Scribd
Upload
25 views17 pages

Notes On Intro To Experimental Design

The document provides an overview of experimental design in engineering, emphasizing the importance of ANOVA (Analysis of Variance) for comparing group means and assessing the significance of differences among treatments. It explains key terminologies, hypotheses for one-way and two-way ANOVA, and outlines the decision-making process based on F-statistics. Examples and exercises illustrate the application of ANOVA in evaluating tensile strength across different materials and factors.

Uploaded by

icharite2003
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
25 views17 pages

Notes On Intro To Experimental Design

The document provides an overview of experimental design in engineering, emphasizing the importance of ANOVA (Analysis of Variance) for comparing group means and assessing the significance of differences among treatments. It explains key terminologies, hypotheses for one-way and two-way ANOVA, and outlines the decision-making process based on F-statistics. Examples and exercises illustrate the application of ANOVA in evaluating tensile strength across different materials and factors.

Uploaded by

icharite2003
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 17

Introduction

to
Experimental Design
Experimental Design
• In engineering, solving practical problems often involves
analyzing data to make informed decisions.
• Experimental design offers an organized method for
organizing experiments and successfully assessing results.
• An essential tool in experimental design is ANOVA
(Analysis of Variance), which enables engineers to compare
group means at the same time and evaluate the impact of
various factors.
Experimental Design
• Most of time, treatment means are different from one
another.
• The question is: Is the difference significant enough to
conclude that one treatment is better (or bigger, smaller,
etc.) than the other?
• The difference is considered significant if the variation
among treatments is proportionally larger than the variation
due to unexplained error.
• ANOVA uses two basic estimates: mean and deviation from
the mean.
Some Useful Terminologies
• Factor: The independent variable being tested (example: type of
alloy).
• Levels: The different groups or categories within the factor
(example: Alloy A, Alloy B, Alloy C).
• Response Variable: The outcome measured in each group
(example: tensile strength).
• Single-Factor (One-Way) ANOVA: Testing the effect of a single
factor with multiple levels.
• Two-Factor and Multi-Factor ANOVA: Analyzing the impact of
two or more factors, along with interactions.
One-Way ANOVA
Hypotheses for One-Way ANOVA:
In one-way ANOVA, we set up the hypotheses as follows:
Null Hypothesis (𝑯𝟎 ): All group means are equal (no significant
difference).
Alternative Hypothesis (𝑯𝒂): At least one group mean is different
from the others.
Mathematically, this can be expressed as:
𝐻0 : 𝜇1 = 𝜇2 = ⋯ = 𝜇𝑘
𝐻𝑎 :At least one 𝜇𝑖 is different,
where 𝜇𝑖 is the mean of group i and k is the number of groups.
ANOVA and F-Statistic
The ANOVA table is as follows:

Source Sum of Degree of Mean Square F-Ratio


Squares Freedom
Between SSB 𝑘 −1 𝑆𝑆𝐵 𝑀𝑆𝐵
𝑀𝑆𝐵 = 𝐹=
Groups 𝑘−1 𝑀𝑆𝑊
Within SSW 𝑁−𝑘 𝑆𝑆𝑊
𝑀𝑆𝑊 =
Groups 𝑁−𝑘
Total SST 𝑁 −1

Where, k is the number of groups being investigated and N


is the total number of observations.
Decision Rule
• Compare the calculated F-statistic to the critical
F-value from an F-distribution table for a given
significance level (𝛼) and degrees of freedom.
This F-value is obtained by using the degree of
freedom 𝑘 − 1, 𝑁 − 𝑘 at the given significance
level.
• If 𝐹-statistic > F-critical, reject 𝐻0 suggesting that
at least one group mean is significantly different.
Example
An engineer tested the tensile strength of three metal alloys (Alloy A,
Alloy B and Alloy C), where each type tested with 5 samples, and the
tensile strength (in Mpa) are as follows:
Alloy A: 30,32,31,29,28
Alloy B: 28,27,29,26,25
Alloy C: 35,34,36,33,32
Analyze the data at 5% significance level to see whether there is a
significant difference in the mean tensile strengths among the three
alloys.
Solution: (To be done while lecturing)
Example

ANOVA results can be obtained by simply using


some software tools like SPSS, Python, Excel,…

For instance, by using Excel, the ANOVA table


for the previous example is as follows (next
slide):

(Step-by-Step will be done in class session)


Example
Exercise
A mechanical engineer is evaluating the impact of three
different types of materials (Material A, Material B, Material
C) on the tensile strength of gears. Six samples are tested for
each material, with the following tensile strength (in MPa):
Material 1: 451, 459, 456, 442, 444, 450
Material 2: 471, 479, 476, 464, 467, 470
Material 3: 435, 431, 429, 437, 441, 448
Test at 1% significance level, if there is a significant difference
in the mean tensile strengths among the three materials.
TWO-WAY ANOVA
Two-way ANOVA is used to analyze the effect of two independent categorical
factors on a continuous dependent variable.
With this method, we determine whether each factor independently affects the
dependent variable (assessing the main effects), and also check if there is an
interaction between the factors (assessing the interaction effect).
We have two cases:
• Two-Way ANOVA without replication: Only one observation per
combination of factors
• Two-way ANOVA with replication: Multiple observations per
combination of factors.
TWO-WAY ANOVA
The hypotheses for Two-Way ANOVA are defined as follows:
For factor A:
Null hypothesis( 𝐻0𝐴 ): There is no main effect for factor A
For factor B:
Null hypothesis( 𝐻0𝐵 ): There is no main effect for factor B
For interaction (A and B)
Null hypothesis( 𝐻0𝐴𝐵 ): There is no interaction effect between
factor A and factor B.
TWO-WAY ANOVA
The general layout of a Two-Way ANOVA is as follows

Source of Sum of Degree of Freedom (df) Mean F-Ratio (F)


Variation Squares Square
(SS) (MS)
Factor A 𝑆𝑆𝐴 𝑑𝑓𝐴 = 𝑘 − 1 𝑆𝑆𝐴 𝑀𝑆𝐴
𝑀𝑆𝐴 = 𝐹𝐴 =
(main 𝑑𝑓𝐴 𝑀𝑆𝐸
effect)
Factor B 𝑆𝑆𝐵 𝑑𝑓𝐵 = 𝑚 − 1 𝑆𝑆𝐵 𝑀𝑆𝐵
𝑀𝑆𝐵 = 𝐹𝐵 =
(main 𝑑𝑓𝐵 𝑀𝑆𝐸
effect)

Interaction 𝑆𝑆𝐴𝐵 𝑑𝑓𝐴𝐵 = (𝑘 − 1)(𝑚 − 1) 𝑀𝑆𝐴𝐵 𝑀𝑆𝐴𝐵


𝑆𝑆𝐴𝐵 𝐹𝐴𝐵 =
(A&B)
= 𝑀𝑆𝐸
𝑑𝑓𝐴𝐵
Error 𝑆𝑆𝐸 𝑑𝑓𝐸 = 𝑁 − 𝑘𝑚 𝑆𝑆𝐸 -
𝑀𝑆𝐸 =
(residual) 𝑑𝑓𝐸
Total 𝑆𝑆𝑇 𝑑𝑓𝑇 = 𝑁 − 1 - -
TWO-WAY ANOVA
• For the table in previous slide;
k: Number of levels of Factor A,

m: Number of levels of Factor B and


N: Total number of observations.

• For decisions, we reject the Null hypothesis if F


(calculated) is greater than the F-critical (from the
F-table).
Example
Consider the following marks obtained by pupils at a primary a
school in a Maths test. Analyze the results at 5% S.L.

Gender B B B G G G B B B

Score 5 6 8 4 7 8 9 6 7

Age 10 10 10 10 10 10 11 11 11

Gender G G G B B B G G G

Score 7 10 13 8 9 13 12 14 16

Age 11 11 11 12 12 12 12 12 12
Example
Sol:
Here we have two factors: Gender and Age
Our Null Hypotheses will be:
𝐻0𝐴 : Gender has no significant effect on students’ score
𝐻0𝐵 : Age has no significant effect on students’ score
𝐻0𝐴𝐵 : Gender and age interaction have no significant
effect on students’ score

You might also like