Scientific Research & Introduction to Research
What is Science?
A systematic, organized body of knowledge acquired using the scientific method. Divided into Natural Science (physics, chemistry, biology, earth sciences) and Social Science (psychology, sociology, economics).
Also classified by purpose: Basic/Pure Science (explains fundamental laws) vs Applied Science (applies knowledge to solve real problems).
What is Research?
A careful, systematic investigation aimed at discovering new knowledge or validating existing knowledge. It is a voyage of discovery — movement from the known to the unknown.
Research is NOT: mere information gathering, simple transportation of facts, self-enlightenment, or a catchphrase.
8 Characteristics of Good Research
- Originates with a question or problem
- Requires clear articulation of the problem
- Follows a specific plan or procedure
- Divides the problem into manageable sub-problems
- Guided by objectives, questions, or hypotheses
- Accepts critical assumptions
- Involves collection, analysis, and interpretation of data
- Is cyclical/helical in nature
Purposes of Research
- Discover new knowledge
- Gain familiarity (Exploratory)
- Portray accurately (Descriptive)
- Enable prediction (Hypothesis-testing)
- Enable control
- Theory development
Scientific Method Characteristics
- Replicability – others can repeat it
- Precision – concepts clearly defined
- Parsimony – simplest explanation preferred
- Falsifiability – must be testable/disprovable
Inductive vs Deductive Research
Data → Pattern → Theory. "Bottom-up." Used when few prior theories exist. Theory-building.
Theory → Hypothesis → Data. "Top-down." Tests existing theories against new data. Theory-testing.
The Research Problem
What is a Research Problem?
A difficulty, unanswered question, or unresolved matter the researcher wants to solve. It must lead to new knowledge — not mere information gathering.
Sources of Research Problems
- Reading the literature (finding research gaps)
- Attending professional conferences
- Seeking advice from experts
- Practical field problems
- Request for Proposals (RFPs)
- Personal experience/background
Guidelines for Selecting a Problem
- Must lead to new knowledge (not self-enlightenment)
- Avoid simple comparisons or yes/no answers
- Must sustain your interest over time
- Avoid emotionally loaded or overly ambitious topics
- Consider feasibility: time, cost, ethical constraints
Components of Chapter 1 – Introduction
- Background to the Study – global/regional/national context (3 steps: general → specific → gap)
- Problem Statement – precise, complete, in grammatically correct sentences
- Operational Definitions – define all key terms as used in the study
- Purpose of the Study – broad declarative statement
- Research Objectives / Questions / Hypotheses – SMART criteria
- Scope & Delimitations – what is and is not included
- Assumptions – facts taken for granted
- Limitations – constraints beyond researcher's control
- Significance of the Study – who benefits and how
Dividing into Sub-Problems
Each sub-problem must be: (1) completely researchable on its own, (2) clearly tied to data interpretation, (3) small in number (3–5), and (4) add up to the totality of the main problem.
Sub-problems are expressed as Objectives (declarative), Questions (interrogative), or Hypotheses (testable).
Research Topic & Research Title
Topic vs Title
Topic: The broad subject area of study.
Title: A concise phrase that summarises the specific focus and objective of the project.
Length
Rule of thumb: 12–15 words maximum. Summarises aim and purpose economically.
Elements of a Good Title
- Topic – the general area
- Method – how the research will be done (optional)
- Sample/Population – who is studied
- Outcome/Product – e.g. a model, framework, system, algorithm
Title Don'ts
- No periods, dashes, or question marks
- No acronyms or abbreviations
- No vague terms, technical jargon, or uncommon words
- No openings like "A Study of…", "Results of…"
- No italics or exclamation marks
Research Hypotheses
What is a Hypothesis?
A tentative, intelligent guess that directs thinking towards a solution. It is an educated conjecture that provides a tentative explanation for a phenomenon. Hypotheses are not proved or disproved — they are supported or not supported by data.
Characteristics of a Good Hypothesis
- Clear and precise
- Capable of being tested
- States the relationship between variables
- Limited in scope (narrow = more testable)
- Consistent with established facts
- Testable within reasonable time
Null Hypothesis (H₀)
States no difference between groups or no relationship. Used in statistical testing — we seek to reject it. Represents what we wish to disprove.
Alternative Hypothesis (H₁)
States that a difference or relationship does exist. If H₀ is rejected, H₁ is accepted. Represents what we wish to prove.
Types of Errors in Hypothesis Testing
| Error Type | Also Called | What it Means |
|---|---|---|
| Type I Error | False Positive | Rejecting H₀ when it is actually true |
| Type II Error | False Negative | Failing to reject H₀ when it is actually false |
Literature Review
What is a Literature Review?
Systematic identification, location and analysis of documents related to the research problem. It is not a chain of isolated summaries — it evaluates, analyses, and synthesises.
Purposes of the Literature Review
- Avoids unnecessary duplication
- Forms the Conceptual Framework
- Demonstrates familiarity with existing knowledge
- Suggests methodologies and approaches
- Reveals data sources and measurement tools
- Helps interpret findings
- Identifies research gaps
Structure of the LR (Inverted Pyramid)
- Historical Development (optional)
- Theoretical Review – underpinning theories
- Empirical Review – what others have found
- Research Gap(s) – what is yet to be done
- Conceptual Framework – researcher's mind map of variables
- Summary
Types of Literature Reviews
Narrative: Traditional; critiques and summarises; identifies gaps. Systematic: Rigorous, reproducible, uses explicit inclusion criteria. Scoping: Maps the breadth of a topic, often for emerging areas.
Plagiarism
Using someone else's exact words or ideas without referencing. Avoid by paraphrasing, summarising, and citing correctly. Detected using software (Turnitin, Urkund). Similarity index at KCA ≤ 15%.
Systematic Reviews
Key Definitions
Research Synthesis: A review of primary research to integrate findings and create generalisations.
Systematic Review: Uses explicit, reproducible methods to identify, select, appraise, and analyse all relevant research on a narrow, clearly defined question.
Meta-Analysis: Statistical methods for combining effect sizes across multiple studies addressing the same question.
Meta-Synthesis: Non-statistical integration of qualitative research findings.
Steps in a Systematic Review
- Formulate the research question (use PICO: Population, Intervention, Comparison, Outcome)
- Develop a review protocol
- Conduct a scoping search, then a full search
- Screen for inclusion/exclusion
- Assess quality of primary studies
- Extract data
- Analyse and synthesise (quantitative/qualitative)
- Interpret and present results (PRISMA flow diagram)
Research Design & Methodology
Research Design Defined
The complete strategy of attack on the central problem — the blueprint for data collection, measurement, and analysis. "Plan the flight, then fly the plan."
Research Methodology (the general approach) ≠ Research Design (the specific methods used).
Key Concepts
- Target Population: The group the researcher wants to generalise to
- Accessible/Survey Population: The reachable, representative subset
- Sample: A smaller group drawn from the accessible population
- Sampling Error: Discrepancy between sample and population characteristics (smaller sample = bigger error)
- Primary Data: Collected from the field (most valid)
- Secondary Data: Derived from existing sources
Validity of Measurement Instruments
| Type | Meaning |
|---|---|
| Face Validity | Looks like it measures what it should |
| Content Validity | Representative sample of the content domain |
| Criterion Validity | Correlates with a related external measure |
| Construct Validity | Measures an unobservable underlying construct |
Reliability of Measurement Instruments
- Inter-Rater: Two raters give identical judgements
- Inter-Consistency: All items in one instrument yield similar results
- Equivalent Forms: Two versions of the same instrument give similar results
- Test-Retest: Same instrument gives same results on two occasions
Method Validity
Internal Validity: Allows accurate conclusions about cause-and-effect within the data. Enhanced by lab controls, double-blind experiments, unobtrusive measurement, triangulation.
External Validity: Extent to which results generalise beyond the study. Enhanced by real-life settings, representative samples, replication in different contexts.
Ethical Issues in Research
- Protection from harm – no undue physical/psychological risk
- Informed consent – voluntary participation, right to withdraw
- Right to privacy – anonymity and confidentiality
- Honesty with colleagues – no fabrication of data, credit others' work
Writing the Research Proposal
Structure of a Research Proposal (5 Chapters)
- Chapter 1 – Introduction: Background, problem statement, definitions, purpose, objectives/questions/hypotheses, scope, limitations, significance
- Chapter 2 – Literature Review: Inverted pyramid structure → Conceptual Framework
- Chapter 3 – Research Methodology: Design, population, sample, data needs, instruments, analysis procedures
- Chapter 4 – Budget: All required resources and costs
- Chapter 5 – Time Schedule: Gantt-chart style activity timeline
Characteristics of a Good Proposal
- Straight-forward and precise (no irrelevant material)
- Written in clear, academic prose
- Logically organised with proper headings
- Future tense (describes intended activities)
Common Weaknesses in Proposals
- Problem too nebulous, unimportant, or complex
- Data difficult to obtain or inappropriate
- Methods, instruments, or controls inadequate
- Statistical analysis too simplistic
- Investigator unfamiliar with relevant literature
- Institutional setting unfavourable
Methods of Data Collection
The Questionnaire
A special-purpose document to collect data from respondents. Each item must relate to a research objective.
Advantages: Inexpensive, anonymity, quick tabulation, large samples.
Disadvantages: Inflexible, no body language, cannot clarify vague answers.
Question Types: Structured/closed-ended (multiple choice, rating, ranking) and Unstructured/open-ended.
Also: Contingency/filter questions, Matrix questions (Likert scale).
The Interview
Oral administration of questions through face-to-face interaction. Allows probing, observation of non-verbal cues, and higher response rates.
Advantages: Probing possible, non-verbal observation, rephrase questions, in-depth data.
Disadvantages: Time-consuming, costly, requires interpersonal skills, smaller samples.
Types: Unstructured (open-ended, conversational) vs Structured (specific question set).
Tools: Interview schedule/guide, audio/video recording, note-taking.
Observation
Researcher participates in or watches activities to collect data. Can be participant or non-participant.
Advantages: Highly reliable, sees exactly what is done, inexpensive, eliminates subjective bias.
Disadvantages: People behave differently when watched, scheduling inconvenience, subject to interruptions.
Participant Observation (Qualitative): Steps → Jottings → Expanded Notes → Analysis (Open Coding → Focused Coding → Theoretical Coding) → Write-Up.
Types of Research
Classification Summary
| Classification By | Types |
|---|---|
| Method of Inquiry | Qualitative vs Quantitative (vs Mixed) |
| Theoretical Perspective | Inductive vs Deductive |
| Purpose | Basic vs Applied; Action vs Evaluation |
| Method of Analysis | Descriptive vs Analytical/Causal-Comparative |
| Other | Survey, Case Study, Ethnography, Historical, Experimental |
Qualitative vs Quantitative
| Qualitative | Quantitative |
|---|---|
| Words, pictures, rich descriptions | Numbers, statistics |
| Design emerges during study | Design fixed before data collection |
| Researcher is the instrument | Tools/instruments used |
| Subjective, interpretive | Objective, positivist |
| Inductive; theory-building | Deductive; theory-testing |
Basic Research
Adds to the universe of scientific knowledge. Motivated by intellectual curiosity. Usually in controlled lab settings. No concern with practical application.
Applied Research
Tests theory and evaluates usefulness in solving problems. Focuses on knowledge directly useful to practitioners.
Action Research
Solves a specific, immediate, local problem. Results not intended for generalisation.
Evaluation Research
Collects data to facilitate decision-making. Criteria: Utility, Feasibility, Propriety, Accuracy.
Case Study
In-depth, longitudinal examination of a single instance (person, group, event). Provides systematic understanding of why something happened. Used extensively in social sciences.
Qualitative Research & QDA
Types of Qualitative Research (Creswell)
- Biographical/Narrative: Portrait of an individual's life
- Phenomenological: Understanding a concept or phenomenon
- Ethnographic: Portrait of a cultural group
- Case Study: Examining a specific case
- Grounded Theory: Developing a theory from data
Qualitative Data Analysis (QDA)
QDA is iterative — data collection and analysis are interleaved. Main steps: Coding → Memoing → Presentation.
1st Cycle (Open/Initial): Descriptive, In Vivo, Process, Concept, Emotion, Values, Attribute coding.
2nd Cycle (Axial/Focused): Groups 1st cycle codes into categories, themes, concepts (meta-codes).
3rd Cycle (Selective/Theoretical): Validates/challenges the emerging theory against further data.
Memoing
Documenting the researcher's reflections on patterns and themes. Memos may be textual or diagrammatic and include dates, headings, and supporting data.
CAQDAS Tools
Computer Aided Qualitative Analysis Software. The software does not analyse data automatically — the researcher uses it to code, annotate, retrieve, and array data.
Examples: Atlas.ti, NVivo, MAXQDA, Dedoose. (CAT and Weft QDA are free.)
Ethical Issues in Qualitative Research
- Worthiness of project; Competence of researcher
- Informed consent; Benefits, costs, reciprocity
- Harm and risk; Honesty and trust
- Privacy, confidentiality, anonymity
- Ownership of data; Use and misuse of results
Quantitative Data Analysis & Presentation
Data Pre-Processing
Raw data must be cleaned → coded → entered → analysed. Coding assigns numerical values to responses. A codebook documents the coding scheme for all variables.
Descriptive Statistics
Measures of Central Tendency:
- Mode: Most frequent value
- Median: Middle value (50th percentile). Position = (n+1)/2
- Mean: Average; takes all values into account; sensitive to outliers
Measures of Variability/Dispersion:
- Range: Highest − Lowest (quick but not representative)
- Standard Deviation (σ): Average deviation from the mean; sensitive to outliers
- Variance: Square of the standard deviation
Inferential Statistics
Used to generalise results from a sample to a population. Classified as Parametric (assumptions made about data; more powerful) or Non-Parametric (no assumptions made).
| Technique | When Used |
|---|---|
| Pearson Correlation (r) | Two continuous interval/ratio variables; ranges −1 to +1 |
| Chi-Square (χ²) | Two categorical variables; non-parametric |
| Simple Regression | One IV predicting one DV |
| Multiple Regression | Multiple IVs predicting one DV |
Data Presentation
- Frequency Distribution Tables (simple and grouped)
- Histograms: Continuous quantitative data; bars touch (exact limits)
- Frequency Polygon: Line connecting midpoints of class intervals
- Bar Charts: Discrete/categorical data; bars separated
- Pie Charts: Proportions/percentages of a whole
- Percentages: Proportion of subgroup to total (0%–100%)
Writing the Research Report
Interpretation of Data
Interpretation is extracting meaning from data — making the data "speak." It requires great skill and must be done carefully to avoid false generalisations. It is intertwined with analysis and cannot be clearly separated.
- Explain relationships found in the data
- Consider all relevant factors before concluding
- Avoid broad generalisations — specify scope, time, area
- Consult experts and seek feedback before finalising
Layout of the Research Report
Preliminary Pages: Title Page · Dedication · Abstract · Table of Contents · List of Tables · List of Figures · Abbreviations · Executive Summary
Main Body Chapters:
- Chapter 1 – Introduction: Background, problem, objectives, scope, significance
- Chapter 2 – Literature Review: Critique of related studies, gaps, conceptual framework
- Chapter 3 – Research Methodology: Design, population, sample, instruments, analysis plan
- Chapter 4 – Data Analysis, Presentation & Discussion: Descriptive → inferential results; discussion linked to literature
- Chapter 5 – Summary, Conclusions & Recommendations
Suffix: References · Appendices
Steps in Writing the Report
- Logical analysis of subject matter (logical or chronological)
- Prepare the final outline
- Write the rough draft
- Rewrite and polish (check unity, cohesion, grammar, spelling)
- Prepare final references (APA format)
- Write the final report (concise, objective, simple language)
Guidelines for the Results & Discussion Chapter
- Present descriptive statistics first, then inferential
- Use tables and figures only where they add value
- Discussion should appear before or after each table
- Link findings to the literature review
- Explain contradictions with other studies
- Objectives/hypotheses guide which tables to include
Common Recommendations Researchers Make
- Areas for further research (unanswered questions)
- Methodological improvements for future studies
- Actions to address the research problem