Abstract
This study aims at mentioning briefly multiple comparison methods such as Bonferroni, Holm–Bonferroni, Hochberg, Hommel, Marascuilo, Tukey, Benjamini–Hochberg and Gavrilov–Benjamini–Sarkar for contingency tables, through the data obtained from a medical research and examining their performances by simulation study which was constructed as the total 36 scenarios to 2 × 4 contingency table. As results of simulation, it was observed that when the sample size is more than 100, the methods which can preserve the nominal alpha level are Gavrilov–Benjamini–Sarkar, Holm–Bonferroni and Bonferroni. Marascuilo method was found to be a more conservative than Bonferroni. It was found that Type I error rate for Hommel method is around 2 % in all scenarios. Moreover, when the proportions of the three populations are equal and the proportion value of the fourth population is far at a level of ±3 standard deviation from the other populations, the power value for Unadjusted All-Pairwise Comparison approach is at least a bit higher than the ones obtained by Gavrilov–Benjamini–Sarkar, Holm–Bonferroni and Bonferroni. Consequently, Gavrilov–Benjamini–Sarkar and Holm–Bonferroni methods have the best performance according to simulation. Hommel and Marascuilo methods are not recommended to be used because they have medium or lower performance. In addition, we have written a Minitab macro about multiple comparisons for use in scientific research.
Similar content being viewed by others
References
Zar JH (1999) Biostatistical analysis, 4th edn. Prentice-Hall, Upper Saddle River, NJ
Elliott AC, Reisch JS (2006) Implementing a multiple comparison test for proportions in a 2×c crosstabulation in SAS®. Proceedings of the 31st annual SAS users group international conference. SAS Institute Inc, Cary, NC
Gavrilov Y, Benjamini Y, Sarkar SK (2009) An adaptive step-down procedure with proven FDR control under independence. Ann Stat 37:619–629
Blakesley RE, Mazumdar S, Dew MA, Houck PR, Tang G, Reynolds CF 3rd et al (2009) Comparisons of methods for multiple hypothesis testing in neuropsychological research. Neuropsychology 23:255–264. doi:10.1037/a0012850
Oden N, VanVeldhuisen PC, Ingrid US, Ip MS, SCORE Study Investigator Group (2010) SCORE Study report 8: closed tests for all pair-wise comparisons of means. Drug Inf J 44:405–420
Michael GA (2007) A significance test of interaction in 2 × K designs with proportions. Tutor Quant Methods Psychol 3:1–7
Ryan TA (1960) Significance tests for multiple comparison of proportions, variances and other statistics. Psychol Bull 57:318–328
Marascuilo L (1966) Large-sample multiple comparisons. Psychol Bull 65:280–290
Holland BS, Copenhaver MD (1988) Improved Bonferroni-type multiple testing procedures. Psychol Bull 104:145–149
Westfall PH, Young SS (1989) P value adjustments for multiple tests in multivariate binomial models. J Am Stat Assoc 84:780–786
Kim SB, Tsui KL, Borodovsky M (2006) Multiple hypothesis testing in large-scale contingency tables: inferring patterns of pair-wise amino acid association in β-sheets. Int J Bioinform Res Appl 2:193–217
Horne J, Plaehn D (2007) Multiple comparisons on 2×c proportions. SAS conference proceedings: Pacific Northwest SAS Users Group 2007. Pacific Northwest SAS Users Group, Seattle, WA. www.pnwsug.org/content/multiple-comparisons-2xc-proportions. Accessed March 2014
Lancaster MB (1949) The derivation and partition of χ2 in certain discrete distributions. Biometrika 36:117–129
Haberman SJ (1973) The analysis of residuals in cross-classified tables. Biometrics 29:205–220
Hsu JC (1996) Multiple comparisons: theory and methods. CRC Press, Boca Raton
Benjamini Y, Hochberg Y (1995) Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc B 57:289–300
Holm S (1979) A simple sequentially rejective multiple test procedure. Scand J Stat 6:65–70
Hochberg Y (1988) A sharper Bonferroni procedure for multiple tests of significance. Biometrika 75:800–802
Hommel G (1988) A stagewise rejective multiple test procedure based on a modified Bonferroni test. Biometrika 75:383–386
Cangur S, Ankarali H (2014) Examining the probabilities of type i error for unadjusted all pairwise comparisons and Bonferroni adjustment approaches in hypothesis testing for proportions. Int J Stat Med Res 3(4):404–411. doi:10.6000/1929-6029.2014.03.04.9
Author information
Authors and Affiliations
Corresponding author
Electronic supplementary material
Below is the link to the electronic supplementary material.
Rights and permissions
About this article
Cite this article
Cangur, S., Ankarali, H. & Pasin, O. Comparing Performances of Multiple Comparison Methods in Commonly Used 2 × C Contingency Tables. Interdiscip Sci Comput Life Sci 8, 337–345 (2016). https://doi.org/10.1007/s12539-015-0128-5
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s12539-015-0128-5