Amaç: Bu makalenin amacı, normallik varsayımının sağ-lanmadığı faktöriyel deneme desenlerinde etkileşim teriminin önemlilik testi için uygulaması kolay ve etkili olan hizalanmış sıra sayıları dönüşümü işlemini tanıtmaktır. Gereç ve Yöntemler: Bir-çok istatistiksel testin geçerliliği, gözlemlerin normal dağıldığı var-sayımına bağlıdır. Ancak incelenen birçok karmaşık özellik, normal olmayan dağılımlara sahiptir. Dağılımın normal olmadığı durum-larda kullanılan çeşitli yaklaşımlar mevcuttur. Normallik veya varyansların homojenliğini ihlal eden verilerin söz konusu olduğu durumlarda, başlangıçta sıra sayısı dönüşümleri alternatif olarak önerilmiştir. Sıra sayıları dönüşüm işleminin dikkate alındığı çalış-malarda önce orijinal gözlemler sıra sayılarına dönüştürülmekte, ardından da sıra sayıları üzerinde parametrik bir istatistik hesap-lanmaktadır. Bulgular: Hem en küçük kareler tekniğine hem de sıra dönüşüm işlemine alternatif bir yöntem, hizalanmış sıra sayıları dönüşüm işlemidir. Hizalanmış sıralama dönüşümü, hem normallik gibi varsayım ihlallerinin etkisini en aza indirmekte hem de sırala-ma dönüşümü için bir sorun olan etkileşim terimlerinin önemlilik testine olanak sağlamaktadır. Hizalanmış sıra sayıları işlemine da-yalı testler sağlam ve güçlü testlerdir. Bu işlem, varsayım ihlalleri-ne ve aykırı değerlere karşı duyarlı değildir. Sonuç: Hizalanmış sıra sayıları dönüşüm işlemlerinin klasik varyans analizi işlemlerine karşı bilinen en iyi alternatifler olduğu ve normallik varsayımlarının şüpheli olduğu durumlarda kullanılması gerektiği sonucuna varıl-mıştır.
Anahtar Kelimeler: İstatistiksel analiz; nonparametrik etkileşim testleri; hizalanmış sıra sayısı testi; faktöriyel analiz
Objective: The purpose of this article is to introduce aligned rank transform which is easy and effective to apply for the significance test of interaction term in factorial experimental design where normality assumption is not provided. Material and Meth-ods: The validity of many statistical tests depends on the assump-tion that observations are normally distributed. But, many complex traits studied have non-normal distributions. Several approaches exist to respond to non-normality. Rank transformations were ini-tially proposed as an alternative when dealing with data that violat-ed normality or homogeneity of variances. In studies where the rank transformation is taken into consideration, the original obser-vations are first converted into ranks, and then a parametric statis-tics is calculated on the ranks. Results: An alternative method to both the least squares technique and the rank transform process is the aligned rank transform procedure. The aligned rank transform minimizes the effect of violations of assumptions such as normality, but does not suffer some of the same problems of the rank trans-form, such as introducing interactions when they are not present or removing interactions when they are present. The aligned rank transform is a robust and powerful technique. This method is not sensitive to outliers and violations of assumptions on error distribu-tion. Conclusion: It concluded that the aligned rank transform pro-cedures appear to be the best known alternatives to the classical analysis of variance procedures, and should be used in situations where the assumptions of normality is suspect.
Keywords: Statistical analysis; nonparametric interaction tests; aligned rank test; factorial analysis
- Leys C, Schumann S. A nonparametric method to analyze interactions: The adjusted rank transform test. Journal of Experimental Social Psychology. 2010;46(4):684-8. [Crossref]
- Durner E. Effective analysis of interactive effects with non-normal data using the aligned rank transform, ARTool and SAS® University edition. Horticulturae. 2019;5(3):57; doi: 10.3390/ horticulturae5030057. [Crossref]
- Higgins JJ, Tashtoush S. An aligned rank transform test for interaction. Nonlinear World, 1994; 1 (2): 201-11. [Link]
- Saste SV, Sananse SL, Sonar CD. On parametric and nonparametric analysis of two factor factorial experiment. Int J Appl Res. 2016;2(7):653-6. [Link]
- Higgins JJ, Blair RC, Tashtoush S. The aligned rank transform procedure. 2nd Annual Conference Proceedings. Conference on Applied Statistics in Agriculture. New Prairie Press; 1990. p.185-95 [Crossref]
- Beasley TM, Zumbo BD. Aligned rank tests for interactions in split-plot designs: distributional assumptions and stochastic heterogeneity. Journal of Modern Applied Statistical Methods. 2009;8(1):16-50. [Crossref]
- Salter KC, Fawcett RF. The art test of interaction: a robust and powerful rank test of interaction in factorial models. Communications in Statistics-Simulation and Computation. 1993;22(1):137-53. [Crossref]
- Beasley TM. Multivariate aligned rank test for interactions in multiple group repeated measures designs. Multivariate Behavioral Research. 2002;37(2):197-226. [Crossref]
- Peterson K. Six modifications of the aligned rank transform test for interaction. Journal of Modern Applied Statistical Methods. 2002;1(1):100-9. [Crossref]
- Lei X, Holt JK, Beasley TM. Aligned rank tests as robust alternatives for testing interactions in multiple group repeated measures designs with heterogeneous covariances. Journal of Modern Applied Statistical Methods. 2004;3(2):462-75. [Crossref]
- Richter SJ, Payton ME. An improvement to the aligned rank statistic for two factor analysis of variance. J Appl Stat Sci. 2005;14(3/4):225-35. [Link]
- Mansouri H. Multifactor analysis of variance based on the aligned rank transform technique. Computational Statistics & Data Analysis. 1999;29(2):177-89. [Crossref]
- Mansouri H, Paige RL, Surles JG. Aligned rank transform techniques for analysis of variance and multiple comparisons. Communications in Statistics-Theory and Methods. 2004;33(9):2217-32. [Crossref]
- Sawilowsky SS. Nonparametric tests of interaction in experimental design. Review of Educational Research. 1990;60(1):91-126. [Crossref]
- Feys J. New nonparametric rank tests for interactions in factorial designs with repeated measures. Journal of Modern Applied Statistical Methods. 2016;15(1): 78-99. [Crossref]
- Oliver-Rodríguez JC, Wang XT. Non-parametric three-way mixed ANOVA with aligned rank tests. Br J Math Stat Psychol. 2015;68(1):23-42. [Crossref] [PubMed]
- Blair RC, Sawilowsky SS, Higgins JJ. Limitations of the rank transform statistic in tests for interactions. Communications in Statistics-Simulation and Computation. 1987;16(4):1133-45. [Crossref]
- Richter SJ, Payton ME. Nearly exact tests in factorial experiments using the aligned rank transform. Journal of Applied Statistics. 1999;26(2):203-17. [Crossref]
- Bryan JJ. Rank transforms and tests of interaction for for repeated measures experiments with various covariance structures. PhD Thesis. Stillwater: Oklahoma State University; 2009.
- Akritas MG, Arnold SF, Brunner E. Nonparametric hypotheses and rank statistics for unbalanced factorial designs. Journal of the American Statistical Association. 1997;92(437):258-65. [Crossref]
- Salazar-Alvarez MI, Tercero Gómez VG, Temblador MDC, Cordero Franco AE, Conover W. Nonparametric analysis of interactions: A review and gap analysis. In: Guan Y, Liao H, eds. Proceedings of Proceedings of the 2014 Industrial and Systems Engineering Research Conference. Vol. 1. Montreal, Canada: Curran Associates, Inc; 2014. p.2910-7.
- Wobbrock JO, Findlater L, Gergle D, Higgins JJ. The aligned rank transform for nonparametric factorial analyses using only ANOVA procedures. In: Proceeding of ACM CHI 2011. Conference on Human Factors in Computing Systems. 2011. p.143-6. [Crossref]
- Luepsen H. The aligned rank transform and discrete variables: A warning. Communications in Statistics-Simulation and Computation. 2017;46(9):6923-36. [Crossref]
- Sawilowsky SS, Blair RC, Higgins JJ. An investigation of the type I error and power properties of the rank transform procedure in factorial ANOVA. Journal of Educational Statistics. 1989;14(3):255-67. [Crossref]
- Luepsen H. Comparison of nonparametric analysis of variance methods: a vote for van der Waerden. Communications in Statistics-Simulation and Computation. 2018;47(9):2547-76. [Crossref]
- Mangiafico SS. How should we analyze Likert item data? Journal of the NACAA. 2019;12(2): 1-8. [Link]
.: Process List