3. β Did you know βDIETβ
stands for :
DID I EAT THAT?
3
4. Table of
Content
β« Introduction
β« Methodology
β« Our Data
β« Descriptive Statistics
β« Proposed Hypothesis
β« Test for Normality & Homogeneity of Variance
β« Analysis of Variance (ANOVA)
β« Conclusions
β« Post βHoc Tests
β« Inferences
4
5. Designs of
Experiment
Designs of experiment (DOE) is defined as a
branch of applied statistics that deals with
planning, conducting, analyzing, and interpreting
tests to evaluate the factors that control the value
of a parameter or group of parameters. DOE is a
powerful data collection and analysis tool that
can be used in a variety of experimental
situations.
5
7. Analysis of variance (ANOVA) is a statistical technique that is used to check
if the means of two or more groups are significantly different from each
other. ANOVA checks the impact of one or more factors by comparing the
means of different samples.
7
8. ASSUMPTIONS
1. All effects are additive.
2. The observations are independent of each
other.
3. eijβs βΌ N(0, Ο2)
4. The dependent variables should be
measurable.
5. Parent population from which samples are
drawn should be normal.
6. Homogeneity of variance between the groups
should be satisfied. 8
9. A test that allows one to make comparisons
between the means of three or more groups
of data, where one independent variables is
considered.
The one-way ANOVA model is given by,
π₯ππ = π + πΌπ + πππ
Dependent variable = overall mean +
treatment effect + random error
9
One-Way ANOVA
10. 10
Null Hypothesis:
π»0: π1 = π2 = β― = ππ
(There is no difference in the
population means)
VS
Alternative Hypothesis:
π»1: ππ β ππ
(At least two population
means are different)
TSS = SSTr + SSE
πππ = πππ‘ππ ππ’π ππ πππ’ππππ =
π π
(π₯ππ β π₯. . )2
ππππ = ππ’π ππ πππ’ππππ ππ’π π‘π πππππ‘ππππ‘π =
π π
(π₯π. βπ₯. . ) =
π
ππ( π₯π. βπ₯. . )
πππΈ = ππ’π ππ πππ’ππππ ππ’π π‘π πΈπππππ =
π π
(π₯ππ β π₯π. )2
11. Source of
Variation
Sum of
Squared
Degree of
Freedom
Mean Sum
of Squares
Fcal Ftab
Treatment SSTr k-1 πππππ
=
ππππ
π β 1
πΉ
=
πππππ
ππππΈ
πΉ
πΌ(π β 1, π β π)
Error SSE N-k ππππΈ
=
πππΈ
π β π
Total TSS N-1
11
12. A test that allows one to make comparisons
between the means of three or more groups of
data, where two independent variables are
considered.
The two-way ANOVA model is given by,
π₯ππ = π + πΌπ + π½π + πππ
Dependent variable = overall mean + treatment
effect + block effect + random error
12
Two Way ANOVA
13. 13
VS
Null Hypothesis:
π»01: π1 = π2 = β― = ππ
(There is no treatment effect)
π»02: π1 = π2 = β― = πβ
(There is no block effect)
Alternative Hypothesis:
π»11: ππ β ππ
(There is a treatment effect)
π»12: ππ β ππ
(There is a block effect)
If there is an interaction effect between the treatments and blocks, then there is another
hypothesis which is given as
H03 : πΎππ = 0
(There is no interaction effect)
H13 : πΎππ β 0
(There is an interaction effect)
15. METHODOLOGY
1. Designing the experiment
2. Exploratory data analysis
3. Checking for normality
4. Homogeneity of variance
5. ANOVA
6. Conclusion
7. Inferences
15
16. DATA
This is a diet dataset. The dataset contains
information on 78 people who undertook one of
three diet types (Keto diet, Intermittent fasting
diet and Sattvic diet). There is background
information such as age, gender, and height. The
aim of the study is to see which diet is best for
losing weight.
Keto Diet β A
Intermittent Diet β B
Sattvic Diet β C
16
17. Descriptive
Statistics
Summary
for Diet and
Gender
17
n Mean (in
kg)
SD Skewness Kurtosis
Diet A 24 3.3 2.24 0.88 0.65
Diet B 27 3.03 2.52 -0.17 -0.74
Diet C 27 5.15 2.4 -0.34 -0.95
Female 45 3.72 2.59 0.01 -0.92
Male 33 4.02 2.53 0.12 -0.22
20. HYPOTHESIS
H01 : π1 = π2
(There is no effect
of gender on weight
loss)
v/s
H11 : ππ β ππ
(There is an effect
of gender on weight
loss)
H02 : π1 = π2 = π3
(There is no effect
of diet type on
weight loss)
v/s
H12 : ππ β ππ
(There is an effect
of diet type on
weight loss)
H03 : πΎππ = 0
(There is no
interaction effect
between gender
and diet)
v/s
H13 : πΎππ β 0
(There is an
interaction effect
between gender
and diet)
20
24. ANOVA TABLE
24
Degrees
of
Freedom
Sum of
Squares
Mean
Sum of
Squares
F-
calculate
d value
F-table
value
Diet 2 71.09 35.547 6.5925 3.15
Gender 1 1.04 1.035 0.1920 4.00
Diet*Gend
er
2 40.92 20.460 3.7945 3.15
Error 72 388.22 5.392
26. 1. Gender does not significantly affect average
weight loss.
2. Diet alone does not affect weight.
3. But combination of Gender and Diet has a
significant effect on weight loss.
26
31. β« Dieticians can use the result of this study to
analyse which diet is suitable for weight loss
for males and females.
β« Further analysis can be done using the factors
age-group and diet.
31
FUTURE
SCOPE OF
THE STUDY