Example: piglet diet experiment with three littersSST is the “sum of squares treatments”SSB is the “sum of squares blocks”SSE is the “sum of squares error”.F*=MST/MSEANOVA (III) 1 Randomized Complete Block Designs (RCBD) Defn: A Randomized Complete Block Design is a variant of the completely randomized design that we recently learned. In this design, blocks of experimental units are chosen where the units within are block are more similar to each other (homogeneous) than to units in other blocks. In a complete block design, there are at least t experimental units in each block. Examples of blocks: 1) a litter of animals could be considered a block since they all have similar genetic structure, similar prenatal/parental care, etc. 2) a field or pasture that can be divided into quadrants since soil properties, environmental conditions, etc are similar within a field 3) a greenhouse with multiple benches since environmental conditions are usually more similar within a greenhouse than between greenhouses 4) a year in which the experiment is performed since environmental conditions are similar within a year Example of a CRBD: A nutritionist is interested in comparing the effect of three diets on weight gain in piglets. In order to perform the experiment, the researcher chooses 10 litters, each with at least three healthy and similarly sized piglets that have just been weaned. In each litter, three piglets are selected and one treatment is randomly assigned to each piglet. Diets are labeled A, B or C.ANOVA (III) 2 Litter Piglet 1 2 3 1 A C B 2 B C A … 10 C B A In a design without blocking, the researcher would pick 30 piglets from different litters and randomly assign treatments to them. This is known as unrestricted randomization. Blocking designs have restricted randomization since the treatments are randomly assigned WITHIN each block. An RCBD has two factors: the factor of interest that includes the treatments to be studied and the “Blocking Factor” that identifies the blocks used in the experiment. There are several forms of Blocking Designs: 1) the RCBD that we will study 2) incomplete block designs in which not every block has t experimental units 3) block designs in which the blocks have more than t experimental units that are used in the experiment 4) Latin square designs which have very specific forms of randomization of treatments within blocks (example is usually relates to time ordering of treatments)ANOVA (III) 3 Assumptions of the RCBD: 1) Sampling: a. The blocks are independently sampled b. The treatments are randomly assigned to the experimental units within a block. 2) Homogeneous Variance: the treatments all have the same variability, i.e. they all have the same variance 3) Approximate Normality: each population is normally distributed Hypotheses As we will see, the blocking factor is included in the study only as a way of explaining some of the variation in responses (Y) of the experimental units. As such, we are not interested in testing hypotheses about the blocking factor. Instead, just like in a one-way ANOVA, we restrict our attention to the other factor (“research” factor). So, hypothesis testing proceeds similar to the techniques we learned for the one-way ANOVA. The two differences are the calculation of the error variance (MSE) and a calculation of the effect of the blocking factor (MSB).ANOVA (III) 4 Notation t the number of treatments of interest in the “research” factor b the number of blocks containing t experimental units N = t × b, the total sample size yij observed value for the experimental unit in the jth block assigned to the ith treatment, j = 1,2,…,b and i = 1,2,…,t •iy bybjij∑==1, the sample mean of the ith treatment jy• tytiij∑==1, the sample mean of the jth block ••y tbytibjij∑∑===11, the overall sample mean of the combined treatmentsANOVA (III) 5 Example: piglet diet experiment with three litters Diet Block Litter A B C Mean 1 yA1 = 54.3 yB1 = 53.1 yC1 = 59.7 7.551=•y2 yA2 = 53.6 yB2 = 52.4 yC2 = 59.7 2.552=•y3 yA3 = 55.2 yB3 = 57.1 yC3 = 67.2 2.623=•y Treatment Mean 4.54=•Ay 2.55=•By 8.59=•Cy Grand Mean 9.56=••y Model: ijjiijYεβαµ+++= where • µ is the overall (grand) mean, • αi is the effect due to the ith treatment, • βj is the effect due to the jth block, and, • εij is the error term where the error terms, are independent observations from an approximately Normal distribution with mean = 0 and constant variance 2εσANOVA (III) 6 Total variability of all of the Yij, is TSS ∑∑••−=ijijyy2)( which can be broken up into three parts: TSS = SST + SSB + SSE SST∑∑=−=•••iiiibyyb22ˆ)(α is the “sum of squares treatments” SSB∑∑=−=•••jjjjtyyt22ˆ)(β is the “sum of squares blocks” SSE∑∑∑∑=+−−=••••ijijijjiijyyyy22ˆ)(ε is the “sum of squares error”. Like before, we are interested in the Mean Squares: MST 1−=tSST, the Mean Square Treatments MSB 1−=bSSB, the Mean Square Blocks MSE )1)(1( −−=btSSE, the Mean Square Error Here ∑−+=iitbMSTE)1()(22ασε and 2)(εσ=MSEE.ANOVA (III) 7 ANOVA Table for a Randomized Complete Block Design Source Sum of Squares Degrees of Freedom Mean Square F-stat Treatment SST t – 1 MST F*=MST/MSEBlock SSB b – 1 MSB Error SSE (t – 1)(b – 1) MSE Total TSS tb – 1 Again, the test of a treatment effect H0: µ1 = µ2 =…= µt HA: at least one mean differs uses the statistic F*=MST/MSE. If the null hypothesis is true then F* has an F-Distribution on numerator degrees of freedom t – 1 and denominator degrees of freedom (t – 1)(b – 1). In addition to the similarity of the F-test of equality of treatment means, the tests and comparisons of treatment means are done exactly the same as before as well.ANOVA (III) 8 Example: piglet experiment. data pigsblocked; input litter diet$ gain @@; datalines; 1 I 54.3 2 I 53.6 3 I 55.2 1 II 53.1 2 II 52.4 3 II 57.1 1 III 59.7 2 III 59.7 3 III 67.2 run; proc glm data=pigsblocked; class diet litter; model gain = diet litter; quit; Dependent Variable: gain Sum of Source DF Squares Mean Square F Value Pr > F diet 2 125.39 62.69 19.02 0.0091 litter 2 38.46 19.23 5.83
View Full Document
Unlocking...