Sidebar Menu

Experimental Design

A collection of articles on the Definition and Basic Concepts of Experimental Design, Assumptions of Analysis of Variance, Comparison of Means, and several types of Experimental Designs that are often used: Completely Randomized Design, Randomized Block Design, Latin Square Design (RBSL) Factorial Experiment, Split- Plot, Split Split Plot Design, Strip Plot Design

Experimental units in a Completely Randomized Design (CRD) are always assumed to be homogeneous . In reality, this is not necessarily true, so other methods are needed that can describe this diversity. If we conduct an experiment on a plot of land that has different fertility levels, then the effect of the treatment that we ascribe to the treatment we are trying may not be correct, thus creating a Type I Error. this land in the RAL will be included in the JKG ( Within) so that the KTG will get bigger and F (KTP/KTG) will get smaller, as a result the experiment is no longer sensitive. Finally, if we repeat the treatment at locations that have different ( non-homogeneous ) variations, then the additional diversity needs to be removed from the analysis so that we focus more on the diversity caused by the treatment we are trying. If the group factor is included in the design, we can capture the variation it causes into the JK Block. This process will reduce SS Within (Error), compared to Completely Randomized Design.

In the previous discussion we discussed the effect of a single treatment on certain responses . The single treatment is called the factor , and the level or level of the factor is called the level .  The factor is symbolized by a capital letter while the level of the factor is symbolized by a lowercase letter . If we simultaneously observe the effect of several factors in the same study, the experiment is called a factorial experiment .

A factorial experiment is an experiment in which the treatment consists of all possible combinations of levels of several factors. Experiments using f factors with t levels for each factor are symbolized by the factorial experiment f t . For example, a 2 2 factorial experiment means that we use 2 factors and the level of each factor consists of 2 levels. Factorial experiment 2 2It is also often written in the form of a 2x2 factorial experiment. The last symbol is often used for factorial experiments where the level of each factor is different, for example 2 levels for factor A and 3 levels for factor B, the experiment is called a 2x3 factorial experiment. A 2x2x3 factorial experiment means a factorial experiment consisting of 3 factors with levels for each factor of 2, 2, and 3. described previously, and the next stage is the selection of the environmental design, which involves the form of experimental designs such as:

Assumptions for hypothesis testing based on the single factor ANOVA model are actually related to the residual value or error (ε ij ). Many references state that single factor ANOVA is quite reliable for this assumption, for example the F test remains reliable and reliable even though the assumptions are not met. However, the level of reliability is very difficult to measure and also depends on the sample size that must be balanced. The F test can be very unreliable if the sample size is not balanced, especially if it is added to the distribution of data that is not normal and the variance is not homogeneous. Therefore, I highly recommend checking the ANOVA assumptions first before proceeding to the analysis phase.

This article is a continuation of the Completely Randomized Design (CRD) Factorial. Suppose there is an experiment as follows:
There are 3 types of materials for the manufacture of batteries (A, B, C) which were tested at 3 temperatures (15 o F, 70 o F, 125 o F). From this experiment, we want to know whether the type of material and temperature affect battery life? Are certain types of materials suitable for certain temperatures? The following are the steps for calculating the analysis of variance followed by Duncan Multiple Range - Post Hoc Test.

4.1. Introduction

Randomized Complete Block Design (RCBD) is generally used to improve the ability to detect true differences between the treatments we are trying by eliminating the influence of other known variations (groups) of experimental errors. If the idea is applied to eliminate two sources of diversity by grouping in two directions, then the design is called the Latin Square Design (LSD). Thus, LSD is an experimental design with two grouping directions , namely row and column. The number of treatments is equal to the number of replications so that each row and column will contain all treatments. In this design, randomization is limited by grouping it into rows and columns, so that each row and column will only get one treatment.

Factorial experiment with a basic design of Randomized Block Design (RCBD) is an experiment in which more than one factor is tested and using RCBD as the experimental design. This design was chosen if the experimental unit used was not uniform, so it was necessary to group it, while in Completely Randomized Design (CRD) Factorial , the experimental unit was relatively uniform so there was no need for grouping. Basically, the RCBD Factorial experiment is the same as the Randomized Complete Block Design (RCBD) experiment previously discussed, but in this experiment it consists of two or more factors.