Research

What is Research

Rigour and Relevance in Research :

Rigour in research ensures methodological accuracy and reliability, while relevance ensures the study
addresses real-world problems and practical needs.

P Focus Example
Purpose Why you're doing the research Understand students’ digital fatigue
Problem What issue you're exploring Lack of insight into emotional effects
Paradigm Your philosophical approach Interpretivism
Participants Who is involved 20 college students
Processes How data is collected and analyzed Interviews + thematic analysis
Presentation How you report your findings Journal article or thesis

Paradigm types(student performance in online classes)

A positivist would use statistics to measure scores and analyze cause-effect (e.g., Does screen
time reduce grades?).

 An interpretivist would explore how students feel and experience online learning (e.g., What
emotions do students associate with Zoom classes?).

 A critical theorist might examine if low-income students are disadvantaged in online settings
due to digital inequality.
Element Purpose Example
Research What are you trying to What is the impact of social media on academic
Question find out? performance?
Research What is your research aiming To examine the effect of social media use on
Objective to do? academic scores.
What is your H₁: Increased social media use reduces academic
Hypothesis
prediction? performance.

Hypothesis (mainly for quantitative research)

A hypothesis is a testable statement that predicts a relationship between variables.

It is either:

 Null hypothesis (H₀): No relationship or effect

 Alternative hypothesis (H₁): There is a relationship or effect

Example:

 H₀ (Null): There is no significant relationship between social media use and academic
performance.
 H₁ (Alternative): Higher social media use negatively affects academic performance.

LITERATURE REVIEW:
 1. Choose a topic. Define your research question.
2. Decide on the scope of your review
3. Select the database you will use to conduct your searches
4. Conduct your searches and find the literature. Keep track of your searches.
5. Review the literature
6. Organizing/Writing

Example of literature review:

https://libguides.uwf.edu/ld.php?content_id=42292575

https://libguides.uwf.edu/ld.php?content_id=42292186

https://writingcenter.unc.edu/tips-and-tools/literature-reviews/

RESEARCH FRAMEWORK DESIGN AND EXPERIMENTATION

FRAMEWORK

SYSTEMS DIAGRAM

Closed system and Open System

Summary of system diagram Rules (Suggested by PAS)

Mathmatical Modeling ?Why?

RESEARCH FRAMEWORK DESIGN AND EXPERIMENTATION – Summary

A research framework is the backbone of a study. It outlines the structure, components, and processes
that guide your research from start to finish. The design phase involves identifying the variables,
methods, and tools to be used, while experimentation focuses on testing hypotheses and collecting
data under controlled conditions.

It serves as a roadmap that ensures:

 Clarity in research goals

 Alignment between objectives and methods
 Proper data collection and interpretation

The framework helps in organizing your research in a logical and systematic way and also supports
repeatability and transparency of the study.
📐 SYSTEMS DIAGRAM

A systems diagram is a visual representation of how different components within a system are related.
It shows how inputs are transformed into outputs through various processes, often including feedback
loops to indicate learning or change over time.

Components of a Systems Diagram:

1. Inputs – Elements that enter the system (e.g., resources, data, energy)
2. Processes – Actions or transformations (e.g., teaching, decision-making)
3. Outputs – The result or outcome (e.g., student performance, product quality)
4. Feedback – Information returned to improve the process
5. Boundaries – What’s inside the system and what lies outside

This diagram is useful in both research planning and communication as it gives a clear, holistic view of
how a system works.

🔒🔓 Closed and Open Systems

In system thinking, it's important to know whether you're dealing with a closed or an open system.

🔹 Closed System:

 Does not interact with the environment.

 Self-contained with no external inputs or outputs.
 Easier to study and model mathematically.
 Example: A chemical reaction in a sealed beaker.

🔹 Open System:

 Interacts with its environment.

 Exchanges information, energy, or materials.
 Dynamic and adaptable.
 Example: A school, which interacts with students, parents, policies, and society.

Most real-world systems, especially in social sciences and business, are open systems.

📌 Summary of System Diagram Rules (PAS - Publicly Available

Specification)

PAS provides guidelines for designing clear and standardized systems diagrams. These help maintain
consistency, readability, and usefulness.
Key Rules:

1. Clearly define the system boundary – Know what’s included in the system and what’s outside.
2. Use consistent symbols and shapes for components (e.g., rectangles for processes, arrows for
flow).
3. Label all components clearly – no vague or confusing names.
4. Show the direction of flow using arrows (input → process → output).
5. Include feedback loops if the system uses information to adjust itself.
6. Avoid clutter – keep it simple and focused.
7. Use logical sequencing – components should follow a cause-effect structure.

Following these rules helps others understand and use your system diagram more effectively.

🔢 Mathematical Modeling – Why Use It in Research?

Mathematical modeling means representing a real-world system using mathematical concepts such as
equations, graphs, and algorithms.

Why it’s important:

1. Accuracy and Objectivity – Converts complex processes into precise language.

2. Prediction and Forecasting – Allows researchers to simulate future outcomes (e.g., spread of
disease, economic trends).
3. Problem Solving – Helps test “what if” scenarios without needing real-world experiments.
4. Optimization – Helps identify the best possible outcome (e.g., lowest cost, maximum profit).
5. Communication – Provides a clear, standard way to present ideas to others in the field.

In research, mathematical models support decision-making and help validate theories through testing.
RESEARCH AND STRATEGIES

Surveys

Design and Creation

Experiment

Case Studies

Action Research

Ethnography

Surveys Collecting data from many people Quantitative / Qualitative Student

feedback on online classes

Design & Creation Building and testing new systems/models Applied / MixedDeveloping a
study planner app

Experiment Testing variables in controlled settings Quantitative New vs. traditional teaching
method

Case Study In-depth study of a specific case Qualitative / Mixed Studying one school’s digital
transformation

Action Research Improving practice with participants involved Qualitative Teachers

improving lesson delivery

Ethnography Observing and understanding cultures Qualitative Living in a village to study its
cultural traditions

Sampling technique

Probability Sampling Techniques

These techniques give every member of the population a known and equal chance of being selected.
They are more scientific and unbiased.
1. 🌀 Random Sampling

 Every individual in the population has an equal chance of being selected.

 Like drawing names from a hat.
 Best when the population is homogeneous (similar characteristics).

Example: You have 1,000 students. You randomly select 100 using a random number generator.

2. 📏 Systematic Sampling

 Start at a random point, then select every nth individual from a list.
 Still random, but with a pattern.

Example: Start at the 5th person on a list, then pick every 10th person.

3. 🧬 Stratified Sampling

 The population is divided into strata (subgroups) based on characteristics (e.g., gender, age,
department).
 Samples are drawn proportionally from each stratum.

Example: If your population is 60% female and 40% male, your sample of 100 would include 60 females
and 40 males.

4. 🏘️Cluster Sampling (Explained More)

This is useful when the population is geographically spread out or naturally divided into groups
(clusters), like schools, districts, or cities.

✅ Key Points:

 You randomly select whole clusters rather than individuals.

 Then you either:
o Survey everyone in those clusters, or
o Do random sampling within those clusters.

📌 When to Use:

 When it’s too costly or impractical to survey the whole population.

  When natural groupings (clusters) already exist.

📊 Example:

Suppose you want to study high school students’ opinions across a country:

 The country has 500 high schools.

 You randomly select 10 schools (clusters).
 You survey all students in those 10 schools (or just a sample from each).

It’s like zooming into small areas (clusters) instead of trying to cover the entire field.

🧠 Comparison: Stratified vs. Cluster

Feature Stratified Sampling Cluster Sampling

Division Basis By traits (e.g., gender, age, income) By geography/group (e.g., schools, cities)

Sample Select some individuals from every

Select whole groups and survey all/part in them
Selection group

60% boys, 40% girls → Sample reflects Pick 5 random schools, survey students in those
Example
that only

Less accurate, but more practical and

Cost More accurate, but may be expensive
affordable
Non-Probability Sampling Techniques

Unlike probability sampling, not every individual has a known or equal chance of being selected. These
methods are often used in qualitative research or when the population is hard to access.

🔸 1. Purposive Sampling (Judgmental Sampling)

 The researcher deliberately selects participants who have specific characteristics or knowledge.
 Based on the researcher’s judgment of who will be most useful or relevant.

✅ Example: If you're studying expert opinions on climate change, you purposely choose environmental
scientists.

🔸 2. Snowball Sampling

 You start with one or a few participants, then ask them to refer others who fit the criteria.
 Useful for hard-to-reach or hidden populations (e.g., drug users, underground artists, etc.).

✅ Example: You interview one artist, then ask them to suggest another, and it grows like a snowball.

🔸 3. Self-Selection Sampling

 The researcher invites people to participate, and whoever is interested or willing responds
themselves.
 Participants choose to be part of the study.

✅ Example: An online survey is shared on social media. People who are interested click and fill it out.

🔸 4. Convenience Sampling

 The researcher selects people who are easy to access or nearby.

 It’s quick, easy, and often used for pilot studies or early-stage research.

✅ Example: Surveying students in your own class because they’re available.

🧠 Summary Table:
Technique How it works When to use it Example

Purposive Handpick participants based on Studying experts or Selecting 10 software

Sampling purpose or knowledge specific traits engineers for a tech study

Snowball Ask one participant to refer Studying hidden or small Research on rare disease
Sampling others like them groups patients

Self-Selection Public invitation; whoever When wide voluntary Online survey on fitness
Sampling responds is part of the study input is acceptable shared on Instagram

Convenience Select participants who are For quick feedback or Asking your classmates to
Sampling easy to reach pilot studies fill out a survey

This slide from a presentation on survey research strategies is hitting on two key aspects you
need to think about: response rate and non-responses, and sample size.

Regarding response rate and non-responses, while the slide doesn't give specifics, it's flagging
that these are important considerations. Think about it: if you send out a bunch of surveys but
only a tiny fraction get filled out and returned (low response rate), or if certain types of people
consistently don't respond (non-response bias), the results you get might not accurately reflect
the bigger picture you're trying to understand.

Then there's sample size. The slide makes a couple of really crucial points here:

 At least 30: This is a bit of a rule of thumb. Generally, having at least 30 data points
(completed surveys, in this case) makes it more feasible to do meaningful statistical
analysis. With a smaller sample, it becomes much harder to draw reliable conclusions.
 Adequate size for generalization: This is where it gets more nuanced. If your goal is to
take what you learn from your survey and apply it to a larger group of people (the "whole
population"), then your sample needs to be large enough and representative enough of
that population. The size needed depends on how large the population is and how much
variability there is within it.

This slide builds on the importance of sample size and digs into how we talk about the accuracy
of our survey results when we want to apply them to a larger group. It's all about accuracy
range and confidence level.

Think of it this way: when you survey a sample of people, you're not asking everyone. So, the
results you get are an estimate of what's true for the entire population.

 Accuracy range (also called 'margin of error' and 'confidence interval'): This tells
you how much your sample results might differ from the actual truth in the whole
population. For example, if a survey says 60% of people prefer a certain brand with a
margin of error of +/- 3%, it means the actual percentage in the entire population is likely
 somewhere between 57% and 63%. It's a range within which we're reasonably confident
the real answer lies.
 Confidence level: This tells you how sure you are that the true population value falls
within that accuracy range. The slide specifically mentions a confidence level of 95%.
What that means is, if you were to do the same survey over and over again with different
random samples, about 95% of the time, the true population value would fall within the
margin of error you calculated for each of those surveys.

this table really brings the previous points about sample size, confidence level, and accuracy
range together! Let's break down what it's showing us.

Table 1: Target population sizes and sample sizes

The table gives you a concrete idea of how large your sample needs to be depending on the size
of the total population you're interested in. It's specifically set up for a 95% confidence level and
a +/- 3% accuracy range (margin of error). Remember from the last slide, a 95% confidence
level means you're pretty darn sure (95% of the time) that the true population value falls within
that +/- 3% range of what your sample shows.

Let's look at the numbers:

 If your target population is small (say, 50 people): You actually need to survey almost everyone
(47 people) to get that 95% confidence with a +/- 3% margin of error. Makes sense, right? When
the total group is small, your sample needs to be a big chunk of it to be representative.
 As your target population grows (to 5,000 people): The required sample size jumps up to 760.
 When your target population gets really large (like 100,000 or even 900,000): Notice
something interesting? The required sample size increases, but not by nearly as much! It goes
from 760 to 888, and then only to 895.

Key Takeaway (as highlighted in the note):

The note at the bottom is super important: "The sample size does not need to increase at the
same rate as the target population size."

This is a really insightful point! Once your population is large enough, the size of the total
population becomes less of a driving factor in determining the needed sample size (for a specific
confidence level and margin of error). What becomes more important is the absolute size of your
sample.

In simple terms: Whether you're trying to understand the opinions of 100,000 people or
1,000,000 people, if you want a 95% confidence level with a +/- 3% margin of error, your
required sample size will be surprisingly similar (around 800-900 in this example).

Tool:
The slide even gives you a handy tool: www.surveysystem.com/sscalc.htm. You can plug in
different population sizes, desired confidence levels, and margins of error to see how the
required sample size changes. It's a great way to play around with these concepts and get a better
feel for how they relate to each other.

So, this table is a practical illustration of how to determine an appropriate sample size for your
survey research when you have a specific level of accuracy and confidence in mind for
generalizing your findings.