Multivariate analysis of variance followed by descriptive discriminant analysis: An analysis of the acoustic treatments effect on mung beans’ growth

Abstract

In this paper, experimental data of acoustic treatments effect on the growth of mung beans based on three response variables (length of stem, number of leaves, length of the longest root) are analyzed using multivariate analysis of variance (MANOVA) followed by descriptive discriminant analysis (DDA). Musical acoustics has been discovered to be significant in stimulating various plant growth. Most related studies assessed multiple response variables of plant growth measurements. Thus, multiple analysis of variance (ANOVA) has been applied extensively in finding the significant effect. However, ANOVA does not consider correlation among the response variables. Therefore, MANOVA is proposed as the alternative of multiple ANOVA since the method considers correlation among response variables. This will control the likelihood of committing Type I error. Then, DDA is performed as multivariate post hoc analysis instead of the usual univariate comparisons. From the analysis, there is a significant effect of acoustic treatments (p<0.05) on the linear combination of response variables. DDA results indicated that length of stem is clearly the most important variables in distinguishing acoustic treatment groups and soprano group is well separated which the treatment helps stimulating length of stem. In conclusion, MANOVA with DDA is a perfect combination of analyses methods in assessing the effect of an independent variable on correlated response variables.