RESEARCH METHODOLOGY

Module IV Guide
Report Writing in Applied Statistics
Instructor: Prof. D. K. Muriithi, PhD.

Most MSc/PhD theses and research reports follow a similar structure.

1. Title
Definition: A concise and informative statement that reflects the content and scope of the research.
A good title is:
• Concise but specific (usually 15–25 words).
• Reflects the main variables, population, and context.
• Clear and specific
• Includes keywords related to statistical methodology and application area
• Avoids jargon or unnecessary words

Example: "Application of Multivariate Regression Models in Predicting Household Energy

Consumption in Urban Kenya"

Example: “Determinants of Smallholder Maize Yield in Tharaka-Nithi County: A Multilevel

Modeling Approach.”

NB:
✓ Use active voice if possible
✓ Mention the method used (e.g., logistic regression, ANOVA, Survival Analysis)
✓ Include the population or domain studied

2. Abstract
Definition: A brief summary of the entire research report, typically between 150–300 words.

Key Components:
• Background/problem.
• Purpose of the study
• Objectives.
• Methods (design, data, key statistical methods).
• Main results (key numerical findings).
• Conclusions and implications.
NB: Write it last but place it first in the document.

Example: "This study examines the use of multivariate regression models to predict household
energy consumption in urban Kenya. Data was collected from 500 households using stratified
sampling. Multiple linear regression and ridge regression were fitted, with ridge regression
showing better performance based on RMSE. The model can assist policymakers in forecasting
future energy demand."

Illustration:
Section Content
Objective Predict household energy consumption
Method Multivariate regression model
Sample Size 500 households
Findings Ridge regression outperformed OLS
Implication Useful for policy planning

3. Introduction
Definition: Sets the context of the research, explains the problem being addressed, and outlines
the significance of the study.
Structure:
1. Background information
2. Statement of the problem
3. Rationale for the study
4. Gap in existing knowledge
5. Significance of the study
Functions:
• Introduce the topic and context (global, regional, local).
• Present the problem statement and its significance.
• Provide objectives and sometimes research questions/hypotheses.
• Outline structure of the thesis (optional).
Keep it logically flowing from broad to specific.

Example: "Energy demand forecasting is critical for infrastructure development in rapidly

urbanizing regions. While several studies have modeled energy consumption at the national level,
few have focused on predicting it at the household level using statistical modeling techniques. This
study fills this gap by applying multivariate regression models to estimate household-level
electricity usage in Nairobi."

Illustration:
Element Description
Background Rising urbanization increases energy demand
Problem Lack of accurate prediction at household level
Rationale Better forecasting improves grid management
Gap Limited focus on household-level data
Significance Supports sustainable energy planning
4. Objectives
Definition: Clearly defined statements indicating what the researcher aims to achieve through
the study.
Types:
• General Objective: Broad aim of the research
• Specific Objectives: Step-by-step breakdown of how the general objective will be achieved
Example:
General Objective: To apply multivariate regression models to predict household energy
consumption in urban Kenya.

Specific Objectives:
1. To collect and preprocess household energy usage data.
2. To identify significant predictors of energy consumption.
3. To fit multiple linear regression and ridge regression models.
4. To compare model performance using RMSE and R².
NB: Use SMART criteria – Specific, Measurable, Achievable, Relevant, Time-bound.

5. Methods and Materials

Definition: Describes how the research was conducted, including data collection, statistical tools,
software, and analytical procedures.

Methods and materials (Methodology)

Core for applied statistics.
Components:
• Study area and population: geographic and demographic description.
• Study design: cross-sectional, longitudinal, experimental, etc.
• Sampling: Explain who the respondents are, how they are selected & why the sample is
valid and adequate.
✓ Target population, sampling frame, inclusion/exclusion criteria.
✓ Sampling design: simple random, stratified, cluster, multistage, etc.
✓ Sample size determination: formula, parameters, design effect, non-response
adjustment.
• Data collection:
✓ Instruments (questionnaire, measurements), development and validation.
✓ Data collection procedures: pilot, training, fieldwork.
✓ Quality control: supervision, re-visits, data checks.
• Variables:
✓ Dependent and independent variables, covariates, potential confounders.
✓ Operational definitions and measurement scales (nominal, ordinal, interval, ratio).
• Statistical analysis plan:
✓ Data management: coding, entry, cleaning, handling missing data.
✓ Descriptive analysis: measures of central tendency, dispersion, graphs.
✓ Inferential/statistical modeling:
 Choice of models (e.g., linear regression, logistic regression, Poisson,
survival analysis, mixed models, time series models).
Model assumptions (linearity, independence, homoscedasticity, normality,
etc.).
Model selection procedures (e.g., AIC, BIC, cross-validation).
Goodness-of-fit and diagnostics (residual analysis, influence diagnostics).
Handling of clustering or complex survey design if applicable.
✓ Software: specify packages (e.g., R, Stata, SPSS, SAS, Python).
• Ethical considerations:
✓ Ethical approval, consent process, data confidentiality.
The methods section should allow another researcher to replicate your work.

Example:
Study Design: Quantitative observational study using cross-sectional survey data.
Data Collection: Primary data collected from 500 households via structured questionnaires
administered in Nairobi County.
Sampling: Stratified random sampling based on income levels.
Statistical Methods:
• Descriptive and Inferential statistics
• Correlation Analysis
• Multiple Linear Regression
• Ridge Regression etc
• Machine learning methods
Software: R, Python, STATA, SPSS etc
Illustration using R:
i. Dependent Variable: Monthly kWh consumed
ii. Independent Variables: Income, Number of appliances, House size
iii. Model Fitting: lm() and glmnet() functions
iv. Validation: 10-fold cross-validation

6. Results
Definition: Presents the findings obtained from the statistical analysis without interpretation.
(Presents analyzed findings, not raw data).
Guidelines:
• Follow the order of objectives or research questions.
• Start with descriptive results:
• Sample characteristics (tables of demographic or baseline variables).
• Key summary statistics and figures.
• Then present inferential/model results:
• Regression tables with estimates, standard errors, confidence intervals, p-values.
• Model fit measures and diagnostics summaries where relevant.
• Use appropriate tables and figures:
• Clear titles, numbered, with notes explaining abbreviations and measures.
 • Interpret numbers briefly in text but save detailed discussion for the Discussion section.

Example:
“In multivariable logistic regression, access to extension services was associated with higher odds
of adoption (adjusted odds ratio 2.102.10, 95% CI 1.451.45–3.043.04, p<0.001p<0.001).”

Discussion (often expected, even if not listed)

Key tasks:
• Interpret your results in light of objectives and existing literature.
• Compare with similar studies (agreement, differences, possible reasons).
• Discuss implications for theory, methods, and practice.
• Acknowledge limitations (e.g., measurement error, residual confounding).
• Suggest areas for further research.
Conclusions and recommendations
• Summarize key findings in relation to objectives.
• Provide clear, evidence-based recommendations for policy, practice, or further research.
• Avoid over-generalization beyond your sampling frame and methods.

Example: "The correlation matrix showed a strong positive relationship between income and
energy consumption (r = 0.62, p < 0.001). Ridge regression had a lower RMSE (2.1) compared to
multiple linear regression (RMSE = 2.7)."
Illustration:
Table: Model Performance Comparison
Model RMSE R²
Multiple Regression 2.7 0.72
Ridge Regression 2.1 0.79

Figure: Scatter Plot of Predicted vs Actual Values

(You can insert a plot here showing actual vs predicted values with a line of best fit.)
NB:
• Use clear headings for tables and figures
• Reference each table/figure in the text
• Ensure all visualizations are labeled and interpretable

Remarks
As a Master’s student in Applied Statistics, your ability to write a well-structured research report
is essential for communicating your findings effectively. Each section from the title to the results
plays a role in presenting your work clearly, logically, and professionally. By following these
guidelines, you ensure that your research is:
• Reproducible
• Understandable to peers and stakeholders
• Applicable to real-world problems
References
1. Creswell, J. W. (2014). Research Design: Qualitative, Quantitative, and Mixed Methods
Approaches. Sage Publications.
2. Montgomery, D. C. (2020). Design and Analysis of Experiments. Wiley.
3. Field, A. (2018). Discovering Statistics Using IBM SPSS Statistics. Sage.
4. American Psychological Association (APA). (2020). Publication Manual of the American
Psychological Association (7th ed.).

CAT I:
Prepare a research proposal (maximum 15 pages) with appropriate title, abstract, introduction,
literature review, methodology, budget and references and submit by End of March 14, 2026.
Definition of a Seminar
A seminar is an academic or professional meeting where a small group of participants engages in
discussion, presentation, and critical analysis of specific topics, often under the guidance of one or
more instructors or experts. In the context of MSc/PhD program, especially in fields like Applied
Statistics, seminars are used to:
• Encourage critical thinking
• Promote knowledge sharing
• Develop presentation and communication skills
• Prepare students for research projects or thesis work

Key Characteristics of a Seminar

Feature Description
Size Small group (usually 5–20 participants)
Format Combination of presentations and discussions
Focus In-depth exploration of a topic or research area
Interaction High level of participant engagement
Preparation Requires prior reading/research by participants

Types of Seminars (Common in Graduate Studies)

1. Research Seminars
• Focus on presenting ongoing or completed research
• Example: A student presents their statistical modeling approach for analyzing
healthcare data
2. Topic-Based Seminars
• Explore theoretical or methodological issues in depth
• Example: “Machine Learning Techniques in High-Dimensional Data Analysis”
3. Literature Review Seminars
• Involve critical discussion of published research
• Example: Reviewing recent papers on Bayesian hierarchical models
4. Guest/Invited Seminars
• Led by external experts or professionals
• Example: A data scientist from a pharmaceutical company discussing clinical trial
modeling

Purpose of a Seminar
Goal Explanation
Knowledge Sharing Disseminate current research findings or trends
Critical Thinking Encourage questioning and evaluation of methods and results
Skill Development Improve public speaking, analytical, and synthesis skills
Networking Connect with peers and experts in the field
Feedback Get constructive input on research ideas or preliminary findings
Example of a Seminar in Applied Statistics
Title: “Applications of Survival Analysis in Public Health Research”
Presenter: Master’s student in Biostatistics
Duration: 45 minutes
Structure:
1. Introduction to survival analysis
2. Case study: Modeling time-to-recovery in HIV patients using Cox Proportional Hazards
model
3. Comparison with parametric models (Weibull, Exponential)
4. Discussion on assumptions and limitations
5. Q&A session with faculty and peers
How Seminars Differ from Lectures
Aspect Lecture Seminar
Audience Size Large Small
Interaction Minimal High
Content Delivery Instructor-led Student-led or collaborative
Teaching new
Objective Exploring and discussing existing or original content
material
Assessment Often none May be graded based on participation or presentation

Benefits of Attending or Presenting in a Seminar

• Enhances understanding through peer interaction
• Improves presentation and defense of ideas
• Builds confidence in discussing complex statistical concepts
• Offers early exposure to research culture
• Helps identify gaps in your own knowledge or research

Tips for Giving a Successful Seminar (as a Presenter)

1. Know your audience: Tailor technicality to the participants’ background.
2. Be clear and concise: Avoid information overload.
3. Use visuals effectively: Charts, tables, and code snippets should enhance clarity.
4. Engage the audience: Ask questions or invite opinions.
5. Prepare for questions: Anticipate critiques on methodology, assumptions, or conclusions.
References
• Creswell, J. W. (2014). Research Design: Qualitative, Quantitative, and Mixed Methods
Approaches. Sage Publications.
• Denzin, N. K., & Lincoln, Y. S. (Eds.). (2011). The SAGE Handbook of Qualitative
Research. Sage.
• University of North Carolina Writing Center. (n.d.). How to Give a Good Seminar
Presentation.
Seminars (presentation of research)
PhD/MSc students usually present progress and final results in seminars.
Key aspects:
• Structure of presentation:
• Title and authors/institution.
• Background and problem.
• Objectives.
• Methods (emphasize design and statistical methods).
• Key results (well-chosen tables/graphs).
• Conclusions and recommendations.
• Time management: 10–20 minutes, plus Q&A.
• Visual aids:
• Use clear, readable slides (avoid dense text, show key tables/figures).
• Highlight main messages per slide.
• Responding to questions:
• Be prepared to explain modeling choices, assumptions, and limitations.
• Acknowledge valid critiques honestly and indicate possible improvements.
For applied statistics, expect questions about:
• Justification of chosen models over alternatives.
• Assumption checking and robustness.
• Handling of missing data, outliers, and influential observations.
• Interpretation of effect sizes and uncertainty (confidence intervals, prediction intervals).

CAT II: Prepare for a seminar presentation on the 1st week of April 2026 (11/04/26)
Write a Research Article in 40 Steps
MAKE DRAFT

Step 1 Make a working title

Step 2 Introduce the topic and define terminology
Step 3 Emphasize why the topic is important
Step 4 Relate to current knowledge: what has been done
Step 5 Indicate the gap: what needs to be done
Step 6 Pose research questions
Step 7 Give purpose and objectives
Step 8 List methodological steps
Step 9 Explain theory behind the methodology used
Step 10 Describe experimental set-up
Step 11 Describe object of the study (technical details)
Step 12 Give summary results
Step 13 Compare different results
Step 14 Focus on main discoveries
Step 15 Answer research questions (conclusions)
Step 16 Support and defend answers
Step 17 Explain conflicting results, unexpected findings and discrepancies with other research
Step 18 State limitations of the study
Step 19 State importance of findings
Step 20 Establish newness
Step 21 Announce further research
Step 22 ABSTRACT: what was done, what was found and what are the main conclusions

REVISE

Step 23 Is the title clear and does it reflect the content and main findings?
Step 24 Are key terms clear and familiar?
Step 25 Are the objectives clear and relevant to the audience?
Step 26 Are all variables, techniques and materials listed, explained and linked to existing
knowledge — are the results reproducible?
Step 27 Are all results and comparisons relevant to the posed questions/objectives?
Step 28 Do some statements and findings repeat in the text, tables or figures?
Step 29 Do the main conclusions reflect the posed questions?
Step 30 Will the main findings be unacceptable by the scientific community?
Step 31 Is the text coherent, clear and focused on a specific problem/topic?
Step 32 Is the abstract readable standalone (does it reflect the main story)?
POLISH

Step 33 Are proper tenses and voices used (active and passive)?
Step 34 Are all equations mathematically correct and explained in the text?
Step 35 Are all abbreviations explained?
Step 36 Reconsider (avoid) using words “very”, “better”, “may”, “appears”, “more convinced”,
“impression”
Step 37 Are all abbreviations, measurements, units, variables and techniques internationally
recognized (SI)?
Step 38 Are all figures/tables relevant and of good quality?
Step 39 Are all figures, tables and equations listed and mentioned in the text?
Step 40 Are all references relevant, up to date and accessible?