Module title
Module description.
Learning Outcome
Intern Activity
Submission
Aviation / Aerospace Applied Example:
Research Internship
Research Methodology & Scientific Writing Portal
Premium web-based LMS dashboard with 11 modules, full unchanged notes, aviation/aerospace applied examples, module FAQs, instructor details and progress tracking.
Meaning and Importance of Research
Research is a systematic, logical, and objective process of investigation undertaken to discover new facts, verify existing knowledge, or find solutions to a defined problem. It moves beyond guesswork by relying on planned methods of data collection, analysis, and interpretation.
For a UG student, research matters because it builds critical thinking, teaches evidence-based reasoning, contributes to the existing body of knowledge, and is the foundation of every academic degree, patent, product innovation, and policy decision.
Academic Writing vs. Research Writing
| Aspect | Academic Writing | Research Writing |
|---|---|---|
| Purpose | Explain or summarize existing knowledge | Generate or test new knowledge |
| Basis | Textbooks, lecture notes | Primary/secondary data, experiments, literature |
| Structure | Essay-like, descriptive | IMRaD (Intro, Methods, Results, Discussion) |
| Evaluation | Clarity and understanding | Originality, rigor, reproducibility |
Types of Research
- Basic (Pure) Research — driven by curiosity, aims to expand theoretical understanding without immediate application (e.g., studying a new material's molecular behaviour).
- Applied Research — aims to solve a specific, practical problem (e.g., designing a low-cost water filter for a village).
- Qualitative Research — explores meanings, experiences, and perceptions using non-numeric data (interviews, case studies, observations).
- Quantitative Research — relies on numeric data, measurement, and statistical analysis to test hypotheses.
- Mixed-Method Research — combines qualitative and quantitative approaches in a single study for a more complete picture.
Characteristics of a Good Research Topic
- Relevant — connects to a real academic, industrial, or social need.
- Original/Novel — offers a new angle, not a repetition of settled work.
- Feasible — achievable within the time, budget, and resources available.
- Clear and specific — narrow enough to be investigated in depth.
- Measurable — allows data to be collected and analysed.
- Ethical — does not cause harm to people, animals, or the environment.
How to Identify a Research Gap
A research gap is an unanswered question or an unexplored area found by comparing what has already been studied against what is missing, contradictory, or outdated in the literature.
- Read "limitations" and "future scope" sections of published papers.
- Look for conflicting results between different studies.
- Note variables, tools, populations, or geographies that no one has tested yet.
- Identify old studies that need re-validation with modern tools or data.
Framing a Research Problem
A research problem is a precise statement of the issue that the study intends to address. A well-framed problem answers: what is being studied, who/what is affected, and why it matters.
Example: "The mechanical strength of concrete reduces significantly when fly ash exceeds 30% replacement — this study investigates the optimum fly-ash percentage for structural safety."
Developing Research Objectives and Research Questions
Objectives are broad statements of intent, usually written using action verbs (to analyse, to evaluate, to compare, to design). Research questions are the specific queries the study will answer, phrased as questions.
Objective: To evaluate the effect of fly ash content on concrete compressive strength.
Research Question: How does increasing fly ash percentage affect the 28-day compressive strength of concrete?
Research Question: How does increasing fly ash percentage affect the 28-day compressive strength of concrete?
Aim vs. Objectives vs. Hypothesis vs. Research Question
| Term | Definition | Example |
|---|---|---|
| Aim | The single overarching goal of the study | To improve the durability of concrete using industrial waste. |
| Objectives | Specific, measurable steps to achieve the aim | To test 3 fly-ash ratios; to compare strength at 7/14/28 days. |
| Hypothesis | A testable, predictive statement about the relationship between variables | Increasing fly ash beyond 30% decreases compressive strength. |
| Research Question | An open enquiry the study seeks to answer | What is the optimum fly-ash ratio for maximum strength? |
Quick Recap
- Research is systematic and evidence-based, unlike casual academic writing.
- Five types: basic, applied, qualitative, quantitative, mixed-method.
- A good topic is relevant, original, feasible, specific, measurable, and ethical.
- Aim → broad goal; Objectives → measurable steps; Hypothesis → testable prediction; Research Question → open enquiry.
Meaning of Research Methodology
Research methodology is the overall strategy and rationale behind a study — it explains how research will be conducted, why particular methods were chosen, and how the approach connects to the research objectives.
Research Method vs. Research Methodology
| Research Method | Research Methodology |
|---|---|
| The specific tools/techniques used to collect and analyse data (e.g., survey, t-test) | The overall logic, design, and justification of the research process |
| Answers "how was the data gathered?" | Answers "why was this approach chosen?" |
| A part of methodology | The umbrella that includes methods |
Types of Research Design
- Exploratory Design — used when the problem is not well defined; aims to generate insights and hypotheses (e.g., preliminary interviews on a new technology).
- Descriptive Design — describes characteristics of a population or phenomenon systematically (e.g., a survey of student internet usage).
- Experimental Design — manipulates one or more variables under controlled conditions to observe cause-and-effect (e.g., testing a new fertilizer's effect on crop yield).
- Analytical Design — uses existing facts or data already available to make a critical evaluation (e.g., analysing secondary data on GDP trends).
- Case-Study Design — an in-depth investigation of a single unit, organisation, or event to draw detailed insights.
Qualitative vs. Quantitative Approaches
| Qualitative | Quantitative |
|---|---|
| Explores "why" and "how" | Measures "how much" and "how many" |
| Data: words, themes, narratives | Data: numbers, statistics |
| Tools: interviews, focus groups | Tools: surveys, experiments, sensors |
| Analysis: thematic coding | Analysis: statistical tests |
Types of Variables
- Independent Variable — the variable manipulated or changed by the researcher (the presumed cause).
- Dependent Variable — the outcome measured, which is expected to change due to the independent variable (the effect).
- Control Variable — kept constant across the study to isolate the effect of the independent variable.
- Moderating Variable — a variable that influences the strength or direction of the relationship between the independent and dependent variables.
Example: Studying the effect of study hours (independent) on exam scores (dependent), while keeping the subject difficulty constant (control), and noting that the effect may vary depending on the student's sleep quality (moderating).
Sampling Techniques
- Random Sampling — every member of the population has an equal chance of selection; minimises bias.
- Stratified Sampling — population is divided into sub-groups (strata) based on a shared characteristic, and samples are drawn proportionally from each stratum.
- Convenience Sampling — samples chosen based on ease of access; quick but prone to bias.
- Purposive Sampling — samples are deliberately selected because they possess specific characteristics relevant to the study objective.
Tools and Techniques Used in Research
Common tools include questionnaires, interview schedules, observation checklists, laboratory instruments, statistical software (SPSS, R, Excel, MATLAB), simulation software, and reference managers (Mendeley, Zotero).
Framework for Writing the Methodology Section
- Restate the research design chosen and justify why it fits the objectives.
- Describe the population and sampling technique used.
- List the tools/instruments and materials used for data collection.
- Explain the procedure step-by-step, in past tense, in enough detail for replication.
- State the data analysis technique/statistical tests applied.
- Mention any ethical considerations and limitations of the method.
Quick Recap
- Methodology = the "why" and overall plan; Method = the specific "how" tool.
- Five design types: exploratory, descriptive, experimental, analytical, case-study.
- Four variable types: independent, dependent, control, moderating.
- Four sampling techniques: random, stratified, convenience, purposive.
Meaning and Purpose of Literature Review
A literature review is a critical, organised summary of existing published work related to a research topic. Its purpose is to show what is already known, how it was studied, and where gaps remain — positioning the new study within the existing body of knowledge.
Importance of Literature Review
- Prevents duplication of already-completed research.
- Helps refine the research problem and objectives.
- Provides theoretical grounding and terminology for the study.
- Identifies suitable methods, tools, and datasets used by others.
- Builds the case for why the proposed study is needed (the research gap).
Sources of Literature
Journal papers, conference proceedings, books and book chapters, doctoral/master's theses, government and institutional reports, and authentic websites (government portals, standards bodies, recognised organisations).
Searching Research Papers
Common academic databases: Google Scholar (broad, free), ResearchGate (author-shared papers), Scopus-indexed journals (quality-vetted, citation-tracked), IEEE Xplore (engineering/technology), SpringerLink, and ScienceDirect (Elsevier).
Keyword and Boolean Search Techniques
Effective search combines core keywords with Boolean operators to narrow or widen results:
AND— narrows results (e.g., "solar energy AND rural electrification")OR— widens results (e.g., "machine learning OR deep learning")NOT— excludes a term (e.g., "battery NOT lithium")- Quotation marks — searches an exact phrase (e.g., "compressive strength of concrete")
How to Read a Research Paper Effectively
Follow a layered reading strategy rather than reading linearly start-to-finish:
- Read the title and abstract to judge relevance.
- Skim the conclusion for the main outcome.
- Study the methodology to understand the approach used.
- Examine results and figures for the actual findings.
- Only then read the full introduction and discussion for context and interpretation.
Preparing a Literature Review Table
A structured table makes it easy to compare multiple papers at a glance and spot patterns or gaps.
| Author/Year | Research Topic | Method Used | Key Findings | Limitations | Possible Research Gap |
|---|---|---|---|---|---|
| Paper 1 | |||||
| Paper 2 | |||||
| Paper 3 |
Identifying Research Gaps from Existing Studies
Compare the "Key Findings" and "Limitations" columns across multiple papers in the review table. Repeated limitations across several studies, or a variable/tool/context nobody has tested, is a strong indicator of a genuine research gap.
Avoiding Copy-Paste and Plagiarism
- Always paraphrase ideas in your own words instead of copying sentences.
- Cite every source using a consistent referencing style (APA, IEEE, etc.).
- Use quotation marks only for short, essential direct quotes, with citation.
- Run your final draft through a plagiarism-checking tool before submission.
Tip: A good literature review is not a list of summaries — it is a synthesis that groups papers by theme, method, or finding, and ends by clearly stating the gap your research will fill.
Quick Recap
- Literature review situates your study within existing knowledge and justifies its need.
- Use Boolean operators (AND/OR/NOT) to refine database searches.
- Read papers in the order: abstract → conclusion → methodology → results → introduction/discussion.
- Always paraphrase and cite to avoid plagiarism.
Searching Papers Using Keywords
Start with 3–5 core keywords from your topic, run searches on Google Scholar/Scopus/IEEE, then refine using synonyms and Boolean combinations until you get 15–25 relevant, recent (preferably last 5–10 years) papers.
Reading Each Section Purposefully
| Section | What to Look For |
|---|---|
| Title | Topic relevance at a glance |
| Abstract | Problem, method, and key result in summary |
| Introduction | Background, motivation, and stated objectives |
| Methodology | How the study was conducted — replicate or compare |
| Results | Actual data, tables, graphs, and numeric outcomes |
| Conclusion | Summary claim and stated limitations/future scope |
Identifying Limitations and Future Scope
Authors almost always disclose what their study could not achieve. These lines are the fastest route to a research gap — read them first before deciding whether a full read of the paper is needed.
Comparing Methods Across Papers
Tabulate the method used by each paper side by side. Patterns to note: which method is most common, which is outdated, and which combination of methods has not yet been tried together.
Identifying Repeated Problems in Multiple Studies
If three or more independent papers report the same limitation (e.g., "small sample size" or "not tested under real-world conditions"), that repeated problem is a validated, high-confidence research gap.
Finding Missing Variables, Tools, Datasets, or Case Studies
- Variables not yet tested in combination.
- Newer tools/technologies not yet applied to an older problem.
- Datasets or geographic regions not yet covered.
- Industries or case-study contexts left unexplored.
Using Review Papers for Broader Gaps
Review/survey papers summarise dozens of studies at once. Their "open challenges" or "future research directions" sections are an efficient shortcut to identifying broad, well-recognised gaps in a field.
Data Extraction Table Format
Once relevant papers are shortlisted, extract structured data into a master table for systematic comparison:
| Sl. No. | Author & Year | Research Topic | Method Used | Input Parameters | Output Parameters | Dataset/Sample | Key Findings | Limitations | Research Gap |
|---|---|---|---|---|---|---|---|---|---|
| 1 |
Practical exercise: Pick your working topic, shortlist 8–10 papers, and fill both the Literature Review Table (Module 3) and the Data Extraction Table above. Then write a 3–4 sentence gap statement based purely on the patterns you observe.
Quick Recap
- Reading order should be purposeful — not linear — to save time.
- Repeated limitations across multiple papers = strong research gap evidence.
- Review/survey papers are a shortcut to field-wide open challenges.
- The Data Extraction Table converts scattered reading into structured, comparable data.
Meaning of Data Extraction from Literature
Literature-based data extraction is the process of systematically pulling out specific, comparable pieces of information (parameters, methods, results) from multiple published papers into a structured format for analysis.
Importance in Research Paper Writing
- Converts scattered reading into organised, comparable evidence.
- Makes trends and gaps across studies visible at a glance.
- Provides the raw material for a strong, evidence-based literature review and gap statement.
Reading a Paper vs. Extracting Research Data
Reading is understanding what a paper says. Extraction is deliberately pulling specific, predefined data fields out of that understanding and placing them into a table for cross-paper comparison — a more disciplined, goal-driven activity than general reading.
Role of Parameter Study in Research
A parameter study identifies which independent factors (inputs) were varied in past studies and which outcomes (outputs) were measured, helping the researcher decide which parameters still need investigation.
Types of Information to Extract from Research Papers
| Field | What it Captures |
|---|---|
| Author and Year | Source identification and recency |
| Research Objective | What the paper set out to investigate |
| Materials/Dataset/Sample Details | What was studied and its scale |
| Input Parameters | The variables manipulated/tested |
| Output Parameters | The measured outcomes/results |
| Methodology Used | The design/approach followed |
| Tools/Software Used | Instruments, software, or analytical tools |
| Results and Findings | The key numerical or qualitative outcome |
| Limitations | What the study could not address |
| Research Gap | What remains unexplored, based on the above |
Note: This field list matches the Data Extraction Table introduced in Module 4 — use that table as your working template while extracting.
Quick Recap
- Extraction is more targeted and structured than ordinary reading.
- Ten standard fields to extract: author/year, objective, materials, inputs, outputs, method, tools, findings, limitations, gap.
- Parameter study focuses specifically on what inputs/outputs each study covered.
Meaning and Importance of Data Collection
Data collection is the systematic gathering of information relevant to the research objectives. The quality of a study's conclusions depends directly on how accurately and appropriately its data was collected.
Types of Data
Primary data — collected directly by the researcher for the current study. Secondary data — collected previously by others and reused for a new purpose.
Primary Data Collection Methods
- Survey — collects data from a large group using standardised questions.
- Questionnaire — a self-administered set of written questions, often part of a survey.
- Interview — direct, often open-ended, verbal questioning of respondents.
- Observation — recording behaviour or events as they naturally occur, without direct interference.
- Experiment — controlled manipulation of variables to test cause-and-effect relationships.
- Field study — data collected in a real-world, natural setting rather than a controlled lab.
Secondary Data Collection Sources
- Published research papers
- Government and organisational reports
- Government databases (census, statistical bureaus)
- Company records
- Online open datasets (e.g., Kaggle, UCI, government open-data portals)
- Institutional repositories (university thesis databases, digital libraries)
Qualitative vs. Quantitative Data
Qualitative data is descriptive and non-numeric (opinions, themes, observations), while quantitative data is numeric and measurable (counts, scores, physical measurements). Many studies collect both to triangulate findings.
Ethical Issues in Data Collection
- Consent — participants must knowingly and voluntarily agree to take part.
- Privacy — personal or sensitive data must be protected and anonymised where possible.
- Accuracy — data must be recorded honestly, without alteration.
- Avoiding data manipulation — no fabrication, selective reporting, or falsification of collected data.
Quick Recap
- Primary data is collected first-hand; secondary data is reused from existing sources.
- Six primary methods: survey, questionnaire, interview, observation, experiment, field study.
- Ethics of data collection rest on consent, privacy, accuracy, and honesty.
Meaning of Parameter Study
A parameter study systematically examines how changes in one or more input parameters affect an output/result, either within a single experiment or across multiple published studies.
Comparing Parameters from Multiple Papers
Once data has been extracted (Module 9), list every input and output parameter used across the shortlisted papers side by side to see which parameters are studied often, and which combinations have never been tested together.
Identifying Common and Missing Parameters
- Common parameters — appear across most studies; considered the "standard" variables in that field.
- Missing parameters — rarely or never tested; a strong candidate for a research gap if they are plausible influencers of the outcome.
Understanding Parameter Influence on Results
For each parameter, examine whether increasing or decreasing it caused the output to rise, fall, or show no clear pattern. Contradictory findings between papers about the same parameter often point to unexplored interacting factors (a moderating variable — see Module 2).
Finding Trends from Literature
Plotting extracted values (e.g., input parameter vs. reported output) from several papers on a single chart can reveal an overall trend line across the field, even when no single paper covered the full range on its own.
Preparing Comparison Tables for Research Papers
| Author/Year | Input Parameter(s) | Range Tested | Output Parameter | Observed Trend |
|---|---|---|---|---|
This table becomes the evidence base for stating, in your own paper, exactly which parameter range remains unexplored — the final, most specific form of a research gap.
Quick Recap
- Parameter study links specific inputs to specific outputs across the literature.
- Common parameters define the field's baseline; missing parameters define the opportunity.
- Comparison tables turn scattered findings into a visible, field-wide trend.
Importance of Statistics in Academic Research
Statistics converts raw observations into interpretable evidence. It allows a researcher to summarise large datasets, test whether observed patterns are meaningful or due to chance, and support conclusions with objective, quantifiable proof rather than opinion.
Role of Statistics in Methodology and Results
In the methodology section, statistics justifies sample size and the analytical tests to be used. In the results section, it presents the actual computed values (mean, standard deviation, p-values) that either support or reject the hypothesis.
Data, Information, and Statistical Evidence
| Term | Meaning |
|---|---|
| Data | Raw, unprocessed facts or numbers (e.g., individual test scores) |
| Information | Data that has been organised or summarised to give it meaning (e.g., average score of a class) |
| Statistical Evidence | Information analysed using statistical tests to support or refute a claim (e.g., a significant difference between two groups' averages) |
Types of Data
- Primary Data — collected first-hand by the researcher specifically for the study (surveys, experiments).
- Secondary Data — already collected by someone else and reused (published reports, databases, prior papers).
Types of Variables (Statistical Perspective)
- Independent/Dependent — cause and effect, as covered in Module 2.
- Control — held constant to isolate effects.
- Categorical — represents categories or labels (e.g., gender, material type).
- Continuous — can take any value within a range (e.g., temperature, weight).
- Discrete — countable, whole-number values (e.g., number of defects).
- Ordinal — categories with a meaningful order but unequal intervals (e.g., satisfaction: low/medium/high).
Population and Sample
The population is the entire group a study wants to draw conclusions about. A sample is a smaller, manageable subset of that population, selected so that conclusions from the sample can be generalised back to the population.
Measurement Scales
| Scale | Description | Example |
|---|---|---|
| Nominal | Categories with no order | Blood group, gender |
| Ordinal | Ordered categories, unequal gaps | Rank in class, satisfaction level |
| Interval | Ordered, equal intervals, no true zero | Temperature in °C |
| Ratio | Ordered, equal intervals, true zero exists | Weight, height, age, income |
How Statistics Supports Research Conclusions
Statistics allows a researcher to move from "it looks like there's a difference" to "there is a statistically significant difference (p < 0.05)" — giving conclusions credibility, replicability, and the ability to withstand peer scrutiny.
Quick Recap
- Data → Information → Statistical Evidence is a chain of increasing meaning.
- Primary data is collected first-hand; secondary data is reused from existing sources.
- Four measurement scales: nominal, ordinal, interval, ratio (increasing precision).
- A sample must represent the population for conclusions to generalise correctly.
Meaning and Importance of Descriptive Statistics
Descriptive statistics summarise and organise a dataset's main features — its centre, spread, and shape — into a form that is easy to understand, without yet drawing conclusions beyond the data itself.
Measures of Central Tendency
Mean — the arithmetic average; sum of all values divided by their count. Sensitive to outliers.
Median — the middle value when data is arranged in order; not affected by extreme values.
Mode — the most frequently occurring value in the dataset; useful for categorical data.
Measures of Dispersion
- Range — difference between the highest and lowest value.
- Variance — average of the squared deviations from the mean; shows overall spread.
- Standard Deviation — square root of variance; expressed in the same unit as the data, so it is more interpretable.
- Coefficient of Variation (CV) — standard deviation expressed as a percentage of the mean; useful for comparing variability across datasets with different units or scales.
Frequency Distribution and Percentage Analysis
A frequency distribution groups data into class intervals and shows how many observations fall into each — the basic building block of a histogram. Percentage analysis converts these frequencies into proportions of the total, making comparisons across groups of different sizes easier.
Ranking and Comparison
Ranking arranges data points from highest to lowest (or vice versa) on a chosen criterion, useful for identifying top performers, priority parameters, or best-fit conditions in an experiment.
Data Visualization for Research Papers
- Bar chart — comparing discrete categories.
- Pie chart — showing proportion of a whole (best with ≤6 categories).
- Line graph — showing a trend over a continuous variable, usually time.
- Histogram — showing the frequency distribution of continuous data.
- Scatter plot — showing the relationship/correlation between two continuous variables.
Preparing Tables and Figures for a Research Paper
- Number tables and figures sequentially (Table 1, Figure 1...) and refer to them in the text.
- Give every table/figure a clear, self-explanatory caption/title.
- Label all axes with units; keep decimal places consistent.
- Avoid 3D or overly decorative charts — clarity beats decoration in a research paper.
Writing Interpretation for Statistical Tables and Graphs
Never leave a table or graph to "speak for itself." Follow every visual with 2–3 sentences explaining what the pattern means in relation to the research objective.
Example: "As shown in Figure 3, compressive strength increased steadily up to 20% fly-ash replacement, after which it declined — indicating an optimum replacement ratio of 20%."
Quick Recap
- Central tendency: mean, median, mode; Dispersion: range, variance, SD, CV.
- Choose the chart type based on the data relationship you want to show.
- Every table/figure needs a caption, unit labels, and a written interpretation.
Meaning of Error in Research Data
Error is the natural, often unavoidable, difference between a measured or predicted value and the true or accepted value. Recognising and reporting error honestly is a mark of rigorous research.
Error vs. Mistake
An error arises from natural limitations of instruments, methods, or sampling and can be quantified. A mistake is an avoidable human blunder (e.g., wrong data entry, misreading a scale) that should be corrected, not reported as "error."
Accuracy and Precision
| Accuracy | Precision |
|---|---|
| Closeness of a measurement to the true value | Closeness of repeated measurements to each other |
| Affected by systematic error | Affected by random error |
Reliability and Validity
Reliability is the consistency of results across repeated trials. Validity is whether an instrument or test actually measures what it claims to measure. A tool can be reliable without being valid (consistently wrong), but a valid tool must also be reasonably reliable.
Types of Statistical Errors
- Random error — unpredictable fluctuations from uncontrollable factors; reduced by averaging repeated readings.
- Systematic error — a consistent, repeatable bias in one direction (e.g., a miscalibrated instrument).
- Human error — mistakes from the observer (misreading, recording errors).
- Instrumental error — arises from the limitations or faults of the measuring device.
- Sampling error — the difference between a sample statistic and the true population value, due to the sample not perfectly representing the population.
- Measurement error — the difference between an observed and true value during data recording.
- Experimental error — arises from limitations in the experimental setup or conditions.
Quantifying Error
- Absolute error = |measured value − true value|
- Relative error = absolute error ÷ true value
- Percentage error = relative error × 100
- Mean Absolute Error (MAE) — average of absolute errors across all data points; robust to large outliers.
- Mean Squared Error (MSE) — average of squared errors; penalises larger errors more heavily.
- Root Mean Square Error (RMSE) — square root of MSE; expressed in the original data unit, commonly used to judge model/prediction accuracy.
- Standard Error (SE) — an estimate of how much a sample mean is expected to vary from the true population mean.
Confidence Interval Basics
A confidence interval (e.g., 95%) gives a range within which the true population value is expected to lie, with a stated level of certainty — it communicates the precision of an estimate rather than a single fixed number.
Bias and Variance
Bias is a systematic deviation from the true value in one consistent direction. Variance is the spread or inconsistency of results across repeated measurements/trials. Good research aims to minimise both.
Outliers and Their Effect on Results
Outliers are data points that lie far outside the normal spread of the dataset. They can distort the mean, inflate variance, and mislead conclusions if not identified and justified (either corrected, explained, or excluded transparently).
How to Report Error Analysis in Research Papers
- State the type(s) of error present and their likely source.
- Report quantified error values (e.g., "RMSE = 2.3 units") alongside results, not hidden in an appendix.
- Mention steps taken to minimise error (calibration, repeated trials, larger sample size).
- Acknowledge remaining uncertainty honestly in the limitations section.
Quick Recap
- Error is natural and quantifiable; a mistake is an avoidable blunder.
- Accuracy relates to systematic error; precision relates to random error.
- MAE, MSE, RMSE, and Standard Error are the standard tools to quantify error/uncertainty.
- Always report error transparently rather than hiding or ignoring it.
Structure of the Results Section
- Restate briefly what was tested/measured.
- Present the data — tables, figures, key statistics.
- State the outcome factually, without interpretation or opinion.
- Order findings logically — usually following the sequence of objectives.
Results vs. Analysis vs. Discussion
| Section | Purpose |
|---|---|
| Results | Presents the raw findings/data objectively — "what was found" |
| Analysis | Applies statistical tests/comparisons to the results — "what the numbers show" |
| Discussion | Interprets meaning, compares with existing literature, explains implications — "what it means" |
Writing Statistical Findings Professionally
Use precise, neutral, past-tense language. Avoid vague words like "a lot," "huge," or "very good" — replace them with exact numbers and units.
Weak: "The results were much better with the new method."
Strong: "The new method achieved a mean accuracy of 92.4% (SD = 1.8), compared to 84.1% (SD = 2.6) for the baseline."
Strong: "The new method achieved a mean accuracy of 92.4% (SD = 1.8), compared to 84.1% (SD = 2.6) for the baseline."
Reporting Mean, SD, Error, p-value, and Confidence Interval
Standard convention: report the mean with its standard deviation, e.g., "M = 45.2, SD = 3.1." For hypothesis tests, report the test statistic and p-value, e.g., "t(28) = 2.14, p = 0.041," and where relevant, the confidence interval, e.g., "95% CI [1.2, 8.7]."
Explaining Graphs and Tables in Text
Point the reader to the specific trend, not just the existence of the figure. Reference it by number ("As Table 2 shows...") and describe the pattern, not simply re-list every value already visible in the table.
Connecting Statistical Results to Research Objectives
Every results paragraph should loop back to the objective it addresses, closing the logical circle between what was planned (Module 1) and what was found.
Statistical Significance vs. Practical Significance
Statistical significance (e.g., p < 0.05) indicates the result is unlikely due to chance. Practical significance asks whether the difference is large enough to matter in the real world. A result can be statistically significant but practically trivial, especially with very large sample sizes.
Ethical Use of Statistics
- Report all relevant results — not just the ones supporting the hypothesis (avoid cherry-picking).
- Never alter raw data to fit a desired outcome.
- Disclose sample size, exclusions, and any data cleaning performed.
Avoiding Data Manipulation and Misleading Graphs
- Do not truncate a y-axis to visually exaggerate small differences.
- Keep scale intervals consistent and clearly labelled.
- Do not selectively omit data points that contradict the hypothesis.
Common Statistical Errors in Undergraduate Research Papers
- Reporting the mean without any measure of spread (SD/variance).
- Confusing correlation with causation.
- Using an inappropriate statistical test for the data type.
- Overstating conclusions from a very small sample size.
- Copy-pasting interpretation language instead of writing original analysis.
Quick Recap
- Results = facts; Analysis = statistical processing; Discussion = interpretation and meaning.
- Always report a spread measure alongside a mean, and a p-value alongside any significance claim.
- Statistical significance ≠ practical significance.
- Report all data honestly — never manipulate axes or cherry-pick results.