RESEARCH DESIGN

Points Topic Coverage

1 Introduction to Research Design – Need & Importance

2 Features of a Good Research Design

Types of Research Designs (Exploratory, Descriptive, Experimental,

3
Analytical, Diagnostic)

4 Induction & Deduction in Research

5 Hypothesis – Meaning, Role & Formulation
Important Terms in Hypothesis (H₀, H₁, Significance, Confidence
6
Interval, Power)
Types of Research Hypotheses (Simple, Complex, Directional, Non-
7
directional, Null, Statistical)
8 Hypothesis Testing Methods – Z-test, t-test
Hypothesis Testing Methods – F-test, Decision Making in Hypothesis
9
Testing
Errors in Hypothesis Testing (Type I & II), ROC Graphics and
10
Applications
 WHAT IS RESEARCH DESIGN?

• It is a plan, structure, and strategy of investigation

conceived to obtain answers to research questions.

• It specifies:
• What data is to be collected
• From whom or where the data is to be collected
• How the data will be collected
• How the data will be analyzed
WHAT IS RESEARCH DESIGN?
 RESEARCH DESIGN
Research Design is the detailed plan or framework that
guides the entire research process — from data
collection to analysis — ensuring the research problem is
addressed systematically and effectively.
Purpose:
To provide a clear roadmap for conducting research,
ensuring validity, reliability, and efficiency.
A research design is the overall plan, structure, and
strategy of investigation conceived to obtain answers to
research questions.
It acts as a blueprint for data collection, measurement,
and analysis.
Helps in ensuring that the research problem is studied
logically, accurately, and economically.
ELEMENTS OF RESEARCH
IMPORTANCE OF RESEARCH DESIGN
IN CIVIL ENGINEERING
• Ensures accuracy and validity of results.
• Helps in efficient resource utilization (time, materials,
manpower).
• Facilitates replicability and standardization of
research.
• Supports informed decision-making in design,
construction, and maintenance.
 NEED FOR RESEARCH DESIGN

• Clarity in Direction: Provides a roadmap for the

researcher.
• Avoids Wastage of Resources: Saves time, money, and
effort by minimizing errors.
• Ensures Validity & Reliability: A good design helps in
drawing valid and reliable conclusions.
• Helps in Hypothesis Testing: Lays out procedures for
testing assumptions.
• Facilitates Data Collection & Analysis: Specifies what
data to collect, from where, and how to analyze.
• Ensures Objectivity: Minimizes bias and subjectivity.
• Guides Decision Making: Assists in choosing correct tools,
methods, and techniques.
IMPORTANCE OF RESEARCH DESIGN

• Framework for Research: Provides a structured

approach to solve a research problem.
• Improves Accuracy: Reduces chances of errors and
ensures scientific validity.
• Facilitates Smooth Operation: Clear procedures
help avoid confusion during execution.
• Ensures Logical Flow: Maintains logical sequence
from problem formulation to conclusions.
• Enhances Reliability of Results: Ensures findings can
be replicated.
 FEATURES OF A GOOD RESEARCH
DESIGN
A good research design is essential for conducting
effective and reliable research.
It ensures that the research objectives are met with
accuracy, efficiency, and clarity.
The quality of research largely depends on the
quality of its design.
The design should clearly define the research
problem, objectives, and hypotheses.
It must specify what data is needed and how it will
be collected and analyzed.
Avoids ambiguity and confusion during the
research process.
TYPES OF RESEARCH DESIGN
 TYPES OF RESEARCH DESIGN

• Exploratory Research
• Used when the problem is not clearly defined.
• Helps in gaining insights and understanding.
• Example: Investigating new construction materials.
• Descriptive Research
• Describes characteristics or functions.
• Example: Studying soil properties in a region.
• Explanatory (Causal) Research
• Examines cause-effect relationships.
• Example: Effect of load on beam deflection.
 EXPLORATORY RESEARCH DESIGN

• To explore a research problem or situation where

little information is available.
• Helps to gain insights, understand phenomena, and
formulate hypotheses or research questions.
• Characteristics
• Flexible and open-ended approach.
• Often qualitative in nature.
• Does not aim to provide conclusive answers but to
clarify concepts and gather preliminary data.
• Useful for identifying variables, generating ideas,
and understanding context.
EXPLORATORY RESEARCH DESIGN
DESCRIPTIVE RESEARCH DESIGN

• To describe characteristics, functions, or

phenomena systematically.
• Answers questions like “what,” “where,” “when,”
and “how.”
• Characteristics
• Structured and well-planned.
• Can be quantitative, qualitative, or mixed methods.
• Does not test hypotheses but provides a detailed
picture of the subject.
• Often involves large samples to generalize findings.
 INDUCTION AND DEDUCTION.

• In research, inductive reasoning moves from

specific observations to broad generalizations
and theory development in a "bottom-up"
approach,
• Deductive reasoning starts with general
theories or hypotheses and tests them with
specific observations in a "top-down"
approach.
• Inductive research is exploratory and aims to
find new insights, whereas deductive research
is focused on validating existing theories.
 HYPOTHESIS

• A hypothesis is a testable statement or a proposed

explanation for a phenomenon, acting as an
educated guess or prediction that can be proven
true or false through scientific investigation and
evidence.
• It's a fundamental part of the scientific method,
guiding research by providing a clear direction for
study and a way to connect observations to
potential outcomes. Key characteristics of a good
hypothesis include being clear, testable, verifiable,
and often structured in an "if-then" format to show a
predicted cause-and-effect relationship.
• Key aspects of a hypothesis:
• Educated guess: It's a smart guess or tentative
assumption made based on existing knowledge or
limited evidence.
• Testable and falsifiable: A hypothesis must be
capable of being tested and potentially proven
wrong by real-life evidence or experimentation.
• Directional: It provides a clear prediction about the
outcome of a study or experiment.
• Clear and specific: It should be stated clearly and
precisely before research begins.
• Basis for further study: It serves as a starting point for
deeper investigation and helps to structure the
research process.
 Z TEST HYPOTHESIS

• The Setup
• Claim: A fast-food chain claims that the mean time to order
food is 60 seconds, with a known standard deviation.

• Researcher's Action: A researcher takes a sample of 36

customers and finds the average order time was 75
seconds.

• Z-Test: The researcher uses a one-sample Z-test to determine

if the sample data contradicts the company's claim.
• Formulating the Hypotheses
• Null Hypothesis (H0): This is a statement of no effect or no
difference. In this case, it's the claim being tested: the
mean order time is 60 seconds.
• H0: μ = 60 seconds

• Alternative Hypothesis (H1): This is what the researcher is

trying to find evidence for. It could be a one-tailed
(greater than or less than) or two-tailed (not equal to)
hypothesis, depending on the question. For instance, the
researcher might want to know if service is slower than
claimed.
• Collecting Data and Calculating the Z-Statistic
• Sample Data: The researcher collected a sample of 36
customers, resulting in a sample mean of 75 seconds.

• Z-Test Formula: The Z-statistic is calculated using the formula: Z

= (sample mean - population mean) / (standard deviation /
sqrt(sample size)).

• Result: If the calculated Z-statistic is extreme enough (e.g., far

from 0), it suggests the observed sample mean is unlikely to
have occurred if the null hypothesis were true.
• Making a Decision
• Compare to Critical Value: The Z-statistic is compared to a
critical value determined by the chosen significance level
(alpha).

• Conclusion: If the calculated Z-score falls in the rejection region

(e.g., Z > 1.96 for a two-tailed test), the null hypothesis is
rejected, providing evidence for the alternative hypothesis.
Otherwise, the null hypothesis is not rejected, meaning there
isn't enough evidence to support the claim that service is
slower.
• a company claims that their new smartphone has an
average battery life of 12 hours. A consumer group
tests 100 phones and finds an average battery life of
11.8 hours with a known population standard
deviation of 0.5 hours.
KEY CHARACTERISTICS OF A Z-TEST

• Normal Distribution: The z-test is based on the

normal probability distribution.
• Known Population Standard Deviation: A crucial
requirement is the knowledge of the population
standard deviation.
• Large Sample Size: The test is most reliable when the
sample size is 30 or greater.
• Purpose: It assesses if a sample mean is significantly
different from a hypothesized population mean or if
two population means are significantly different.
T TEST

KEY TAKEAWAYS
• A t-test can shed light on a statistically
significant difference between the means
of two data sets.
• It is used for hypothesis testing in statistics.
• Calculating a t-test requires the difference
between the mean values from each data
set, the standard deviation of each group,
and the number of data values.
• T-tests can be dependent or independent.
Degree of
freedom is n-1
EXAMPLES USING T-TEST FORMULA

EXAMPLE 1: CALCULATE A T-TEST FOR THE FOLLOWING DATA OF THE

NUMBER OF TIMES PEOPLE PREFER COFFEE OR TEA IN FIVE TIME INTERVALS.
 F TEST

An F-test in hypothesis testing compares two population

variances to see if they are significantly different. The F-
statistic, calculated as the ratio of two sample
variances, follows an F-distribution. The hypotheses are
typically stated as H₀: σ₁² = σ₂² (variances are equal)
against H₁: σ₁² ≠ σ₂² (variances are unequal). If the
calculated F-statistic exceeds a critical value
determined by the significance level and degrees of
freedom, the null hypothesis is rejected, suggesting the
population variances are unequal.
F-Statistic Calculation
• The F-statistic is calculated as the ratio of the two
sample variances (s₁² and s₂²).
• Formula: F = s₁² / s₂², where s₁² is the larger variance
and s₂² is the smaller variance.
Interpretation and Decision Making
• The calculated F-statistic is compared to a critical F-
value from an F-distribution table.
• If F > Critical F-Value: The null hypothesis is rejected,
indicating there is sufficient evidence to conclude that
the population variances are different.
• If F ≤ Critical F-Value: The null hypothesis is not
rejected, meaning there is insufficient evidence to
conclude that the population variances are different.
EXAMPLE 1: A RESEARCH TEAM WANTS TO STUDY THE EFFECTS OF A NEW DRUG ON
INSOMNIA. 8 TESTS WERE CONDUCTED WITH A VARIANCE OF 600 INITIALLY. AFTER 7
MONTHS 6 TESTS WERE CONDUCTED WITH A VARIANCE OF 400. AT A SIGNIFICANCE
LEVEL OF 0.05 WAS THERE ANY IMPROVEMENT IN THE RESULTS AFTER 7 MONTHS?
EXAMPLE 2: A TOY MANUFACTURER WANTS TO GET BATTERIES FOR TOYS. A
TEAM COLLECTED 41 SAMPLES FROM SUPPLIER A AND THE VARIANCE WAS
110 HOURS. THE TEAM ALSO COLLECTED 21 SAMPLES FROM SUPPLIER B WITH
A VARIANCE OF 65 HOURS. AT A 0.05 ALPHA LEVEL DETERMINE IF THERE IS A
DIFFERENCE IN THE VARIANCES.
TYPES OF ERRORS IN HYPOTHESIS
• In hypothesis testing, the two types of
errors are Type I error, a false positive,
where a true null hypothesis is incorrectly
rejected, and Type II error, a false
negative, where a false null hypothesis is
incorrectly not rejected. These are
fundamental to statistical inference, as
using sample data to understand a
population introduces the possibility of
drawing inaccurate conclusions.
Why These Errors Occur
• These errors are inherent to hypothesis
testing because researchers use sample
data to make inferences about an entire
population, which is often impractical or
impossible to measure directly. Samples,
by their nature, may not perfectly
represent the population, leading to
errors in statistical conclusions.
 ROC GRAPHICS
• A ROC curve plots True
Positive Rate (Sensitivity) (y-
axis) against False Positive
Rate (1-Specificity) (x-axis)
across all possible test
thresholds. The curve's
position reveals a
diagnostic test's accuracy:
curves in the top-left
corner (near 0,1) are more
accurate than those closer
to the diagonal line (near
0.5), which indicates
random guessing. The Area
Under the Curve (AUC), a
single scalar value,
quantifies overall
discrimination, with 1.0
being perfect and 0.5
being no better than
chance.
• Key Components & How to Interpret Them Axes:
• Y-axis (Sensitivity): The True Positive Rate (TPR), or how
often a test correctly identifies positive cases (e.g.,
diseased).
• X-axis (False Positive Rate): The False Positive Rate (FPR),
or how often a test incorrectly identifies negative cases
as positive.

• The Diagonal Line (Line of Equality):

• A 45-degree diagonal line indicates a test with no
discriminatory ability, meaning its performance is no
better than random chance (AUC = 0.5).
The Curve's Shape:
• Top-Left Corner (0,1):
The ideal point,
representing a test
with 100% sensitivity
and 100% specificity
(or 0% false positives).
• Closer to Top-Left:
Curves situated further
from the diagonal line
and closer to the top-
left corner
demonstrate better Interpretation:
1.0: A perfect model.
predictive power and 0.9–0.99: Excellent.
discrimination 0.8–0.89: Good.
between classes. 0.7–0.79: Fair.
0.5 or less: No discriminatory ability.