Transcript Chapter 11
Business Statistics, 4e by Ken Black Chapter 11 D iscrete D istributions Analysis of Variance & Design of Experiments Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. 11-1 Learning Objectives • Understand the differences between various experimental designs and when to use them. • Compute and interpret the results of a one-way ANOVA. • Compute and interpret the results of a random block design. • Compute and interpret the results of a two-way ANOVA. • Understand and interpret interaction. • Know when and how to use multiple comparison techniques. Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. 11-2 Introduction to Design of Experiments, #1 Experimental Design - a plan and a structure to test hypotheses in which the researcher controls or manipulates one or more variables. Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. 11-3 Introduction to Design of Experiments, #2 Independent Variable • Treatment variable is one that the experimenter controls or modifies in the experiment. • Classification variable is a characteristic of the experimental subjects that was present prior to the experiment, and is not a result of the experimenter’s manipulations or control. • Levels or Classifications are the subcategories of the independent variable used by the researcher in the experimental design. Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. 11-4 Introduction to Design of Experiments, #3 Dependent Variable - the response to the different levels of the independent variables. Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. 11-5 Three Types of Experimental Designs • Completely Randomized Design • Randomized Block Design • Factorial Experiments Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. 11-6 Completely Randomized Design 1 Machine Operator 2 3 4 Valve Opening Measurements . . . . . . . . . Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. . . . 11-7 Valve Openings by Operator 1 2 3 4 6.33 6.26 6.44 6.29 6.26 6.36 6.38 6.23 6.31 6.23 6.58 6.19 6.29 6.27 6.54 6.21 6.4 6.19 6.56 6.5 6.34 6.19 6.58 6.22 Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. 11-8 Analysis of Variance: Assumptions • Observations are drawn from normally distributed populations. • Observations represent random samples from the populations. • Variances of the populations are equal. Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. 11-9 One-Way ANOVA: Procedural Overview H o: 1 2 3 k H a : A t least one of th e m eans is different from the others F If F > If F Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. F F c c M SC M SE , reject H o . , do not reject H o . 11-10 One-Way ANOVA: Sums of Squares Definitions + between = error sum of squares total sum of squares sum of squares SST = SSC + SSE X ij X nj i =1 C 2 C n j 1 j= 1 where : j X j i particular C = number j number X = grand X X j ij i 1 X C j 1 member ij X 2 j of a treatment of treatment levels of observatio ns in a given trea level tment level mean = mean of a treatment individual Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. nj level j = a treatment n X 2 group or level value 11-11 Partitioning Total Sum of Squares of Variation SST (Total Sum of Squares) SSC (Treatment Sum of Squares) Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. SSE (Error Sum of Squares) 11-12 One-Way ANOVA: Computational Formulas C j j 1 SSE X SST nj C i 1 j 1 nj C j X ij X MSC MSE 2 ij X j 1 i 1 F n X j X SSC df 2 df N C E 2 SSC df C 1 C C df N 1 T w here: i = a particu lar m em ber of a treatm ent leve l j = a treatm ent level SSE C = num ber of treatm ent levels df n j = num ber of observations in a g iven treatm ent level E MSC MSE X = grand m ean X X ij Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. j colum n m e an = individua l value 11-13 One-Way ANOVA: Preliminary Calculations 1 2 3 4 6.33 6.26 6.44 6.29 6.26 6.36 6.38 6.23 6.31 6.23 6.58 6.19 6.29 6.27 6.54 6.21 6.4 6.19 6.56 6.5 6.34 6.19 6.58 6.22 Tj T1 = 31.59 T2 = 50.22 T3 = 45.42 T4 = 24.92 T = 152.15 nj n1 = 5 n2 = 8 n3 = 7 n4 = 4 N = 24 6.318000 6.277500 6.488571 6.230000 6.339583 Mean Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. 11-14 One-Way ANOVA: Sum of Squares Calculations C SSC n X j X j j 1 [ 5 ( 6 . 318 6 . 339583 ) 7 ( 6 . 488571 SSE 2 2 8 ( 6 . 2775 6 . 339583 ) 2 6 . 339583 ) 2 6 . 339583 ) 2 ( 6 . 31 6 . 318 ) 2 4 ( 6 . 23 0 . 23658 X nj C i 1 j 1 ij X 2 j ( 6 . 33 6 . 318 ) 2 ( 6 . 4 6 . 318 ) 2 ( 6 . 26 6 . 318 ) 2 ( 6 . 26 6 . 2775 ) ( 6 . 22 6 . 230 ) 0 . 15492 Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. 2 2 ( 6 . 29 6 . 318 ) ( 6 . 36 6 . 2775 ) ( 6 . 19 6 . 230 ) 2 2 11-15 2 One-Way ANOVA: Sum of Squares Calculations SST nj C i 1 j 1 X ij X 2 ( 6 . 33 6 . 339583 ) 2 ( 6 . 26 6 . 339583 ) ( 6 . 31 6 . 339583 ) 2 ( 6 . 19 6 . 339583 ) 0 . 39150 Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. 2 ( 6 . 22 6 . 339583 ) 2 2 11-16 One-Way ANOVA: Mean Square and F Calculations C 1 4 1 3 df C df E df T N C 24 4 20 N 1 24 1 23 MSC MSE F SSC df MSC MSE Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. . 078860 3 C SSE df . 23658 . 15492 . 007746 20 E . 078860 10 . 18 . 007746 11-17 Analysis of Variance for Valve Openings Source of Variance df Between Error Total 3 20 23 Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. SS MS 0.23658 0.078860 0.15492 0.007746 0.39150 F 10.18 11-18 A Portion of the F Table for = 0.05 F . 05 , 3 , 20 df1 df 2 1 2 3 4 5 6 7 8 9 1 161.45 199.50 215.71 224.58 230.16 233.99 236.77 238.88 240.54 … … … … … … … … … … 18 4.41 3.55 3.16 2.93 2.77 2.66 2.58 2.51 2.46 19 4.38 3.52 3.13 2.90 2.74 2.63 2.54 2.48 2.42 20 4.35 3.49 3.10 2.87 2.71 2.60 2.51 2.45 2.39 21 4.32 3.47 3.07 2.84 2.68 2.57 2.49 2.42 2.37 Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. 11-19 One-Way ANOVA: Procedural Summary Ho : 1 2 3 4 H a : At least one of the means is different If F > If F FF c c from the others 1 Rejection Region 3 2 20 3 . 10 , reject H o . 3 . 10 , do reject H o . Since F = 10.18 > F c 3 . 10 , reject H o . Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. Non rejection Region F . 05 , 9 ,11 3 . 10 Critical Value 11-20 Excel Output for the Valve Opening Example Anova: Single Factor SUMMARY Groups Count Sum Average Variance Operator 1 5 31.59 6.318 0.00277 Operator 2 8 50.22 6.2775 0.0110786 Operator 3 7 45.42 6.488571429 0.0101143 Operator 4 4 24.92 6.23 0.0018667 ANOVA Source of Variation SS df MS Between Groups 0.236580119 3 0.07886004 Within Groups 0.154915714 20 0.007745786 Total 0.391495833 23 Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. F P-value F crit 10.181025 0.00028 3.09839 11-21 Multiple Comparison Tests An analysis of variance (ANOVA) test is an overall test of differences among groups. Multiple Comparison techniques are used to identify which pairs of means are significantly different given that the ANOVA test reveals overall significance. • Tukey’s honestly significant difference (HSD) test requires equal sample sizes • Tukey-Kramer Procedure is used when sample sizes are unequal. Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. 11-22 Tukey’s Honestly Significant Difference (HSD) Test H SD q M SE ,C ,N -C n w here: M S E = m ean square error n = sam ple size q ,C ,N -C = critical value of the studentized range distribution from T able A .10 Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. 11-23 Data from Demonstration Problem 11.1 PLANT (Employee Age) Group Means nj 1 29 27 30 27 28 2 32 33 31 34 30 3 25 24 24 25 26 28.2 5 32.0 5 24.8 5 C=3 dfE = N - C = 12 Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. MSE = 1.63 11-24 q Values for = .01 Number of Populations Degrees of Freedom 1 2 3 4 5 90 135 164 186 2 14 19 22.3 24.7 3 8.26 10.6 12.2 13.3 4 6.51 8.12 9.17 9.96 11 4.39 5.14 5.62 5.97 12 4.32 5.04 5.50 5.84 ... q . 01 , 3 ,12 5.04 . . Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. 11-25 Tukey’s HSD Test for the Employee Age Data H SD X 1 X 1 X 2 X 2 X 3 X 3 q M SE ,C , N C n 5.0 4 1.6 3 5 2 .8 8 2 8 .2 3 2 .0 3.8 2 8 .2 2 4 .8 3.4 3 2 .0 2 4 .8 7 .2 Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. 11-26 Tukey-Kramer Procedure: The Case of Unequal Sample Sizes H SD q M SE ,C ,N -C 2 ( 1 n r 1 n ) s w here: M S E = m ean square error n n q r s ,C ,N -C = sam ple size for = sam ple size for r s th th sam ple sam ple = critical value of the studentized range distribution from T able A .10 Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. 11-27 Freighter Example: Means and Sample Sizes for the Four Operators Operator 1 2 3 4 Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. Sample Size 5 8 7 4 Mean 6.3180 6.2775 6.4886 6.2300 11-28 Tukey-Kramer Results for the Four Operators Pair 1 and 2 Critical Difference .1405 |Actual Differences| .0405 1 and 3 .1443 .1706* 1 and 4 .1653 .0880 2 and 3 .1275 .2111* 2 and 4 .1509 .0475 3 and 4 .1545 .2586* *denotes significant at .05 Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. 11-29 Partitioning the Total Sum of Squares in the Randomized Block Design SST (Total Sum of Squares) SSE (Error Sum of Squares) SSC (Treatment Sum of Squares) SSR (Sum of Squares Blocks) Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. SSE’ (Sum of Squares Error) 11-30 A Randomized Block Design Single Independent Variable . Individual observations . Blocking Variable . . . . . . . . . . . . . . . Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. 11-31 Randomized Block Design Treatment Effects: Procedural Overview Ho : 1 2 3 k H a : At least one of the means is different F If F > If F from the others M SC M SE F F c c , reject H o . , do not reject H o . Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. 11-32 Randomized Block Design: Computational Formulas C SSC n j 1 n ( X j X ) 2 ( X i X ) 2 SSR C i 1 n n j 1 i 1 n n ( X ij X ) SST j 1 i 1 M SC SSC C 1 SSR M SR SSE N nC 1 trea tm en ts b lo cks C R 2 df E C 1 n1 C 1 n 1 N n C 1 2 df E N 1 w here: i = block group (row ) j = a treatm ent level (colum n) C = num ber of treatm ent levels (colum ns) n = num ber of observations in each treatm en t level (num ber of blocks - row s) n1 M SE F df ( X ij X i X i X ) SSE F df M SC M SE M SR M SE X ij X j X i individual observation SSC sum of squares colum ns (treatm ent) treatm ent (colum n) m ean SSR = sum of squares row s (blocking) block (row ) m ean SSE = sum of squares error SST = sum of squares total X = grand m ean N = total num ber of observations Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. 11-33 Randomized Block Design: Tread-Wear Example Speed Supplier Slow Medium Fast Block Means ( X ) i n=5 1 3.7 4.5 3.1 3.77 2 3.4 3.9 2.8 3.37 3 3.5 4.1 3.0 3.53 4 3.2 3.5 2.6 3.10 5 3.9 4.8 3.4 4.03 3.54 4.16 2.98 3.56 Treatment Means( X ) j N = 15 X C=3 Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. 11-34 Randomized Block Design: Sum of Squares Calculations (Part 1) C SSC n j 1 ( X j X ) 2 5[ (3.5 4 3.5 6 ) 2 (4.1 6 3.5 6 ) 2 (2.9 8 3.5 6 ) 2 (3.5 3 3.5 6 ) 2 3.4 8 4 n SSR C i 1 ( X i X ) 2 3[ (3.7 7 3.5 6 ) 2 (3.3 7 3.5 6 ) 2 (3.1 0 3.5 6 ) 2 (4.0 3 3.5 6 ) 1.5 4 9 Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. 11-35 2 ] Randomized Block Design: Sum of Squares Calculations (Part 2) SSE n C i 1 j 1 ( X ij X j X i X ) 2 (3 . 7 3 . 54 3 . 77 3 . 56 ) 2 ( 2 . 6 2 . 98 3 . 10 3 . 56 ) 0 . 143 SST n C i 1 j 1 ( X ij X ) (3 . 7 3 . 56 ) 5 . 176 2 2 (3 . 4 3 . 54 3 . 37 3 . 56 ) 2 (3 . 4 2 . 98 4 . 03 3 . 56 ) 2 2 (3 . 4 3 . 56 ) Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. 2 ( 2 . 6 3 . 56 ) 2 (3 . 4 3 . 56 ) 2 11-36 Randomized Block Design: Mean Square Calculations M SC M SR SSC C 1 SSR n1 M SE 3.484 2 1.549 SSE 4 1.742 0 .387 0 .143 N nC 1 8 M SC 1.742 F 96 .78 M SE 0 .018 Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. 0 .018 11-37 Analysis of Variance for the Tread-Wear Example Source of VarianceSS df Treatment 3.484 Block 1.549 Error 0.143 Total 5.176 Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. MS 2 4 8 14 F 1.742 0.387 0.018 96.78 21.50 11-38 Randomized Block Design Treatment Effects: Procedural Summary H o: 1 2 3 H a : A t least one of th e m eans is different from the others F MSC MSE 1 . 742 96 . 78 0 . 018 F = 96.78 > F Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. .01,2,8 = 8.65, reject H o . 11-39 Randomized Block Design Blocking Effects: Procedural Overview H o: 1 2 3 4 5 H a: A t least one of the blocking m eans is different from the others F MSR MSE F = 2 1 .5 > . 387 21 . 5 . 018 F .0 1, 4 ,8 Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. = 7 .0 1 , reject H o . 11-40 Excel Output for Tread-Wear Example: Randomized Block Design Anova: Two-Factor Without Replication SUMMARY Suplier 1 Suplier 2 Suplier 3 Suplier 4 Suplier 5 Count Sum 11.3 10.1 10.6 9.3 12.1 Average 3.7666667 3.3666667 3.5333333 3.1 4.0333333 Variance 0.4933333 0.3033333 0.3033333 0.21 0.5033333 5 17.7 5 20.8 5 14.9 3.54 4.16 2.98 0.073 0.258 0.092 3 3 3 3 3 Slow Medium Fast ANOVA Source of Variation SS df MS F P-value F crit Rows 1.5493333 4 0.3873333 21.719626 0.0002357 7.0060651 Columns 3.484 2 1.742 97.682243 2.395E-06 8.6490672 Error 0.1426667 8 0.0178333 Total 5.176 Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. 14 11-41 Two-Way Factorial Design Column Treatment . . Row Treatment Cells . . . . . . . . . . . . . . . Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. 11-42 Two-Way ANOVA: Hypotheses R ow E ffects: H o : R ow M eans are all equal. H a : A t least one row m ean is different from the others. C olum ns E ffects: H o : C olum n M eans are all equal. H a : A t least one colum n m ean is different from the others. Interaction E ffects: H o : T he interaction effects are zero. H a : T here is an interaction effect. Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. 11-43 Formulas for Computing a Two-Way ANOVA R SSR nC i 1 C SSC nR R ( X i X ) ( X j X ) (X SST (X c 1 r 1 a 1 M SR M SC M SI M SE ijk j 1 k 1 R n ijk R R 1 2 df ( X ij X i X j X ) i 1 j 1 R C n i 1 C df j 1 C SSI n SSE 2 X ij ) X) C 2 df I w h ere : C 1 n = n u m b er o f o b serv atio n s p er cell R 1 C 1 C = n u m b er o f co lu m n treatm en ts R = n u m b er o f ro w treatm en ts 2 df E R C n 1 i = ro w treatm en t lev el j = co lu m n treatm en t le v el 2 SSR R 1 SSC C 1 SSI R 1 C 1 df F F F T R C I N 1 M SR M SE M SC M SE M SI M SE k = cell m em b er X X X X ijk ij i j = in d iv id u a l o b serv atio n = cell m ean = ro w m ean = co lu m n m ean X = gran d m ean SSE R C n 1 Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. 11-44 A 2 3 Factorial Design with Interaction Row effects Cell Means R1 R2 C1 Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. C2 Column C3 11-45 A 2 3 Factorial Design with Some Interaction Row effects Cell Means R1 R2 C1 Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. C2 Column C3 11-46 A 2 3 Factorial Design with No Interaction Row effects Cell Means R1 R2 C1 C2 C3 Column Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. 11-47 A 2 3 Factorial Design: Data and Measurements for CEO Dividend Example Location Where Company Stock is Traded How Stockholders are Informed of Dividends Annual/Quarterly Reports Presentations to Analysts Xj NYSE AMEX 2 1 2 1 X11=1.5 2 3 1 2 X21=2.0 2 3 3 2 X12=2.5 3 3 2 4 X22=3.0 1.75 2.75 Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. OTC Xi 4 3 4 2.5 3 X13=3.5 4 4 3 2.9167 4 X23=3.75 X=2.7083 N = 24 n=4 3.625 11-48 A 2 3 Factorial Design: Calculations for the CEO Dividend Example (Part 1) R SSR nC ( X i X ) i 1 2 ( 4 )( 3 )[( 2 .5 2 .7 0 8 3 ) ( 2 .9 1 6 7 2 .7 0 8 3 ) ] 2 2 1.0 4 1 8 C SSC nR j 1 ( X j X ) 2 ( 4 )( 2 )[(1.7 5 2 .7 0 8 3 ) ( 2 .7 5 2 .7 0 8 3 ) ( 3.6 2 5 2 .7 0 8 3 ) ] 2 2 2 1 4 .0 8 3 3 R SSI n i 1 C ( X ij X i X j X ) 2 j 1 4 [(1.5 2 .5 1.7 5 2 .7 0 8 3 ) ( 2 .5 2 .5 2 .7 5 2 .7 0 8 3 ) 2 2 ( 3.5 2 .5 3.6 2 5 2 .7 0 8 3 ) ( 2 .0 2 .9 1 6 7 1.7 5 2 .7 0 8 3 ) 2 2 ( 3.0 2 .9 1 6 7 2 .7 5 2 .7 0 8 3 ) ( 3.7 5 2 .9 1 6 7 3.6 2 5 2 .7 0 8 3 ) ] 2 2 0.0 8 3 3 Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. 11-49 A 2 3 Factorial Design: Calculations for the CEO Dividend Example (Part 2) R SSE n (X ijk X ij ) ( 2 1.5) (1 1.5) i 1 C j 1 k 1 2 2 2 ( 3 3.7 5) 2 ( 4 3.7 5) 2 7 .7 5 0 0 C SST R n (X c 1 r 1 a 1 ijk ( 2 2 .7 0 8 3) 2 2 .9 5 8 3 X) 2 2 (1 2 .7 0 8 3) Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. 2 ( 3 2 .7 0 8 3) 2 ( 4 2 .7 0 8 3) 2 11-50 A 2 3 Factorial Design: Calculations for the CEO Dividend Example (Part 3) M SR SSR 1.0418 1 14 .0833 M SC 7 .0417 C 1 2 SSI 0 .0833 M SI 0 .0417 R 1 C 1 2 SSE 7 .7500 M SE 0 .4306 R C n 1 18 R 1 SSC 1.0418 Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. F R F C F I M SR M SE M SC M SE M SI M SE 1.0418 0 .4306 7 .0417 0 .4306 0 .0417 0 .4306 2 .42 16 .35 0 .10 11-51 Analysis of Variance for the CEO Dividend Problem Source of VarianceSS df Row 1.0418 Column 14.0833 Interaction 0.0833 Error 7.7500 Total 22.9583 *Denotes MS 1 2 2 18 23 F 1.0418 2.42 7.0417 16.35* 0.0417 0.10 0.4306 significance at = .01. Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. 11-52 Excel Output for the CEO Dividend Example (Part 1) Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. Anova: Two-Factor With Replication SUMMARY NYSE ASE OTC Total AQReport Count 4 4 4 12 Sum 6 10 14 30 Average 1.5 2.5 3.5 2.5 Variance 0.3333 0.3333 0.3333 1 Presentation Count Sum Average Variance 4 8 2 0.6667 4 12 3 0.6667 4 15 3.75 0.25 8 14 1.75 0.5 8 22 2.75 0.5 8 29 3.625 0.2679 12 35 2.9167 0.9924 Total Count Sum Average Variance 11-53 Excel Output for the CEO Dividend Example (Part 2) ANOVA Source of Variation Sample Columns Interaction Within SS 1.0417 14.083 0.0833 7.75 Total 22.958 Business Statistics, 4e, by Ken Black. © 2003 John Wiley & Sons. df 1 2 2 18 MS 1.0417 7.0417 0.0417 0.4306 F P-value F crit 2.4194 0.1373 4.4139 16.355 9E-05 3.5546 0.0968 0.9082 3.5546 23 11-54