RESEARCH GUIDE
Developing a Dissertation Research Plan
Sample, Measures and Analysis Settled First
A dissertation research plan sets out how a study will be carried out, and its statistical part fixes the sample, the measures and the analysis before any data are collected. Gaps in those three are the ones a committee or the data itself exposes later, so they are closed while the plan can still change.
A dissertation research plan describes the design, the participants or data source, the measures, the procedure for collecting data and the analysis that will answer each question. Its statistical part ties these together: the sample must be large enough to estimate what the question asks, each measure must capture its variable in this population, and each analysis must follow from the design. A dissertation methods plan written at this level can be checked against the research questions item by item, and a dissertation data collection plan shows how records will arrive in a form the analysis can use. OnlineSPSS, a specialist statistical consultancy for academic, clinical and professional research, supports the sample size reasoning, the measurement plan and the analysis plan, while the design choices, the ethics approval and the chapter remain the researcher’s.
PLAN COMPONENTS
Components of the Plan and Their Statistical Content
Each part of a plan carries a statistical commitment, and reading the plan component by component shows where a decision is missing or cannot be carried out.
Design and Sampling How participants or records are selected, how many are needed and what population the results can describe, which is where the sample size reasoning belongs.
Measures and Scoring Each instrument or record field, how it is scored and what evidence supports it in this population, since a score is only as useful as the variable it represents.
Data Collection Procedure Who collects which data, when and in what format, with identifiers that let records be linked, so the data arrive in a form the analysis can use.
Analysis for Each Question One named analysis for each question, with its outcome, explanatory variables and main assumptions to examine, so that nothing is chosen after the data are seen.
Data Management and Ethics How data are stored, cleaned and protected, who has access and how consent limits use, which an ethics board often reviews alongside the statistical plan.
Timeline and Milestones When recruitment, collection and analysis take place, with room for slower recruitment, because a delayed sample is a common reason a plan has to change.
PLAN COMPARISON
Quantitative, Qualitative and Mixed Plans Compared
A plan is built around the evidence the study will produce, and the statistical questions change with it, as the table shows for the three main approaches.
| Feature | Quantitative | Qualitative | Mixed methods |
|---|---|---|---|
| How the sample is justified | Sample size reasoning from the precision or power needed | Purposive selection, with the amount of data judged against the aim and the richness of the material | Each strand has its own logic, and the link between strands is stated |
| Data collected | Measures, records or scores | Interviews, observations or documents | Both, with a plan for how they relate |
| Analysis plan | A named statistical model for each question | A coding approach such as thematic or framework analysis | Each strand analyzed on its own terms and then integrated |
| Quality criteria | Validity, reliability and precision | Trustworthiness, reflexivity and transparent coding | Both sets, plus the quality of the integration |
| Main planning risk | A sample too small for the question | Material too thin or too narrow for the aim | Strands that never meet in the analysis |
KEY PLANNING DECISIONS
Statistical Decisions to Settle in the Plan
These decisions turn a plan from a description into something that can be carried out and checked, and each is easier to settle before recruitment than after.
Justify the Sample Size Base the target on the precision or power needed for the primary question, using an effect size justified from earlier work or practical importance, not on convention.
Choose Measures With Evidence Prefer instruments with reliability and validity evidence in a similar population, and record how each is scored, including reverse-coded items and missing responses.
Fix the Primary Outcome Name one primary outcome for each question and label the others secondary, so the plan states in advance which result the main conclusion rests on.
Plan for Missing Data State the expected amount of missing data, how it will be prevented and how it will be handled, because a plan silent on this leaves the choice to the analysis stage.
Link the Analysis to the Design Say whether the data are independent, paired, clustered or repeated, since this decides whether the planned model matches the way the data are collected.
Set Rules for Deviations Note how departures from the plan will be recorded and reported, which keeps later changes visible and the prespecified analyses distinct from the rest.
COMMON ISSUES
Gaps That Surface Once Data Collection Starts
Gaps are easy to close on paper and hard to close in the field, and each of the following is easier to fix in a draft than in a dataset.
Sample Size by Convention A target taken from a rule of thumb or from another study may give an interval too wide to answer the question, so it is checked against the precision needed.
An Unvalidated Instrument A homemade or newly translated questionnaire has no known reliability in this population, so the plan adds a pilot or reports reliability from the collected data.
Analysis Chosen From the Data A plan that says the analysis will depend on what the data look like lets choices follow the results, so the main analysis is named in advance with its fallback.
No Way to Link Records Data gathered at several points or from several sources without a shared identifier cannot be paired, which removes the stronger analyses from the options.
Recruitment Taken for Granted Plans often assume the sample will arrive on schedule, although response and access vary, so the plan states a realistic yield and what follows if it falls short.
Measurement Points Left Vague A plan that says outcomes will be measured after the intervention, without fixing when, lets follow-up drift between participants, so each measurement point is given a date or a window.
REPORTING
How the Plan Becomes the Methods Section
A well-formed plan converts almost directly into the methods and results, and the report then states where the study followed the plan and where it did not.
The sample size reasoning, the measures with their scoring and the named analyses move into the methods chapter largely as planned. For the methods chapter, the APA Journal Article Reporting Standards list the sampling procedure, the sample size and how it was determined, the measures with their reliability in the sample, the handling of missing data and the analyses carried out. Where the plan was preregistered or approved by a committee, the report refers to it and describes any change. A short deviation log kept during data collection makes that description accurate.
STATISTICAL METHODS
Analyses a Research Plan Commonly Names
The analysis named for each question depends on its outcome and design, and plans with a quantitative strand most often name one of these, along with a reliability check and a missing-data approach.
Independent Samples t-Test
Comparing two groups on a measure
Repeated Measures ANOVA
The same people on several occasions
Multiple Linear Regression
Predictors of a continuous outcome
Logistic Regression
Predictors of a yes or no outcome
Reliability Analysis (Cronbach's Alpha)
Consistency of multi-item scales
Missing Data Analysis
Planning for incomplete responses
ANALYSIS SERVICES
From the Sample Size to the Final Tables
The order of the cards matches the order in which a dissertation reaches each stage, from the sample size argument to tables built on the collected data.
Research Planning & Study Design
Sample size and precision reasoning, a primary outcome for each question, a measurement plan and an analysis plan, recorded before recruitment or data extraction.
Quantitative Data Analysis
The planned analyses carried out on the collected data, with diagnostics, effect estimates and confidence intervals, and any departure from the plan noted.
Results Reporting & Interpretation
Methods wording and tables that follow the plan, so each planned analysis appears in the results in the same terms and in the same order.
Independent Statistical Review
A plan already drafted, or approved by a committee, is read for sample adequacy, evidence for the measures and whether each analysis fits the design.
FREQUENTLY ASKED QUESTIONS
Questions About the Dissertation Research Plan
A recruitment target that cannot be met, piloting, evidence for an instrument and the data management section are the points that arise most while a plan is drafted.
What if the sample I can recruit is smaller than the calculation requires?
The plan should say so openly and show what the smaller sample can still estimate, instead of presenting an underpowered test as if it could settle the question. Options include narrowing the question or the number of outcomes, using a more efficient design such as paired or repeated measurements, extending recruitment, or reporting estimates with their intervals. The calculation for the original target can be run in G*Power, with the assumed effect size stated and justified.
Should I pilot the instruments and procedures before the main study?
A pilot is worth planning when an instrument is new or translated, when the procedure is complex or when recruitment and completion rates are uncertain, because it tests whether the study can run as described. Its purpose is feasibility: whether items are understood, how long collection takes and how many people agree and finish. A small pilot gives too imprecise an effect size to fix the sample size on its own, and it is not used as evidence for the hypotheses.
What evidence do I need that my instrument is suitable?
Reliability and validity evidence from earlier studies in a similar population is the usual starting point, together with reliability computed in your own sample. A translated or adapted instrument needs more, such as a documented translation procedure and a check of its structure in the new setting. Where no suitable evidence exists, a pilot or a validation step belongs in the plan.
What belongs in the data management part of a plan?
How data are stored, de-identified and backed up, who has access, how long they are kept and how the dataset is documented with a codebook, all consistent with the consent given and with institutional policy. Saved scripts for cleaning and analysis also belong here, because they let the work be rerun and checked.
How do you protect my research data and confidentiality?
Your dissertation data and project details are handled confidentially. Project information is not published, and researcher names and institutions are not used publicly without permission. An NDA is available where required and can be arranged before sensitive materials are shared. Read more about confidentiality and data security.
What happens when I request an assessment, and how long does it take?
You share your research question, design and relevant materials, such as a protocol, an analysis plan or a dataset, and a statistically trained Research Coordinator reviews the request and responds the same day. The analysis, deliverables, timeline and one fixed project fee are then confirmed before any work begins, so the timeline follows from the agreed scope.
Still have a question? Find more answers in our full FAQ.
NEXT STEP
Stress-Test Your Plan
Before You Collect Data
A draft plan, the research questions and the instruments or data source are enough to start. The reply names the statistical gaps worth closing before recruitment and the scope of any support.
No Commitment Required
Strictly Confidential
Fixed Project Fee
In-House PhD-Level Statisticians
Same-Day Response
Serving Researchers Since 2014