Practical research guidance, explained in detail.

Objectives should state exactly what the study intends to measure, compare, describe, determine or evaluate. They should be consistent with the research title, study design, variables and planned analysis.

• Use specific action verbs such as assess, compare, determine, evaluate or identify.
• Avoid vague objectives such as “to know” or “to understand”.
• Separate the primary objective from supporting secondary objectives.
• Each important objective should have corresponding variables and an analysis plan.

Practical check: if you cannot identify what data are needed to answer an objective, the objective is probably too broad or unclear.

A research gap is a meaningful unanswered question, limitation, inconsistency or under-studied area within the existing evidence. A gap should emerge from critical reading rather than from simply stating that “few studies are available”.

• Look for inconsistent findings between previous studies.
• Identify populations, outcomes or settings that have not been adequately studied.
• Check whether earlier studies had important methodological limitations.
• Distinguish between a genuine scientific gap and a merely local absence of data.

A defensible gap should lead logically to the problem statement, research question and study objectives.

Study design is selected according to the research question, whether an intervention is involved, the timing of measurements, availability of comparison groups and feasibility.

• Cross-sectional: useful for prevalence, characteristics and associations measured at one period.
• Case-control: useful for comparing previous exposures between participants with and without an outcome.
• Cohort: useful for following exposed and unexposed groups over time.
• Experimental/interventional: used when the investigator assigns an intervention under an appropriate protocol.

The most complex design is not automatically the best design. The correct design is the one that answers the objective with acceptable bias, ethics and feasibility.

Sample size should be based on the primary outcome and the statistical question, not on an arbitrary target. The calculation may require expected prevalence, effect size, variability, confidence level, statistical power and group allocation.

• Define the primary outcome before calculating sample size.
• Use assumptions supported by prior literature or pilot data where possible.
• Allow for expected non-response, attrition or incomplete records.
• Report the assumptions and method used so the calculation is reproducible.

Every important variable should have an operational definition that explains exactly how it will be identified, measured, categorized or calculated. This improves consistency during data collection and reduces ambiguity during analysis.

• Define exposure/intervention variables, outcomes, predictors and potential confounders.
• Specify units, cut-offs and scoring methods.
• Use validated instruments where appropriate.
• Distinguish primary outcomes from secondary or exploratory outcomes.

For most original research articles, the IMRaD structure—Introduction, Methods, Results and Discussion—provides the central organization, accompanied by title, abstract, references and required declarations.

• Introduction: establishes context, research gap and objective.
• Methods: describes design, participants, measurements and analysis sufficiently for evaluation or replication.
• Results: reports findings without unnecessary interpretation.
• Discussion: interprets findings, compares evidence and addresses strengths and limitations.

The Results section answers “what did the study find?” whereas the Discussion answers “what do those findings mean?”. Mixing interpretation into the results can make a manuscript difficult to follow.

• Results should present objective-wise findings using concise text, tables and figures.
• Avoid repeating every number from a table in the narrative.
• Discussion should begin with the most important findings, not repeat the entire Results section.
• Compare findings with relevant literature and explain plausible reasons for similarities or differences.

A strong abstract is a faithful miniature of the full work. It should contain enough information for a reader to understand the question, approach, major findings and conclusion without overstating the evidence.

• Follow the journal’s required structured or unstructured format.
• Include actual key numerical findings when appropriate.
• Do not introduce methods, results or conclusions that do not appear in the manuscript.
• Avoid citations unless specifically allowed.

A thesis Discussion should be organized around the study objectives or major findings. Each major result should be interpreted, compared with previous evidence and placed in the context of the study design and population.

• State the important finding.
• Explain its scientific or clinical interpretation.
• Compare with supportive and contrasting studies.
• Discuss plausible explanations and methodological considerations.
• Address strengths, limitations and implications without making claims beyond the data.

Test selection depends on the research objective, outcome type, number of groups, independence or pairing of observations, distributional assumptions and whether adjustment for other variables is required.

• Continuous versus categorical outcome.
• Two groups versus multiple groups.
• Independent versus paired/repeated observations.
• Normally distributed versus non-normal data where relevant.
• Unadjusted comparison versus multivariable modelling.

A p-value describes how compatible the observed data are with the null hypothesis under the assumptions of the statistical model. It does not measure the magnitude, clinical importance or probability that the hypothesis is true.

Interpret p-values together with effect sizes, confidence intervals, study design, sample size and clinical/scientific relevance. A very small difference can be statistically significant in a large sample yet have little practical importance.

A confidence interval provides a range of values compatible with the estimated effect under the statistical model and gives information about precision. Wide intervals usually indicate greater uncertainty.

• Report confidence intervals with effect estimates such as mean differences, odds ratios, relative risks or regression coefficients.
• Do not rely only on whether the interval crosses a null value.
• Consider whether the entire interval contains clinically important or unimportant effects.

Tables and figures should reduce cognitive load and communicate a specific message. They should not merely decorate the report.

• Use clear titles and consistent units.
• Avoid displaying the same information in both a table and graph without a reason.
• Use appropriate denominators and clearly state missing data where relevant.
• Keep decimal places and statistical notation consistent.
• Choose graph types that match the variable and comparison being shown.

Start with journal scope and readership. A highly ranked journal is not suitable if it rarely publishes the manuscript’s subject, methodology or article type.

• Check aims and scope and review recent related articles.
• Confirm whether the journal accepts the intended article type.
• Review indexing in databases relevant to your academic purpose.
• Understand open-access or subscription models and legitimate publication charges.
• Assess publication ethics and avoid journals displaying predatory characteristics.

Indexing indicates inclusion in a bibliographic database, while journal metrics summarize citation-related performance using different calculation methods. Quartiles usually position journals within subject categories according to a particular database or metric.

These concepts should not be treated as interchangeable. Always verify the journal directly in the relevant database rather than relying only on claims displayed on a journal website.

Requirements vary by journal, but a submission package may include the main manuscript, title page, cover letter, tables/figures, highlights, graphical abstract, supplementary files and declarations.

• Authorship and contribution statement.
• Conflict-of-interest disclosure.
• Funding information.
• Ethics approval and consent statements where applicable.
• Data availability statement where required.

Responses should be respectful, specific and point-by-point. Each reviewer comment should be quoted or clearly identified, followed by the response and location of the revision.

• Thank the reviewer for constructive points without excessive repetition.
• State exactly what was changed and where.
• If a suggestion is not adopted, explain the scientific or methodological reason.
• Ensure that all changes described in the response are actually present in the revised manuscript.

When human participants are involved, informed consent generally ensures that individuals understand the study purpose, procedures, possible risks/benefits, confidentiality and voluntary nature of participation. Exact requirements depend on the study and ethics approval.

Consent is a process, not merely a signature. Participant information should be understandable and coercion should be avoided.

Authorship should reflect substantial intellectual contribution and accountability for the work. Gift, honorary or purchased authorship undermines research integrity.

• Discuss authorship criteria early.
• Document contributions transparently.
• Ensure all authors review and approve the submitted manuscript.
• Acknowledge contributions that do not qualify for authorship appropriately.

Plagiarism involves presenting another person’s words, ideas or work without appropriate acknowledgement. Similarity software can help identify text overlap, but the numerical similarity percentage alone does not determine plagiarism.

• Paraphrase by genuinely rewriting the idea in your own scholarly expression.
• Cite the original source for ideas and evidence.
• Use quotation sparingly and according to academic conventions.
• Do not manipulate text merely to reduce a similarity score.

Research findings must reflect the data actually collected. Fabrication, falsification, selective deletion of inconvenient data or undocumented changes to outcomes can invalidate a study and create serious ethical problems.

• Maintain clear data dictionaries and audit trails.
• Document cleaning rules and exclusions.
• Preserve original/raw data according to applicable policies.
• Report missing data and deviations transparently.

A research presentation should tell a coherent story rather than reproduce a thesis on slides. Use short sections and make the main finding visually obvious.

• Title and context.
• Research gap/objective.
• Essential methodology.
• Key results with readable tables/figures.
• Interpretation, conclusion and implications.

Keep slide text concise and reserve detailed explanation for the speaker.

A poster should be readable from a reasonable distance and allow the viewer to understand the study in a few minutes. Use a clear visual hierarchy, concise text and well-labelled figures.

• Use a strong title and clear section order.
• Prioritize key results rather than including every analysis.
• Use sufficiently large fonts.
• Maintain consistent alignment and spacing.
• Avoid overcrowding with long paragraphs.

A journal club should move beyond summarizing the paper. It should evaluate the research question, methodology, validity, results, limitations and relevance.

• Identify the clinical/scientific question.
• Explain the study design and population.
• Highlight the most important numerical findings.
• Critically assess bias and limitations.
• Conclude with applicability and discussion points.