Symmetric and Unsymmetric Balanced Incomplete Block Designs: A Comparative Analysis

This paperdiscusses a comparat ive analysis on balanced incomplete block designs by using the classical analysis of variance (ANOVA) method. Fortunately, the data co llected for the analysis were in two groups of the balanced incomplete-block designs (BIBD’s), that is, symmetric, and unsymmetric (BIBD’s). In this paper, the basic interest is to apply classical ANOVA on the two types of BIBD’s and check whether they are significant and also minimizes error. A secondary data from N.R.C.R.I, Umudike, Abia State was used. To ach ieve this, we shall consider treatment (adjusted), b lock (adjusted) treatment (not adjusted) in the classical ANOVA method on the available data. Though, symmetric balanced incomplete block design (SBIBD) and unsymmetric balanced incomplete block design (USBIBD) are significant, it is pertinent to note that the SBIBD classical ANOVA method is found to be preferable to the USBIBD with reference to their variances at different level of significance.


Introduction
The basic concepts of the statistical design of experiments and data analysis were discovered in the early part of the 20 th century as a cost effective research design tool to help improve y ields in farming. Since then, many types of designs experiments and analysis techniques have been developed to meet the diverse needs of researchers and engineers. One of such experimental designs is called the balance inco mplete block design (BIBD). Even though, there are missing data in BIBD, it also possesses fine qualities, like having the ability to handle local control or minimize error. BIBDs can be used to eliminate block differences giving equal accuracy for comparing treatments. Also during the analysis of BIBD they are relatively less complicated than unbalanced designs.
To the best of my knowledge complete b lock design is widely used by researchers more than the balanced incomp lete block designs because the missing data are computed before analysis. It is pertinent to note that, Classical ANOVA method can be used to analyzed BIBD without computing the missing data see, for example [1], [2], [3]. In this paper, the basic interest is on balanced incomp lete design (BIBD's). Here, the Classical analysis-of-variance method was applied on SBIBD and USBIBD to check if they are significant. In addition, to know which of the BIBD's with minimal error.
The objective of this study is to check if there is any significance of the relationship between the variables at 0.01 and 0.05 levels with reference to minimu m variance on symmetric balanced incomp lete block design (SBIBD) and unsymmetric balanced inco mplete b lock design (USBIBD).

Literature
A block design is a set together with a family of subsets (repeated subsets are allo wed at times) whose members are chosen to satisfy some set of properties that are deemed useful for a particular application. These applications come fro m many areas, including experimental design, fin ite geometry, software testing, cryptography, and algebraic geometry. Many variations have been examined, but the most intensely studied are the balanced incomplete block designs (BIBDs or 2-designs) which historically were related to statistical issues in the design of experiments, see, hppt://en.wikipedia.org/wiki/ Block_design [4], [5], [6].
A complete-b lock design is one of the most widely used designs. It is used when it is possible to run all the treat ment combinations in each book. In situations where it is not possible to run all the treatment co mbinations in each book due to shortage of experimental apparatus or facilities or the physical size of the block, an inco mplete-b lock design is used: see, fo r examp le, [7], [2]. A balanced incomp lete block design in which the number b of b locks equals the number v of elements arranged among the blocks is said to be symmetric, see www.answer.co m/topic / symmetric -designmathematics.
Some o f the leading studies on balance incomplete block designs are [7], [8], [9], [10], [11], [12], [13] and [14]. However, this paper briefly d iscusses the balanced incomplete-block designs (BIBD's) and further analysis symmetric balanced incomp lete block design (SBIBD) and unsymmetric balanced incomplete block design (USBIBD) with respect to their means. The Classical analysis-of-variance methods a statistical method for making simu ltaneous comparisons between two or more means. These methods were differently applied to data collected and were represented in various analysis-of-variance tables. The case study data emp loyed here were obtained fro m the Nat ional Root Crops Research Institute (NRCRI) Umudike, Abia State, Nigeria. The data are on varieties of cassava with different ratios of Nitrogen, Phosphorous and potassium (NPK). The data are secondary data, whose entries are abinito arranged as BIBD by the experimenter.

Symmetric Bal anced Incomplete-bl ock Design
Here, we assume that there are t treatments and b blocks. In addition, we assume each block contains k treatment each of which occurs r times in the design (or is replicated r times), and that there are N = tr = b k total observations; see, for example [9]. Furthermore, the number of t imes each pair of treatments appears in the same block is if t = b, the design is said to be symmetric. The parameter must be an integer. The symmetry property of a design occur when the row effect and the column effects of the design are interchanged and still the error sum of squares so calculated in each remains the same, see, [2].

Unsymmetric Bal anced Incomplete -bl ock Design
we assume that there are t treat ments and b blocks. The only difference between SBIBD and USBIBD is that t ≠ b, the design is said to be unsymmet ric.
a.The Classical Model: So me existing data sets are evaluated to see if they satisfy the assumption for application of chosen models. There are three conceptual classes of such models in designs: •Fixed-effects models •Random -effects models •Mixed-effects models In this work, fixed effects model is employed. The fixed effect model analysis of variance applies to situations in which the experimenter has subjected his experimental material to several treatments, each of which affects only the mean of the underly ing normal distribution of the "response variable", see, [15].
b. The classical ANOVA: In statistics, analysis of variance is a collection of statistical models and their associated procedures which compare means by splitting the overall observed variance into different parts see. [8].The analysis of variance for classical ANOVA is summarized in Table 1, Table 2 and Table 3.

Expe riment A
A balanced incomplete-block design (BIBD) that is symmetric was used to study the yield on four variet ies of cassava with four d ifferent rates of NPK. These rates were administered in addition to the natural manure. The data collected fro m the experiment are tabulated in the layout of table 4.  Since Fcal ≥Ftab at different levels of probability, we conclude that the fertilizer applied has a significant effect on the cassava. The analysis of variance in Table 8 shows that at different levels of probability, we conclude that the fertilizer applied has a significant effect on the cassava. Experi ment B Consider example 2, an unsymmetric balanced incomp lete block design (UBIBD) was used to study the yield in nine varieties of cassava with three different rates of NPK. These rates were administered in addition to the natural manure. The data collected from the experiment are tabulated in the layout of table 9. A C Notice that each pair of treatment in Table 9 (AB,AC.. ) occurs together twice 9in exactly 2 blocks) The analysis of variance when treat ment is adjusted for Experiment 2 is show in Tab le 11. Since at different levels of probability, we conclude that the fertilizer applied has a significant effect on the cassava.

Result Presentation
The symmetric and unsymmetric balanced incomp lete block designs (Experiment A&B) are summarized in the table 14&15 below.

Summary and Conclusions
Considering the results obtained from the experiment used in this research work, different observation were made. In experiment 1, that is, a symmetric balanced incomp lete block design (SBIBD), the classical ANOVA method, when treatments are not adjusted. Also when the treatments are adjusted using classical ANOVA method was significant at 0.05 and 0.01 levels of probability. However, the classical and vector space ANOVA methods were considered and found to be significant at 0.05 and 0.01 levels of probability. In summary the classical ANOVA was all significant at 0.05 and 0.01 levels of probability. On the basis of comparison, using min imization of error in BIBD, we have that SBIBD leads to a minimu m error than USBIBD whether the treatment is adjusted or not. Also it is better to adjust the treatments in SBIBD and USBIBD because the adjustment leads to minimu m erro r.
To have an effective comparison of the two methods, the following observations were made: (ⅰ) Based on minimization of error in BIBD, we have that SBIBD leads to a minimu m error than USBIBD whether the treatment is adjusted or not.
(ⅱ) When treatments are adjusted or not adjusted, the classical ANOVA method was applied on SBIBD and USBIBD; both yielded the same result, that is, both are significant at 0.05 and 0.01 levels of p robability.
(ⅲ) The classical ANOVA method on SBIBD seems easier and more convenient to handle than the classical ANOVA method on USBIBD.
(ⅳ) Block is being adjusted to know if the treat ment is SBIBD or USBIBD by checking whether the erro r su m of squares for the treat ments when adjusted equals the error sum of squares for the blocks when adjusted.