Research Methods !

class: title-slide

.row[
.col-7[
.title[
# Research Methods 
]
.subtitle[
## Research Methods
]
.author[
### Laxmikant Soni [blog](https://laxmikants.github.io) [](https://github.com/laxmiaknts) [](https://twitter.com/laxmikantsoni09)
]

.affiliation[
]

]

.col-5[

.logo[

<img src="rmarkdown.png" width="480" />
]

]
]

---

# Understanding Research

.pull-left[

## Meaning of Research

* **Definition**: Research is a systematic investigation aimed at discovering new information, verifying existing knowledge, or solving a specific problem.
* **Key Concepts**:
  - **Systematic Approach**: Involves structured methods and procedures.
  - **Knowledge Expansion**: Enhances understanding in various fields.
  - **Problem-Solving**: Provides solutions to practical and theoretical problems.
* **Purpose**: The primary goal of research is to generate reliable and valid knowledge that can be applied in real-world contexts.

]

.pull-right[

## Objectives of Research

* **Primary Objectives**:
  - **Exploration**: To explore new areas and generate initial insights.
  - **Description**: To describe characteristics of a specific phenomenon or population.
  - **Explanation**: To explain the reasons behind observed phenomena.
  - **Prediction**: To forecast future occurrences based on current data.
  - **Application**: To apply findings for practical problem-solving.
* **Example**: A study aimed at identifying the factors affecting student performance in online learning environments.

]

---

# Understanding Research

.pull-left[

## Types of Research

* **Classification by Purpose**:
  - **Basic Research**: Focuses on expanding fundamental knowledge without immediate practical application.
  - **Applied Research**: Aimed at solving specific, practical problems.
* **Classification by Method**:
  - **Quantitative Research**: Involves numerical data and statistical analysis (e.g., surveys, experiments).
  - **Qualitative Research**: Focuses on non-numerical data, such as interviews and observations, to understand concepts, experiences, or social contexts.
  - **Mixed Methods**: Combines both qualitative and quantitative approaches to provide a comprehensive view.

]

.pull-right[

## Significance of Research

* **Key Contributions**:
  - **Advances Knowledge**: Enhances theoretical and practical understanding in various fields.
  - **Supports Decision-Making**: Provides evidence-based insights for policy-making, business strategies, and societal development.
  - **Solves Real-World Problems**: Offers solutions to challenges in health, education, technology, and more.
  - **Fosters Innovation**: Drives technological advancements and new product development.
* **Example**: Research in healthcare can lead to the discovery of new treatments, improving patient outcomes.

]

---

# Research Problem

.pull-left[

## Definition of Research Problem

* **Definition**: A research problem is a specific issue, difficulty, or gap in existing knowledge that a researcher aims to address through a systematic investigation.
* **Key Concepts**:
  - **Clarity**: The problem should be clearly defined and well-articulated.
  - **Relevance**: It must be significant and relevant to the field of study.
  - **Feasibility**: The problem should be researchable within the available resources and time frame.
* **Example**: Identifying the reasons behind the high dropout rates in online education.

]

.pull-right[

## Necessity of Defining a Research Problem

* **Importance**:
  - **Guides the Research**: Provides direction and focus to the study.
  - **Sets the Scope**: Helps in determining the boundaries of the research.
  - **Ensures Relevance**: Ensures that the research addresses a significant issue.
* **Benefits**:
  - **Efficient Use of Resources**: Optimizes time, effort, and resources by targeting specific issues.
  - **Improves Outcomes**: Leads to more meaningful and impactful research findings.

]

---

# Research Problem

.pull-left[

## Techniques for Defining a Research Problem

* **Key Techniques**:
  - **Literature Review**: Analyze existing research to identify gaps or unexplored areas.
  - **Brainstorming**: Generate multiple ideas and refine them to a specific problem.
  - **Stakeholder Analysis**: Understand the needs and challenges faced by relevant stakeholders.
  - **SWOT Analysis**: Evaluate the strengths, weaknesses, opportunities, and threats related to the topic.
* **Example**: Conducting interviews with industry experts to identify pressing challenges in sustainable agriculture.

]

.pull-right[

## Formulation of a Research Problem

* **Steps for Formulation**:
  1. **Identify the Broad Area**: Start with a general topic of interest.
  2. **Narrow Down the Focus**: Specify the issue within the broader context.
  3. **State the Problem Clearly**: Use precise language to define the problem.
  4. **Set Boundaries**: Define the scope and limitations of the study.
* **Outcome**: A well-formulated research problem that serves as the foundation for developing hypotheses and research questions.

]

---

# Research Problem

.pull-top[

## Objectives of a Research Problem

* **Primary Objectives**:
  - **Understanding the Issue**: Gain deeper insights into the problem's causes and effects.
  - **Finding Solutions**: Develop strategies or interventions to address the issue.
  - **Contributing to Knowledge**: Add to the academic or practical knowledge in the field.
  - **Informing Policy and Practice**: Provide evidence to support decision-making and policy formulation.
* **Example**: Aiming to reduce employee turnover by studying factors affecting job satisfaction.

]

---

# Research Design

.pull-left[

## Meaning of Research Design

* **Definition**: Research design is a structured framework or blueprint that guides the entire research process, from data collection to analysis.
* **Key Concepts**:
  - **Structure**: Provides a clear plan for conducting the study.
  - **Purpose**: Ensures that the research problem is addressed efficiently and effectively.
  - **Validity**: Enhances the reliability and validity of the research findings.
* **Example**: Using a randomized controlled trial (RCT) to test the effectiveness of a new drug.

]

.pull-right[

## Need for Research Design

* **Importance**:
  - **Guides the Research Process**: Provides a systematic approach to answer research questions.
  - **Minimizes Errors**: Reduces biases and ensures accurate data collection.
  - **Enhances Reliability**: Leads to replicable and consistent results.
* **Benefits**:
  - **Optimal Resource Allocation**: Helps in the efficient use of time, budget, and resources.
  - **Clear Focus**: Keeps the research aligned with its objectives and scope.

]

---

# Research Design

.pull-left[

## Features of a Good Research Design

* **Key Features**:
  - **Clarity**: Clearly defined research problem and objectives.
  - **Flexibility**: Ability to adapt to changes during the research process.
  - **Control**: Mechanisms to reduce the impact of extraneous variables.
  - **Validity**: Ensures internal and external validity to support the generalization of results.
  - **Ethical Considerations**: Ensures the ethical treatment of participants and data.
* **Example**: A well-structured survey design that includes random sampling to enhance validity.

]

.pull-right[

## Types of Research Designs

* **Classification by Purpose**:
  - **Exploratory Research**: Aims to explore a phenomenon without prior knowledge.
  - **Descriptive Research**: Focuses on describing characteristics or functions.
  - **Explanatory (Causal) Research**: Investigates cause-and-effect relationships.
* **Classification by Method**:
  - **Experimental Design**: Involves manipulation of variables to test hypotheses.
  - **Observational Design**: No manipulation; relies on observing subjects in natural settings.
  - **Cross-Sectional Design**: Studies a sample at one point in time.
  - **Longitudinal Design**: Observes the same subjects over an extended period.

]

---

# Research Design

.pull-top[

## Basic Principles of Experimental Designs

* **Key Principles**:
  - **Replication**: Repeating experiments to confirm findings.
  - **Randomization**: Random assignment of subjects to control and experimental groups to eliminate biases.
  - **Control**: Use of control groups to compare results.
* **Benefits**:
  - **Increases Accuracy**: Enhances the precision of the experimental results.
  - **Reduces Bias**: Mitigates the influence of confounding variables.
* **Example**: A double-blind clinical trial where neither the participants nor the researchers know who receives the treatment.

]

---

# Research Design

.pull-top[

## Design of Experiments

* **Steps in Designing an Experiment**:
  1. **Define the Objective**: Clearly state the aim of the experiment.
  2. **Select Factors and Levels**: Identify independent variables and their levels.
  3. **Choose the Experimental Design**: Options include completely randomized design, factorial design, etc.
  4. **Randomize Assignments**: Allocate subjects to different groups randomly.
  5. **Collect and Analyze Data**: Use statistical methods to interpret results.
* **Types of Experimental Designs**:
  - **Completely Randomized Design**: Subjects are randomly assigned to all experimental conditions.
  - **Factorial Design**: Examines the effect of two or more factors simultaneously.
  - **Block Design**: Subjects are divided into blocks based on certain characteristics before being randomly assigned.
* **Example**: Using a factorial design to study the interaction effects of different fertilizers and watering schedules on crop yield.

]

---

# Example Research Problem: Performance Evaluation of Library Code

.pull-top[

## Definition of Research Problem

* **Definition**: Evaluating the performance of a specific software library (e.g., a sorting algorithm library) to identify its efficiency, speed, and resource utilization.
* **Key Concepts**:
  - **Performance Metrics**: Focuses on metrics like execution time, memory usage, and CPU utilization.
  - **Optimization**: Aims to optimize the code or choose the best-performing library for a given use case.
* **Example**: Analyzing the performance of different Python libraries for matrix multiplication (e.g., NumPy vs. TensorFlow).

]

---

# Example Research Problem: Performance Evaluation of Library Code

.pull-top[

## Necessity of Defining a Research Problem

* **Importance**:
  - **Guides the Research**: Provides a clear focus on evaluating specific performance aspects.
  - **Sets the Scope**: Helps in determining which performance metrics to measure and under what conditions.
  - **Ensures Relevance**: Addresses practical issues like reducing computation time in software applications.
* **Benefits**:
  - **Efficient Use of Resources**: Targets specific areas for performance improvement, optimizing time and computational resources.
  - **Improves Software Efficiency**: Leads to more efficient code, benefiting both developers and end-users.

]

---

# Example Research Problem: Performance Evaluation of Library Code

.pull-top[

## Techniques for Defining a Research Problem

* **Key Techniques**:
  - **Literature Review**: Examine existing performance studies on similar libraries to identify gaps.
  - **Benchmarking**: Conduct preliminary benchmarks to identify performance issues in the library.
  - **User Feedback**: Collect insights from developers who use the library in real-world projects.
  - **SWOT Analysis**: Evaluate the strengths, weaknesses, opportunities, and threats of the library's performance.
* **Example**: Running initial tests to determine if a popular library has performance bottlenecks when handling large datasets.

]

---

# Example Research Problem: Performance Evaluation of Library Code

.pull-top[

## Formulation of a Research Problem

* **Steps for Formulation**:
  1. **Identify the Broad Area**: Focus on software performance evaluation.
  2. **Narrow Down the Focus**: Specify the library and performance metrics of interest (e.g., execution time, memory usage).
  3. **State the Problem Clearly**: Define the performance evaluation objective, such as "Evaluating the execution speed of NumPy for large-scale matrix operations."
  4. **Set Boundaries**: Define the scope, such as testing under specific hardware configurations or dataset sizes.
* **Outcome**: A well-defined research problem to guide benchmarking and optimization efforts.

]

---

# Example Research Problem: Performance Evaluation of Library Code

.pull-top[

## Objectives of a Research Problem

* **Primary Objectives**:
  - **Performance Measurement**: Assess the library's efficiency in different scenarios.
  - **Comparison**: Compare the performance of multiple libraries to identify the best option.
  - **Optimization**: Provide recommendations for improving the library's performance.
  - **Documentation**: Generate detailed performance reports for developers.
* **Example**: Comparing the execution time of sorting algorithms in different Python libraries (e.g., `sorted()` vs. `numpy.sort()`).

]

---

# Research Design: Performance Evaluation of Library Code

.pull-top[

## Meaning of Research Design

* **Definition**: A structured plan to systematically evaluate the performance of the library code, covering aspects like data collection, testing, and analysis.
* **Key Concepts**:
  - **Structure**: Includes steps like setting up benchmarks, running tests, and analyzing results.
  - **Purpose**: To ensure accurate and reliable evaluation of the library's performance.
  - **Validity**: Ensures that the performance tests reflect real-world usage scenarios.
* **Example**: Designing a series of benchmarks to test library performance on different data sizes.

]

---

# Example Research Problem: Performance Evaluation of Library Code

.pull-top[

## Need for Research Design

* **Importance**:
  - **Guides the Evaluation**: Provides a structured approach to assessing library performance.
  - **Minimizes Errors**: Reduces biases in testing conditions and data selection.
  - **Enhances Reliability**: Ensures replicable and consistent performance results.
* **Benefits**:
  - **Optimal Use of Resources**: Efficiently utilizes testing environments and computational power.
  - **Clear Focus**: Keeps the evaluation aligned with specific performance goals.

]

---

# Example Research Problem: Performance Evaluation of Library Code

.pull-top[

## Features of a Good Research Design

* **Key Features**:
  - **Clarity**: Clearly defined performance metrics and testing scenarios.
  - **Flexibility**: Allows for adjustments in testing conditions based on initial results.
  - **Control**: Controls for variables like hardware specifications and background processes.
  - **Validity**: Uses standard benchmarks to ensure the accuracy of results.
  - **Ethical Considerations**: Ensures no misuse of resources or intellectual property.
* **Example**: A robust benchmarking design that includes testing the library on different hardware setups.

]

---

# Example Research Problem: Performance Evaluation of Library Code

.pull-top[

## Types of Research Designs

* **Classification by Purpose**:
  - **Exploratory Research**: Investigates unknown performance issues in a newly released library version.
  - **Descriptive Research**: Provides a detailed report on the library's performance metrics.
  - **Explanatory (Causal) Research**: Examines how changes in code affect the library's performance.
* **Classification by Method**:
  - **Experimental Design**: Compares different optimization techniques on the library.
  - **Observational Design**: Monitors the library's performance in real-world applications.
  - **Cross-Sectional Design**: Tests library performance at a single point in time with various datasets.
  - **Longitudinal Design**: Evaluates performance changes over multiple library updates.

]

---

# Example Research Problem: Performance Evaluation of Library Code

.pull-top[

## Basic Principles of Experimental Designs

* **Key Principles**:
  - **Replication**: Re-run performance tests on different datasets to verify consistency.
  - **Randomization**: Randomly select datasets and test scenarios to avoid biases.
  - **Control**: Use control variables like fixed hardware and software environments.
* **Benefits**:
  - **Increases Accuracy**: Provides a precise measurement of performance differences.
  - **Reduces Bias**: Ensures unbiased results by controlling external factors.
* **Example**: Comparing the performance of library code across different Python versions.

]

---

# Example Research Problem: Performance Evaluation of Library Code

.pull-top[

## Design of Experiments

* **Steps in Designing an Experiment**:
  1. **Define the Objective**: Evaluate the performance of a specific library function.
  2. **Select Factors and Levels**: Factors could include dataset size, hardware configuration, or optimization flags.
  3. **Choose the Experimental Design**: Use a factorial design to test different combinations of factors.
  4. **Randomize Assignments**: Randomly assign test datasets to different performance scenarios.
  5. **Collect and Analyze Data**: Use performance metrics like execution time and memory usage to analyze results.
* **Types of Experimental Designs**:
  - **Completely Randomized Design**: Randomly assign datasets to different performance tests.
  - **Factorial Design**: Explore interactions between factors like dataset size and hardware.
  - **Block Design**: Group tests based on software versions before evaluating performance.
* **Example**: An experiment to test the impact of different compiler optimizations on library execution speed.

]

---

class: inverse, center, middle
# Thanks