One Missing Observation in Graeco Latin Square Design: An Approximate Analysis of Variance
Author(s)
Kupolusi, Joseph A. , Ojo, Oluwadare O. ,
Download Full PDF Pages: 01-08 | Views: 760 | Downloads: 168 | DOI:
Abstract
Background: Experimental results can seriously be affected by different degrees of variation that arise from unknown or uncontrollable design factors which is not of interested to the experimenter that may probably has effect on the response. Blocking is an extremely important design technique that can be used to systematically eliminate the effect of the uncontrollable design factors. Graeco-Latin Square design is used to eliminate three sources of variability; that is, it systematically allows blocking in three directions. Thus, rows, columns and Greek letters actually represent three restrictions on randomization. There are situations whereby one observation is occasionally missed in a Graeco Latin Square design of order P x P. In this paper, we proposed an approximate method for one missing observation in Graeco Latin Square Design of any order.
Results: The proposed approximate method was applied to a data set of Graeco Latin Square Design of order 4 and the results are presented in ANOVA tables. Based on this result, Mean Sum of Squares (MSE) reduced drastically compared to that of complete data set. Reduction in MSE is a clear indication that the proposed method can be used to obtain a better result with a minimum variance and unbiased estimate.
Conclusion: The result of the analysis indicated that the proposed approximate method for Graeco Latin Square is appropriate for estimation of missing observation through a simulation study of 1000 experimental runs. The result converged to the real data set. Hence, the method derived in this research is capable to handling the problem of missing observation in Graeco-Latin Square design.
Keywords
Latin Square, Graeco Latin Square, ANOVA, missing data, experimental error
References
i Ai, M., Li, K., Liu, S., & Lin, D. K. (2013). Balanced incomplete Latin square designs. Journal of Statistical Planning and Inference, 143(9), 1575–1582. https://doi.org/10.1016/j.jspi.2013.05.001
ii Allan, F. E., & Wishart, J. (1930). A method of estimating the yield of a missing plot in field experimental work. The Journal of Agricultural Science, 20(3), 399–406. https://doi.org/10.1017/S0021859600006912
iii Bartlett, M. S. (1937). Some examples of statistical methods of research in agriculture and applied biology. Supplement to the Journal of the Royal Statistical Society, 4, 137–183. https://doi.org/10.2307/2983644
iv Casselgren, C. J., & Häggkvist, R. (2013). Completing partial Latin squares with one filled row. Column and Symbol. Discrete Mathematics, 313, 1011–1017. https://doi.org/10.1016/j.disc.2013.01.019
v Cochran, W. G. (1957). Analysis of covariance: Its nature and uses. Biometrics, 13(3), 261–281. https://doi.org/10.2307/2527916
vi Coons, I. (1957). The analysis of covariance as a missing plot technique. Biometrics, 13(3), 387–405. https://doi.org/10.2307/2527922
vii Dempster, A. P., Laird, N. M., & Rubin, D. B. (1977). Maximum likelihood from incomplete data via the EM algorithm. Journal of the Royal Statistical Society Series B (Methodological), 39, 1–38.
viii Donovan, D. M., & Şule Yazıcı, E. (2014). A polynomial embedding of pairs of orthogonal partial Latin squares. Journal of Combinatorial Theory, Series A, 126, 24–34. https://doi.org/10.1016/j.jcta.2014.04.003
ix Euler, R. (2010). On the completability of incomplete Latin squares. European Journal of Combinatorics, 31, 535–552. https://doi.org/10.1016/j.ejc.2009.03.036
x Healy, M., & Westmacott, M. (1956). Missing values in experiments analysed on automatic computers. Applied Statistics, 5, 203–206. https://doi.org/10.2307/2985421
xi Kramer, C. Y., & Glass, S. (1960). Analysis of variance of a Latin square design with missing observations. Applied Statistics, 9, 43–50. https://doi.org/10.2307/2985758
xii Kittiwat, S. & Kanogkan, L. | (2017) One missing value problem in Latin square design of any order: Exact analysis of variance, Cogent Engineering, 4:1, 1411222
xiii Little, R. J. A., & Rubin, D. B. (2002). Statistical analysis with missing data (2nd ed.). Hoboken: John Wiley & Sons. https://doi.org/10.1002/9781119013563
xiv Mead, R., Gilmour, S. G., & Mead, A. (2012). Statistical principles for the design of experiments: Applications to real experiments, Vol. 36. Cambridge: Cambridge University Press. https://doi.org/10.1017/CBO9781139020879
xv Montgomery, D. C. (2008). Design and analysis of experiments (7th ed.). New York, NY: John Wiley & Sons.
xvi Ott, R. L., & Longnecker, M. (2010). An introduction to statistical methods and data analysis (6th ed.). Boston: Cengage Learning.
xvii Rubin, D. B. (1972). A non-iterative algorithm for least squares estimation of missing values in any analysis of variance design. Applied Statistics, 136–141. https://doi.org/10.2307/2346485
xviii Rubin, D. B. (1987). Multiple imputation for non-response in surveys. Hoboken: John Wiley & Sons. https://doi. org/10.1002/SERIES1345
xix Sirikasemsuk, K. (2016a). A review on incomplete latin square design of any order. In N. Rusli, W. M. Zaimi, K. A. Khazali, M. J. Masnan, W. S. Daud, N. Abdullah, …, Y. N. Yusuf (Eds.),
xx AIP Conference Proceedings 2016, (vol. 1775, pp. 030022). Melville: AIP Publishing.
xxi Sirikasemsuk, K. (2016b). One missing value problem in Latin square design of any order: Regression sum of squares. In 2016 Joint 8th International Conference on Soft computing and Intelligent Systems (SCIS) and 17th International Symposium on Advanced Intelligent Systems (pp. 142–147). Japan: IEEE Press.
xxii Speed, T. P. (1992). Introduction to fisher (1926) the arrangement of field experiments. In S. Kotz & N. L. Johnson (Eds.), Breakthroughs in statistics, Vol. 2 (pp. 71–81). New York, NY: Springer. https://doi. org/10.1007/978-1-4612-4380-9
xxiii Wilkinson, G. N. (1958). Estimation of missing values for the analysis of incomplete data. Biometrics, 14(2), 257–286. https://doi.org/10.2307/2527789
xxiv Yates, F. (1933). The analysis of replicated experiments when the field results are incomplete. Empire Journal of, Experimental Agriculture, 2(1), 129–142.
xxv Yates, F. (1936). Incomplete randomized blocks. Annals of eugenics, 7, 121–140. https://doi. org/10.1111/j.1469-1809.1936.tb02134.x
xxvi Youden, W. J. (1937). Use of incomplete block replications in estimating tobacco-mosaic virus. Contributions from Boyce Thompson Institute, 9(1), 41–48.
xxvii Zhang, H. (2013). 25 new r -self-orthogonal Latin squares. Discrete Mathematics, 313, 1746–1753. https://doi.org/10.1016/j.disc.2013.04.021
Cite this Article: