an institution.
Buying options
Tax calculation will be finalised at checkout
Purchases are for personal use only
Learn about institutional subscriptionsReferences and Further Reading
Anderson G (1994) Simple tests of distributional form. J Economet 62:265–276
Boero G, Smith J, Wallis KF (2004) The sensitivity of chi-squared goodness of-fit tests to the partitioning of data. Economet Rev 23:341–370
Bol’shev LN, Mirvaliev M (1978) Chi-square goodness-of-fit test for the Poisson, binomial, and negative binomial distributions. Theory Probab Appl 23:481–494 (in Russian)
Chernoff H, Lehmann EL (1954) The use of maximum likelihood estimates in tests for goodness of fit. Ann Math Stat 25:579–589
Chichagov VV (2006) Unbiased estimates and chi-squared statistic for one-parameter exponential family. In: Statistical methods of estimation and hypotheses testing, vol 19. Perm State University, Perm, Russia, pp 78–89
Cohran G (1954) Some methods for strengthening the common χ 2 tests. Biometrics 10:417–451
Dahiya RC, Gurland J (1972) Pearson chi-squared test of fit with random intervals. Biometrica 59:147–153
Dzhaparidze KO, Nikulin MS (1974) On a modification of the standard statistic of Pearson. Theory Probab Appl 19:851–853
Dzhaparidze KO, Nikulin MS (1992) On evaluation of statistics of chi-square type tests. In: Problem of the theory of probability distributions, vol 12. Nauka, St. Petersburg, pp. 59–90
Fisher RA (1924) The condition under which χ 2 measures the discrepancy between observation and hypothesis. J R Stat Soc 87:442–450
Fisher RA (1925) Partition of χ 2 into its components. In: Statistical methods for research workers. Oliver and Boyd, Edinburg
Habib MG, Thomas DR (1986) Chi-square goodness-of-fit tests for randomly censored data. Ann Stat 14:759–765
Hsuan TA, Robson DS (1976) The goodness-of-fit tests with moment type estimators. Commun Stat Theory Meth A5:1509–1519
Lemeshko BYu, Postovalov SN, Chimitiva EV (2001) On the distribution and power of Nikulin’s chi-squared test. Ind Lab 67:52–58 (in Russian)
McCulloch CE (1985) Relationships among some chi-squared goodness of fit statistics. Commun Stat Theory Meth 14:593–603
Mirvaliev M (2001) An investigation of generalized chi-squared type statistics. Academy of Sciences of the Republic of Uzbekistan, Tashkent, Doctoral dissertation
Molinari L (1977) Distribution of the chi-squared test in non-standard situations. Biometrika 64:115–121
Moore DS (1977) Generalized inverses, Wald’s method and the construction of chisquared tests of fit. J Am Stat Assoc 72:131–137
Moore DS, Spruill MC (1975) Unified large-sample theory of general chisquared statistics for tests of fit. Ann Stat 3:599–616
Nikulin MS (1973a) Chi-square test for continuous distributions. Theory Probab Appl 18:638–639
Nikulin MS (1973b) Chi-square test for continuous distributions with shift and scale parameters. Theory Probab Appl 18:559–568
Nikulin MS, Voinov VG (1989) A chi-square goodness-of-fit test for exponential distributions of the first order. Springer-Verlag Lect Notes Math 1412:239–258
Rao KC, Robson DS (1974) A chi-squared statistic for goodness-of-fit tests within the exponential family. Commun Stat 3:1139–1153
Roy AR (1956) On χ 2 statistics with variable intervals. Technical report N1, Stanford University, Statistics Department
Singh AC (1987) On the optimality and a generalization of Rao–Robson’s statistic. Commun Stat Theory Meth 16, 3255–3273
Spruill MC (1976) A comparison of chi-square goodness-of-fit tests based on approximate Bahadur slope. Ann Stat 2:237–284
Stigler SM (2008) Karl Pearson’s theoretical errors and the advances they inspired. Stat Sci 23:261–171
Voinov V (2006) On optimality of the Rao–Robson–Nikulin test. Ind Lab 72:65–70
Voinov V (2010) A decomposition of Pearson–Fisher and Dzhaparidze–Nikulin statistics and some ideas for a more powerful test construction. Commun Stat Theory Meth 39(4):667–677
Voinov V, Pya N (2010) A note on vector-valued goodness-of-fit tests. Commun Stat 39(3):452–459
Voinov V, Nikulin MS, Pya N (2008) Independently distributed in the limit components of some chi-squared tests. In: Skiadas CH (ed) Recent advances in stochastic modelling and data analysis. World Scientific, New Jersey
Voinov V, Pya N, Alloyarova R (2009) A comparative study of some modified chi-squared tests. Commun Stat Simulat Comput 38:355–367
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2011 Springer-Verlag Berlin Heidelberg
About this entry
Cite this entry
Voinov, V., Nikulin, M. (2011). Chi-Square Goodness-of-Fit Tests: Drawbacks and Improvements. In: Lovric, M. (eds) International Encyclopedia of Statistical Science. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-04898-2_172
Download citation
DOI: https://doi.org/10.1007/978-3-642-04898-2_172
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-642-04897-5
Online ISBN: 978-3-642-04898-2
eBook Packages: Mathematics and StatisticsReference Module Computer Science and Engineering