Slide:

https://www.canva.com/design/DAGJrv-W2bM/bbH2qCV4iXrklOH7LQygPQ/edit?ut
m_content=DAGJrv-W2bM&utm_campaign=designshare&utm_medium=link2&utm
_source=sharebutton

Chap 15
Technique overview

1. Frequency distribution

Definition: A mathematical distribution whose object is to obtain a count of the number of

responses associated with different values of one variable and to express these counts in
percentage terms
Column 1: Different categories
Column 2: Code assigned to each value (ascending level of familiarity)
Column 3: Number of respondents checking each value
Column 4: Percentage of respondents checking each value
Column 5: Percentages calculated by excluding the cases with missing values (if so)
Column 6: Cumulative percentage

X-axis: Values of the variable

Y-axis: Absolute or relative frequencies of the values
Case: Motivational factors influencing the decision the attend the Olympic Games
The results of this study helped planners for the 2008 Olympic Games in Beijing

2. Statistics associated with Frequency Distribution

a. Measures of location (mean, mode, median)
- Describe the center of the distribution, if the sample is changed by adding a
fixed constant to Image
b. Measures of variability
- Range: Measures the spread of the data, which is directly affected by outliers
Image
- Interquartile range: The difference between the 75th and 25th percentile Image
- Variance and Standard Deviation: Image
 + Deviation from the mean: Difference between the mean and an
observed value Image
+ Variance: mean squared deviation
- Coefficient of variation: Ratio of the SD to the mean Image
Image Chú thích dưới ảnh: Mean, standard deviation and coefficient of
variation of age at harvesting
c. Measures of shape (skewness and kurtosis)

3. Cross-tabulation
Definition: While a frequency distribution describes one variable at a time, a
cross-tabulation describes two or more variables simultaneously.
- Size: The number of categories of one variable x the number of categories of
the second variables
EX:
- Since two variables have been cross-classified, percentages could be computed
either column-wise, based on column totals, or row-wise, based on row totals
(Note: hải thanh chia thành 2 slides cho column-wise với row-wise nha)
- Row-wise:

- However, the general rule is to compute the percentages in the direction of the
independent variable, across the dependent variable → The correct way of
calculating percentages is as shown in Table 15.4
- Column-wise:
ANOVA - Analysis of Variance
1. Opening Scenario:
(Tao ghi tiêu đề Anova trước Opening Scenario cho mày dễ hiểu thôi. Lúc làm slide thì làm
slide of OS trước rồi mới tới slide giới thiệu Anova giúp tao nha, để tới đó tao đặt câu hỏi
dẫn vô phần này cho mượt á.)

Imagine you're a researcher investigating the effects of different exercise programs on weight
loss. You have three groups of participants:

● Group 1: Aerobic exercise only

● Group 2: Strength training only

● Group 3: Combined aerobic and strength training

You want to know if there are any differences in average weight loss among these three
groups.

→ Can we use the T-test in this case?

2. Definition:
A statistical technique for examining the differences among means for two or more
populations. It is used to determine if there is a statistically significant difference between
two or more categorical groups by testing for differences of means using a variance.

3. When to Use?
Scenario 1: Comparing Means of Three or More Groups

ANOVA is most commonly used when you want to compare the means of three or more
independent groups.

● Business example: A marketing team wants to compare the effectiveness of three

different advertising campaigns (TV, online, print) in increasing sales revenue. They
randomly assign customers to each campaign and track their subsequent purchases.
One-way ANOVA can determine if there are significant differences in average sales
revenue across the three campaigns.

● Business questions:

○ Marketing: "Does the type of advertising campaign (TV, online, print)

significantly affect average sales revenue?"

○ Retail: "Which store layout (A, B, or C) results in the highest average

customer spending?"
 ○ Human Resources: "Is there a significant difference in employee job
satisfaction ratings across different departments (marketing, sales, operations,
etc.)?"

Scenario 2: One Independent Variable with Multiple Levels (One-Way ANOVA)

The independent variable is the factor you're manipulating or comparing (e.g., study
preparation course). Each group represents a different level of that variable. One-way
ANOVA tells you if the differences between group means are statistically significant.
● Business example: A software company is testing the effectiveness of four different
pricing models for their product. They track the number of new subscribers under
each model and use one-way ANOVA to determine if the pricing model significantly
affects subscriber acquisition.
● Business questions:
○ Product Development: "Do different product packaging designs (A, B, C, D)
significantly impact consumer purchase intention?"
○ Customer Service: "Does the length of customer service training (1 week, 2
weeks, 3 weeks) significantly affect customer satisfaction scores?"
○ Finance: "Is there a significant difference in investment returns across
different portfolio allocations (aggressive, moderate, conservative)?"

Scenario 3: Two Independent Variables (Two-Way ANOVA)

If you have two independent variables, you can use two-way ANOVA to examine not only
the individual effects of each variable but also their interaction (whether the effect of one
variable depends on the level of the other).
● Business example: An e-commerce company investigates the impact of website
design (layout A, layout B) and promotional offers (discount, free shipping, bundle)
on conversion rates. They analyze sales data from different combinations and use
two-way ANOVA to determine the optimal combination for maximizing conversions.
● Business questions:
○ Marketing: "How do different combinations of social media platforms
(Facebook, Instagram, Twitter) and advertising budgets (low, medium, high)
affect brand awareness?"
○ Sales: "Does the effectiveness of different sales training programs (A, B) vary
across different experience levels (new hires, experienced sales reps)?"
○ Operations: "What is the combined effect of different manufacturing
processes (A, B) and raw material suppliers (X, Y) on product quality?"
ANCOVA - Analysis of Covariance
1. Another Problem:
You're a researcher investigating the effectiveness of a new teaching method on students' test
scores. You randomly assign students to either the experimental group (new method) or the
control group (traditional method). However, you realize that the students in the two groups
have varying levels of prior knowledge in the subject, which could significantly influence
their test scores.

→ Can we use the ANOVA in this case?

2. Definition:
ANCOVA is an extension of ANOVA that analyzes the differences between three or more
group means while controlling for the effects of at least one continuous variable (covariates).
It is an advanced analysis of variance procedure in which the effects of one or more
metric-scaled extraneous variables are removed from the dependent variable before
conducting the ANOVA.

3. When to Use?
You should consider using ANCOVA (Analysis of Covariance) when you need to:

● Control for a continuous covariate: ANCOVA is particularly useful when you have a
continuous variable (known as a covariate) that you suspect might influence the relationship
between your independent and dependent variables. By controlling for the covariate,
ANCOVA can help you isolate the true effect of your independent variable, reducing the risk
of confounding.

○ Business Example: A retail company is testing a new store layout to increase sales.
They want to measure the impact of the new layout while accounting for the varying
sizes of their stores, as larger stores might naturally have higher sales.

○ Business Question: Does the new store layout significantly increase sales, even after
considering the size of the store?

○ Covariate: Store size (square footage)

● Increase statistical power: ANCOVA can increase the power of your analysis by accounting
for some of the variability in the dependent variable that is due to the covariate. This can
make it easier to detect statistically significant differences between groups.

○ Business Example: A pharmaceutical company is testing a new drug's effect on

blood pressure. They want to maximize their ability to detect any significant
differences between the drug and a placebo, so they control for the participants' initial
blood pressure levels, which can vary greatly.

○ Business Question: Does the new drug significantly lower blood pressure compared
to a placebo, after adjusting for the participants' baseline blood pressure?
 ○ Covariate: Baseline blood pressure

● Adjust for pre-existing differences between groups: In situations where groups are not
randomly assigned (e.g., quasi-experimental designs), ANCOVA can help adjust for
pre-existing differences between groups on the covariate. This can make the groups more
comparable and lead to more accurate results.

○ Business Example: A software company wants to compare the effectiveness of two

different onboarding programs for new employees. However, the employees in each
group have different levels of prior experience with similar software.

○ Business Question: Does one onboarding program lead to higher employee

productivity than the other, after accounting for the pre-existing differences in
experience?

○ Covariate: Prior experience with similar software (measured in years of use or

proficiency level)
Technique practicing

1. Consumer satisfaction data

2. Consumer shopping well-being
3. Data repurchase intention