UNIT 3:

Problem
Identification
& Formulation
CHEEZI PREET KAUR
CTPET/23/5184
Presentation Overview
• Problem Identification & Formulation
• Research & Investigation Questions
• Hypothesis and Hypothesis Testing
• Variable Selection and Design Matrix
• Factorial Designs (2-Level, 3-Level)
• Analysis of Variance (ANOVA)

2
Introduction
• Identifying and formulating a problem is the foundation of any
research study.
• A well-defined problem leads to precise research questions and
methodology.
• The process involves understanding the issue, defining objectives, and
setting boundaries.

3
Problem Identification
• Process of detecting a gap or issue in existing knowledge or practice.
• Involves reviewing literature, observing phenomena, or real-world
issues.
• Helps in setting research direction and objectives.

4
Problem Formulation
• Transforming a general idea into a specific, researchable problem.
• It includes narrowing the focus and clarifying what needs to be
studied.
• Should be clear, concise, and aligned with study goals.

5
Research Question vs
Investigation Question
• Research Question: Broad question guiding the entire study.
Example: What are the factors affecting employee productivity?

• Investigation Question: Specific aspects derived from the research

question.
Example: Does drinking one cup, two cups, or no coffee change how
many tasks employees complete in an hour?
Example: Does listening to music impact employee productivity?

6
Measurement Issues
• Involve defining how variables will be measured.
• Must ensure reliability, validity, and accuracy.
• Includes choosing correct instruments and scales.
Example: Measuring ‘productivity’ using a Likert scale (rating scale
which includes 5 – 7 points)

7
Hypothesis
• A tentative statement predicting a relationship between variables.
• Basis for conducting experiments and analysis.
Example: ‘Coffee intake improves employee productivity.'

8
Qualities of a Good
Hypothesis
• Clear and precise.
• Testable and falsifiable.
• Specific and limited in scope.
• Based on theoretical concepts.
• Logical and consistent.

9
Null vs Alternative
Hypothesis
• Null Hypothesis (H₀): No relationship exists. Example: H₀: µ₁ = µ₂
Coffee has no effect on productivity.

• Alternative Hypothesis (H₁): Suggests a difference or effect. Example:

H₁: µ₁ ≠ µ₂
Coffee does have an effect on productivity.

• Testing determines whether to reject H₀.

10
Visual Representation

11
Hypothesis Testing:
Logic & Importance
• Involves comparing observed data with expected outcomes under H₀.
• Steps: State hypothesis → Select test → Collect data → Analyze →
Conclude.
• Important for evidence-based decision making and theory validation.
Example: “Employees who drink two cups of coffee complete more
tasks per hour than those who drink none”.
It’s clear, testable, and based on everyday observation.

12
Selection of Variables
• Identify dependent and independent variables.
• Should align with research objectives.
• Variables must be measurable and relevant.
• Helps in designing the experiment accurately.
Example:
Independent variable: Coffee intake (0, 1, 2 cups where each cup is
300ml)
Dependent variable: Tasks completed per hour (Productivity)

13
Design Matrix
• A table showing all possible combinations of variable levels.
• Used in factorial designs to plan experiments.
• Each row represents a unique experimental condition.

Person Coffee (cups) Music (Yes/No) Tasks Completed

A 0 Yes 8
B 1 No 10
C 2 Yes 13
D 0 No 7

14
2-Level Factorial Design
• Each factor has two levels (e.g., High/Low, Yes/No).
• Efficient for studying interaction effects.
• 2 factors → 2² = 4 combinations.
• Allows examination of main and interaction effects.
Example:
Two levels of each factor:
Coffee: Yes/No
Music: Yes/No
4 combinations: Coffee + Music, Coffee only, Music only, Neither

15
3-Level Factorial Design
• Each factor has three levels (e.g., Low, Medium, High).
• Provides more detail but requires more runs.
• 2 factors → 3² = 9 combinations.
• Useful for exploring non-linear relationships.
Example:
Coffee: 0, 1, 2 cups
Music: No music, soft music, loud music
More combinations (9 total), more detailed analysis, but also more
complex.

16
Analysis of Variance
(ANOVA)
• A statistical method to compare means of three or more groups.
• Determines if observed differences are statistically significant.
• Components: Between-group variance, Within-group variance.
• Commonly used in factorial design experiments.
Example:
“Are the differences in average tasks completed between coffee levels
significant, or could they be due to chance”?

17
P-Value Interpretation
Table
P-Value What it Means Conclusion
There's less than a 5%
chance that this result is ✅ Reject H₀ → Coffee
p < 0.05 due to luck or does affect productivity
randomness.

The difference could

❌ Keep H₀ → No strong
easily have happened by
evidence coffee has an
p ≥ 0.05 chance. We keep the null
effect
hypothesis (H₀).

18
Analogy Table: Coin Toss
vs. Coffee Effect
Situation Expected Result Observed Result Interpretation
That’s too unlikely
to be random.
~5 heads out of
Tossing a fair coin 9 heads Maybe the coin is
10
biased.

That’s too unlikely

to be a
Coffee vs. No Similar Coffee: 12, No coincidence —
Coffee test productivity Coffee: 7 coffee probably
helps.

19
Thank You

20