Analysis of variance ANOVA Used in Experimental Design Example: to compare the performance of several materials in manufacturing depending on several factors. In which known sources of variability are varied over a wide range . There are k samples from k populations. A common procedure used to deal with testing population means is called the Analysis of Variance (ANOVA). Example ANOVA An engineer is interested in how the mean absorption of moisture in concrete varies among 5 different concrete aggregates. The samples are exposed to moisture for 48 hours. A sample of size n = 6 observations from each k = 5 = number of populations is to be tested for each aggregate. We want to test 𝜇1 , 𝜇2 , . . , 𝜇5 are the same . Hypothesis test
The goal of ANOVA is to determine if the differences among the 5
sample means are due to - random variation alone or - variation beyond merely random effects, i.e., differences in the chemical composition of the aggregates. The k=5 different populations are classified on the basis of a single treatment (here different aggregates)=⇒ One-Way ANOVA.
Fall 2021 DR. MAHA A. HASSANEIN 4
Analysis of Variance
Fall 2021 DR. MAHA A. HASSANEIN 5
Total Variability
Fall 2021 DR. MAHA A. HASSANEIN 6
Resolution of total variability
Fall 2021 DR. MAHA A. HASSANEIN 7
Fall 2021 DR. MAHA A. HASSANEIN 8 Fall 2021 DR. MAHA A. HASSANEIN 9 Fall 2021 DR. MAHA A. HASSANEIN 10 Fall 2021 DR. MAHA A. HASSANEIN 11 Fall 2021 DR. MAHA A. HASSANEIN 12 Fall 2021 DR. MAHA A. HASSANEIN 13 Fall 2021 DR. MAHA A. HASSANEIN 14 Example The students split into three groups, in each group are randomly assigned to use one of the three exam prep programs for the next three weeks to prepare for an exam. At the end of the three weeks, all of the students take the same exam. The exam scores for each group are shown in table. Perform a one- way ANOVA by hand to determine if the mean exam score is different between the three groups
Fall 2021 DR. MAHA A. HASSANEIN 15
Step 1 First, we will calculate the mean for all three groups along with the overall mean:
Fall 2021 DR. MAHA A. HASSANEIN 16
Step 2 Step 2: Calculate SSA. SSA= = 10(83.4-85.8)2 + 10(89.3- 85.8)2 + 10(84.7-85.8)2 = 192.2 Next, we will calculate the regression sum of squares (SSA) using the following formula:
Fall 2021 DR. MAHA A. HASSANEIN 17
Step 3 Next, we will calculate the error Group 1: (85-83.4)2 + (86-83.4)2 + (88- sum of squares (SSE) using the 83.4)2 + (75-83.4)2 + (78-83.4)2 + (94- following formula: 83.4)2 + (98-83.4)2 + (79-83.4)2 + (71- 83.4)2 + (80-83.4)2 = 640.4 Group 2: (91-89.3)2 + (92-89.3)2 + (93- 89.3)2 + (85-89.3)2 + (87-89.3)2 + (84- 89.3)2 + (82-89.3)2 + (88-89.3)2 + (95- 89.3)2 + (96-89.3)2 = 208.1 SSE: 640.4 + 208.1 + 252.1 = 1100.6 Group 3: (79-84.7)2 + (78-84.7)2 + (88- 84.7)2 + (94-84.7)2 + (92-84.7)2 + (85- 84.7)2 + (83-84.7)2 + (85-84.7)2 + (82- 84.7)2 + (81-84.7)2 = 252.1
Fall 2021 DR. MAHA A. HASSANEIN 18
Step 4 Calculate the total sum df treatment: k-1 = 3-1 = 2 of squares (SST) using df error: n-k = 30-3 = 27 the following formula: df total: n-1 = 30-1 = 29 SST = SSA + SSE MS treatment: SST / df treatment = 192.2 / 2 = 96.1 MS error: SSE / df error = 1100.6 / 27 = 40.8 SST = 192.2 + 1100.6 = 1292.8 F: MS treatment / MS error = 96.1 / 40.8 = 2.358
Fall 2021 DR. MAHA A. HASSANEIN 19
Step 5 Fill in the ANOVA table. Now that we have SSA, SSE, and SST, we can fill in the ANOVA table:
Fall 2021 DR. MAHA A. HASSANEIN 20
Step 6 Interpret the results. α (significance level) = 0.05 DF1 (numerator degrees of freedom) = df treatment = 2 DF2 (denominator degrees of freedom) = df error = 27 The F test statistic for this one-way ANOVA is 2.358 F critical value is 3.3541. 𝐹𝛼 < 𝐹𝑐 Decision: we fail to reject the null hypothesis. This means we don’t have sufficient evidence to say that there is a statistically significant difference between the mean exam scores of the three groups.