1. Factorial Experiments Analysis of Variance Experimental Design

2. Dependent variable Y k Categorical independent variables A, B, C, … (the Factors) Let –a = the number of categories of A –b = the number of categories of B –c = the number of categories of C –etc.

3. The Completely Randomized Design We form the set of all treatment combinations – the set of all combinations of the k factors Total number of treatment combinations –t = abc…. In the completely randomized design n experimental units (test animals, test plots, etc. are randomly assigned to each treatment combination. –Total number of experimental units N = nt=nabc..

4. The treatment combinations can thought to be arranged in a k-dimensional rectangular block A 1 2 a B 12b

5. A B C

6. Another way of representing the treatment combinations in a factorial experiment A B... D C

7. Example In this example we are examining the effect of We have n = 10 test animals randomly assigned to k = 6 diets The level of protein A (High or Low) and The source of protein B (Beef, Cereal, or Pork) on weight gains Y (grams) in rats.

8. The k = 6 diets are the 6 = 3×2 Level-Source combinations 1.High - Beef 2.High - Cereal 3.High - Pork 4.Low - Beef 5.Low - Cereal 6.Low - Pork

9. Table Gains in weight (grams) for rats under six diets differing in level of protein (High or Low) and s ource of protein (Beef, Cereal, or Pork) Level of ProteinHigh ProteinLow protein Source of ProteinBeefCerealPorkBeefCerealPork Diet123456 7398949010749 1027479769582 1185696909773 10411198648086 8195102869881 10788102517497 100821087274106 877791906770 11786120958961 11192105785882 Mean100.085.999.579.283.978.7 Std. Dev.15.1415.0210.9213.8915.7116.55

10. Example – Four factor experiment Four factors are studied for their effect on Y (luster of paint film). The four factors are: Two observations of film luster (Y) are taken for each treatment combination 1) Film Thickness - (1 or 2 mils) 2)Drying conditions (Regular or Special) 3)Length of wash (10,30,40 or 60 Minutes), and 4)Temperature of wash (92 ˚C or 100 ˚C)

11. The data is tabulated below: Regular DrySpecial Dry Minutes92  C100  C92  C100  C 1-mil Thickness 203.43.419.614.52.13.817.213.4 304.14.117.517.04.04.613.514.3 404.94.217.615.25.13.316.017.8 605.04.920.917.18.34.317.513.9 2-mil Thickness 205.53.726.629.54.54.525.622.5 305.76.131.630.25.95.929.229.8 405.55.630.530.25.55.832.627.4 607.26.031.429.68.09.933.529.5

12. Notation Let the single observations be denoted by a single letter and a number of subscripts y ijk…..l The number of subscripts is equal to: (the number of factors) + 1 1 st subscript = level of first factor 2 nd subscript = level of 2 nd factor … Last subsrcript denotes different observations on the same treatment combination

13. Notation for Means When averaging over one or several subscripts we put a “bar” above the letter and replace the subscripts by Example: y 241

14. Profile of a Factor Plot of observations means vs. levels of the factor. The levels of the other factors may be held constant or we may average over the other levels

15. Definition: A factor is said to not affect the response if the profile of the factor is horizontal for all combinations of levels of the other factors: No change in the response when you change the levels of the factor (true for all combinations of levels of the other factors) Otherwise the factor is said to affect the response:

16. Definition: Two (or more) factors are said to interact if changes in the response when you change the level of one factor depend on the level(s) of the other factor(s). Profiles of the factor for different levels of the other factor(s) are not parallel Otherwise the factors are said to be additive. Profiles of the factor for different levels of the other factor(s) are parallel.

17. If two (or more) factors interact each factor effects the response. If two (or more) factors are additive it still remains to be determined if the factors affect the response In factorial experiments we are interested in determining –which factors effect the response and – which groups of factors interact.

18. Factor A has no effect A B

19. Additive Factors A B

20. Interacting Factors A B

21. The testing in factorial experiments 1.Test first the higher order interactions. 2.If an interaction is present there is no need to test lower order interactions or main effects involving those factors. All factors in the interaction affect the response and they interact 3.The testing continues with for lower order interactions and main effects for factors which have not yet been determined to affect the response.

22. Level of ProteinBeefCerealPorkOverall Low79.2083.9078.7080.60 Source of Protein High100.0085.9099.5095.13 Overall89.6084.9089.1087.87 Example: Diet Example Summary Table of Cell means

23. Profiles of Weight Gain for Source and Level of Protein

25. Models for factorial Experiments Single Factor: A – a levels y ij =  +  i +  ij i = 1,2,...,a; j = 1,2,...,n Random error – Normal, mean 0, std-dev. Overall meanEffect on y of factor A when A = i

26. y 11 y 12 y 13 y 1n y 21 y 22 y 23 y 2n y 31 y 32 y 33 y 3n y a1 y a2 y a3 y an Levels of A 123 a observations Normal dist’n Mean of observations 11 22 33 aa  +  1  +  2  +  3  +  a Definitions

27. Two Factor: A (a levels), B (b levels y ijk =  +  i +  j + (  ) ij +  ijk i = 1,2,...,a ; j = 1,2,...,b ; k = 1,2,...,n Overall mean Main effect of AMain effect of B Interaction effect of A and B

28. Table of Means

29. Table of Effects – Overall mean, Main effects, Interaction Effects

30. Three Factor: A (a levels), B (b levels), C (c levels) y ijkl =  +  i +  j +  ij +  k + (  ) ik + (  ) jk +  ijk +  ijkl =  +  i +  j +  k +  ij + (  ik + (  jk +  ijk +  ijkl i = 1,2,...,a ; j = 1,2,...,b ; k = 1,2,...,c; l = 1,2,...,n Main effects Two factor Interacti ons Three factor Interaction Random error

31.  ijk = the mean of y when A = i, B = j, C = k =  +  i +  j +  k +  ij + (  ik + (  jk +  ijk i = 1,2,...,a ; j = 1,2,...,b ; k = 1,2,...,c; l = 1,2,...,n Main effects Two factor Interactions Three factor Interaction Overall mean

32. Levels of C Leve ls of B Levels of A Leve ls of B Levels of A No interaction

33. Levels of C Leve ls of B Levels of A A, B interact, No interaction with C Leve ls of B

34. Levels of C Leve ls of B Levels of A A, B, C interact Leve ls of B

35. Four Factor: y ijklm =  +   +  j + (  ) ij +  k + (  ) ik + (  ) jk + (  ) ijk +  l + (  ) il + (  ) jl + (  ) ijl + (  ) kl + (  ) ikl + (  ) jkl + (  ) ijkl +  ijklm =  +  i +  j +  k +  l + (  ) ij + (  ) ik + (  ) jk + (  ) il + (  ) jl + (  ) kl +(  ) ijk + (  ) ijl + (  ) ikl + (  ) jkl + (  ) ijkl +  ijklm i = 1,2,...,a ; j = 1,2,...,b ; k = 1,2,...,c; l = 1,2,...,d; m = 1,2,...,n where0 =  i =  j =  (  ) ij   k =  (  ) ik =  (  ) jk =  (  ) ijk =   l =  (  ) il =  (  ) jl =  (  ) ijl =  (  ) kl =  (  ) ikl =  (  ) jkl =  (  ) ijkl and  denotes the summation over any of the subscripts. Main effects Two factor Interactions Three factor Interactions Overall mean Four factor InteractionRandom error

36. Estimation of Main Effects and Interactions Estimator of Main effect of a Factor Estimator of k-factor interaction effect at a combination of levels of the k factors = Mean at the combination of levels of the k factors - sum of all means at k-1 combinations of levels of the k factors +sum of all means at k-2 combinations of levels of the k factors - etc. =Mean at level i of the factor - Overall Mean

37. Example: The main effect of factor B at level j in a four factor (A,B,C and D) experiment is estimated by: The two-factor interaction effect between factors B and C when B is at level j and C is at level k is estimated by:

38. The three-factor interaction effect between factors B, C and D when B is at level j, C is at level k and D is at level l is estimated by: Finally the four-factor interaction effect between factors A,B, C and when A is at level i, B is at level j, C is at level k and D is at level l is estimated by:

39. Anova Table entries Sum of squares interaction (or main) effects being tested = (product of sample size and levels of factors not included in the interaction) × (Sum of squares of effects being tested) Degrees of freedom = df = product of (number of levels - 1) of factors included in the interaction.

40. Analysis of Variance (ANOVA) Table Entries (Two factors – A and B)

41. The ANOVA Table

42. Analysis of Variance (ANOVA) Table Entries (Three factors – A, B and C)

43. The ANOVA Table

44. The Completely Randomized Design is called balanced If the number of observations per treatment combination is unequal the design is called unbalanced. (resulting mathematically more complex analysis and computations) If for some of the treatment combinations there are no observations the design is called incomplete. (some of the parameters - main effects and interactions - cannot be estimated.)

45. Example: Diet example Mean = 87.867

46. Main Effects for Factor A (Source of Protein) BeefCerealPork 1.733-2.9671.233

47. Main Effects for Factor B (Level of Protein) HighLow 7.267-7.267

48. AB Interaction Effects Source of Protein BeefCerealPork LevelHigh3.133-6.2673.133 of Protein Low-3.1336.267-3.133

50. Example 2 Paint Luster Experiment

52. Table: Means and Cell Frequencies

53. Means and Frequencies for the AB Interaction (Temp - Drying)

54. Profiles showing Temp-Dry Interaction

55. Means and Frequencies for the AD Interaction (Temp- Thickness)

56. Profiles showing Temp-Thickness Interaction

57. The Main Effect of C (Length)

59. The Randomized Block Design

60. Suppose a researcher is interested in how several treatments affect a continuous response variable (Y). The treatments may be the levels of a single factor or they may be the combinations of levels of several factors. Suppose we have available to us a total of N = nt experimental units to which we are going to apply the different treatments.

61. The Completely Randomized (CR) design randomly divides the experimental units into t groups of size n and randomly assigns a treatment to each group.

62. The Randomized Block Design divides the group of experimental units into n homogeneous groups of size t. These homogeneous groups are called blocks. The treatments are then randomly assigned to the experimental units in each block - one treatment to a unit in each block.

63. Experimental Designs In many experiments were are interested in comparing a number of treatments. (the treatments maybe combinations of levels of several factors.) The objective of Experimental design is to reduce the magnitude of random error resulting in more powerful tests to detect experimental effects

64. The Completely Randomizes Design Treats 123…t Experimental units randomly assigned to treatments

65. Randomized Block Design Blocks All treats appear once in each block

66. The Model for a randomized Block Experiment i = 1,2,…, tj = 1,2,…, b y ij = the observation in the j th block receiving the i th treatment  = overall mean  i = the effect of the i th treatment  j = the effect of the j th Block  ij = random error

67. The Anova Table for a randomized Block Experiment SourceS.S.d.f.M.S.Fp-value TreatSS T t-1MS T MS T /MS E BlockSS B n-1MS B MS B /MS E ErrorSS E (t-1)(b-1)MS E

68. A randomized block experiment is assumed to be a two-factor experiment. The factors are blocks and treatments. The is one observation per cell. It is assumed that there is no interaction between blocks and treatments. The degrees of freedom for the interaction is used to estimate error.

69. Experimental Designs In many experiments were are interested in comparing a number of treatments. (the treatments maybe combinations of levels of several factors.) The objective of Experimental design is to reduce the magnitude of random error resulting in more powerful tests to detect experimental effects

70. The Completely Randomized Design Treats 123…t Experimental units randomly assigned to treatments

71. Randomized Block Design Blocks All treats appear once in each block

72. The matched pair design an experimental design for comparing two treatments Pairs The matched pair design is a randomized block design for t = 2 treatments

73. The Model for a randomized Block Experiment i = 1,2,…, tj = 1,2,…, b y ij = the observation in the j th block receiving the i th treatment  = overall mean  i = the effect of the i th treatment  j = the effect of the j th Block  ij = random error

74. The Anova Table for a randomized Block Experiment SourceS.S.d.f.M.S.Fp-value TreatSS T t-1MS T MS T /MS E BlockSS B n-1MS B MS B /MS E ErrorSS E (t-1)(b-1)MS E

75. Incomplete Block Designs

76. Randomized Block Design We want to compare t treatments Group the N = bt experimental units into b homogeneous blocks of size t. In each block we randomly assign the t treatments to the t experimental units in each block. The ability to detect treatment to treatment differences is dependent on the within block variability.

77. Comments The within block variability generally increases with block size. The larger the block size the larger the within block variability. For a larger number of treatments, t, it may not be appropriate or feasible to require the block size, k, to be equal to the number of treatments. If the block size, k, is less than the number of treatments (k < t)then all treatments can not appear in each block. The design is called an Incomplete Block Design.

78. Comments regarding Incomplete block designs When two treatments appear together in the same block it is possible to estimate the difference in treatments effects. The treatment difference is estimable. If two treatments do not appear together in the same block it not be possible to estimate the difference in treatments effects. The treatment difference may not be estimable.

79. Example Consider the block design with 6 treatments and 6 blocks of size two. The treatments differences (1 vs 2, 1 vs 3, 2 vs 3, 4 vs 5, 4 vs 6, 5 vs 6) are estimable. If one of the treatments is in the group {1,2,3} and the other treatment is in the group {4,5,6}, the treatment difference is not estimable. 1212 2323 1313 4545 5656 4646

80. Definitions Two treatments i and i* are said to be connected if there is a sequence of treatments i 0 = i, i 1, i 2, … i M = i* such that each successive pair of treatments (i j and i j+1 ) appear in the same block In this case the treatment difference is estimable. An incomplete design is said to be connected if all treatment pairs i and i* are connected. In this case all treatment differences are estimable.

81. Example Consider the block design with 5 treatments and 5 blocks of size two. This incomplete block design is connected. All treatment differences are estimable. Some treatment differences are estimated with a higher precision than others. 1212 2323 1313 4545 1414

82. Definition An incomplete design is said to be a Balanced Incomplete Block Design. 1.if all treatments appear in exactly r blocks. This ensures that each treatment is estimated with the same precision 2.if all treatment pairs i and i* appear together in exactly blocks. This ensures that each treatment difference is estimated with the same precision. The value of is the same for each treatment pair.

83. Some Identities Let b = the number of blocks. t = the number of treatments k = the block size r = the number of times a treatment appears in the experiment. = the number of times a pair of treatment appears together in the same block 1.bk = rt Both sides of this equation are found by counting the total number of experimental units in the experiment. 2.r(k-1) = (t – 1) Both sides of this equation are found by counting the total number of experimental units that appear with a specific treatment in the experiment.

84. BIB Design A Balanced Incomplete Block Design (b = 15, k = 4, t = 6, r = 10, = 6)

85. An Example A food processing company is interested in comparing the taste of six new brands (A, B, C, D, E and F) of cereal. For this purpose: subjects will be asked to taste and compare these cereals scoring them on a scale of 0 - 100. For practical reasons it is decided that each subject should be asked to taste and compare at most four of the six cereals. For this reason it is decided to use b = 15 subjects and a balanced incomplete block design to assess the differences in taste of the six brands of cereal.

86. The design and the data is tabulated below:

87. Analysis of Block Experiments

88. The purpose of such experiments is to estimate the effects of treatments applied to some material (experimental units, subjects etc.) grouped into relatively homogeneous groups (blocks). The variability within the groups (blocks) will be considerably less than if the subjects were left ungrouped. This will lead to a more powerful analysis for comparing the treatments.

89. The basic model for block experiments Suppose we have t treatments and b blocks of size k. The Assumption of Additivity

90. Let if treat n is applied to m th unit in j th block otherwise

91. Not of full rank

97. Now define the incidence matrix

98. Note if treat n is applied to m th unit in j th block otherwise Now

99. Thus Also and

100. Thus and

101. Finally

108. Summary: The Least Squares Estimates The Residual Sum of Squares

109. Hence

110. Summary: The Least Squares Estimates The Residual Sum of Squares