Research planning guide · Updated October 14, 2026
Research methodology explains why your design, data collection and analysis can answer your research question. Methods are the specific procedures you use. Start with the question and the evidence you need, then choose feasible procedures.
For example: among first-year university students, does guided AI feedback improve independent writing performance over one semester compared with teacher feedback? Specify the population, comparison, outcome, setting and time frame. An exploratory interview study about students’ experiences would answer a different question.
Check existing work before choosing the design. Record your search terms, databases, dates and screening decisions so another researcher can understand the evidence behind your choices.
Search papers for your research questionThe free tool searches Crossref records. Add subject databases and citation searching when coverage matters.
These categories overlap: an experimental or observational design can use quantitative data, and a review may synthesize qualitative or quantitative evidence.
How much, how often, or what association?
A survey estimates how frequently students use AI feedback. Define the sampling frame, measurement, missing-data handling and analysis before collection.
Interpretation: An association from a survey alone does not establish a causal effect.
How do people understand or experience something?
Interviews explore how students use feedback. Explain recruitment, interview guide, coding, researcher reflexivity and how interpretations are checked.
Interpretation: A small purposive sample supports contextual insight; it does not estimate population prevalence.
Which pattern occurs, and how can it be explained?
Combine a survey with interviews and specify where the two strands connect: sampling, interpretation or both.
Interpretation: Collecting two data types is insufficient without explaining how they are integrated.
Does an intervention cause a change under specified conditions?
Randomly assign a feedback intervention and comparison, predefine outcomes and explain allocation and analysis.
Interpretation: Generalization still depends on the population, setting, adherence and follow-up.
What happens in an existing setting?
A cohort follows students over a semester. Define eligibility, exposures, outcomes and plausible confounders.
Interpretation: Selection, measurement and confounding can affect the interpretation.
What do existing studies collectively establish?
Define a review question and protocol, search relevant databases, screen against criteria and assess study limitations.
Interpretation: A quick keyword search is not an exhaustive systematic review.
“We will compare guided AI feedback with teacher feedback over one semester among eligible first-year students. Writing performance will be assessed using a predefined rubric by assessors blinded to allocation where feasible. The protocol will define recruitment, allocation, primary outcome, sample-size rationale and missing-data procedures before data collection.”
This is a planning example. Replace every detail with your actual protocol; report only approvals and procedures that occurred. Choose the tense to match a proposal or a completed study.
Consult your discipline’s methods literature and institutional requirements. This guide offers a starting structure, not approval of a particular design.
