RESEARCH GUIDE
Choosing a Dissertation Topic
Testing Data, Measures and Sample Before You Commit
Choosing a dissertation topic is a decision for the candidate and the committee, and a statistical check can show early whether the question can be answered with the data the project will be able to obtain. The check looks at access, measures and precision, and never at which topic is worth pursuing.
A researchable dissertation topic is one whose question can be answered with data that exist or can be collected within the limits of the project. For a quantitative topic that means a named population, an outcome that can be measured, a source of data that can be reached and a sample large enough to give a useful answer. Dissertation topic feasibility is judged on these points together, because a topic that is attractive but lacks a measure or has too few cases cannot be rescued later by the analysis, which is why choosing a dissertation topic benefits from a data check before the proposal is written. OnlineSPSS, a specialist statistical consultancy for academic, clinical and professional research, tests shortlisted topics against data access, measures and expected precision as part of the support listed under Services, while the topic, its contribution and the final decision remain with the researcher and the committee.
DATA ROUTES
Types of Dissertation Topic and the Data Each One Depends On
Topics differ less by subject than by where their data come from, and the source decides how much control the researcher has and what can go wrong.
Survey Topic Asks a defined group about attitudes, behavior or experience, so it depends on reaching enough respondents and on measures that suit the construct.
Experimental Topic Manipulates a factor and compares conditions, which can support a causal reading but needs participants, materials and time to run the sessions.
Existing-Dataset Topic Uses data already collected by others, which saves time but fixes the available variables, so the key measure is confirmed before commitment.
Records-Based Topic Draws on clinical, administrative or company records, where access approvals and the quality of the recorded variables decide feasibility.
Qualitative Topic Seeks meanings or experiences through interviews or documents, where feasibility means reachable participants, sufficient material and a fitting method.
Mixed Methods Topic Combines two strands, so feasibility must hold for the numerical part, the qualitative part and the work of bringing the two together.
ROUTE COMPARISON
New Survey, Existing Dataset and Experiment Compared
The same topic can often be approached by more than one route, and the table sets out what each route usually asks of a project.
| Feature | New survey | Existing dataset | Experiment |
|---|---|---|---|
| Control over measures | Full, within what respondents will answer | Limited to what was collected | Full for the manipulated factor |
| Time before analysis | Includes design, approval and fieldwork | Short, though access may need approval | Often the longest, as recruitment and sessions are needed |
| Sample size | Set by recruitment and response | Fixed, and should be checked first | Set by recruitment and dropout |
| Typical risk | Low response and weak measures | A key variable is missing or defined differently | Small groups and attrition |
| Claims supported | Description and association | Depend on the original design | Causal, if participants were randomized |
KEY PLANNING DECISIONS
Questions to Answer Before a Topic Is Fixed
Each question is answered with evidence from the data source and the literature, and a topic that fails one of them is revised while it can still change.
Can the Data Be Reached? Confirm that the dataset, site or participants are available to you, and which approvals are needed, before the topic is written into a proposal.
Is the Key Variable Measured? Check that the outcome and the main explanatory variable exist in usable form, since a topic whose central construct is unmeasured cannot be tested.
How Many Cases Are Realistic? Estimate the sample the project can reach, allowing for refusals and dropout, and see what precision that sample gives for the quantity of interest.
Is a Comparison Group Available? Check that the group your question compares against, such as nonusers, an earlier period or another site, exists in the data source and was measured in the same way.
Will the Outcome Vary Enough? Check that the outcome varies in the population you can reach, since an event almost nobody experiences, or a score most respondents max out, leaves little to explain.
Can the Analysis Be Done in Time? Allow for cleaning, modeling and checking within the schedule, and for any method that you or your committee will need support to run.
COMMON ISSUES
Topic Problems That Surface Only at the Analysis Stage
Each of these can stall a project after approval, and a short exchange with the data owner while the topic is still open usually brings it to light.
A Dataset That Cannot Answer An existing dataset may lack the exposure, outcome or time ordering that the question needs, and discovering this late forces a different question.
Definitions That Differ Across Sites Data pooled from several clinics, schools or firms may define the same variable differently, so a topic that depends on combining them checks each definition before relying on the pooled file.
A Construct Without a Measure Words such as engagement or quality name ideas, and until each is tied to an instrument or record, the topic has nothing to analyze.
A Topic Too Broad for One Outcome A topic needing several outcomes, groups and time points multiplies the analyses and the sample, so it is narrowed to one primary question.
Gaps in Existing Records Records kept for care or administration often leave key fields blank, and more often for some groups than others, so the share and pattern of missing values are checked before a topic depends on them.
Access Assumed, Not Confirmed Permission to use records or recruit at a site can take months or fail, so the topic is not fixed on access that has only been discussed informally.
REPORTING
Recording the Feasibility Case in a Proposal
A proposal is stronger when it shows, and does not merely assert, that the topic is feasible, and the evidence is short and specific.
A feasibility case in a proposal normally names the data source and the access arranged, the measures with evidence of their suitability, the expected sample and the precision or power it gives, and the design with the claims it supports. Setting this evidence down now also gives the later methods chapter its starting point, since the same data source and measures will be described there. Institutional review and the committee usually expect the same elements, set out in the form their guidelines request.
STATISTICAL METHODS
Methods a Topic Commits You To
Choosing the topic also chooses a family of analyses, so it helps to see what the outcome and the design will require before the topic is final.
Multiple Linear Regression
A continuous outcome and several predictors
Logistic Regression
A binary outcome, with enough events
Independent Samples t-Test
Two groups on one measured outcome
Reliability Analysis (Cronbach's Alpha)
Evidence that a scale is consistent
Missing Data Analysis
Gaps in an existing dataset
Comparative & Hypothesis Testing
Comparing groups and conditions
ANALYSIS SERVICES
Statistical Input While the Topic Is Still Open
A feasibility check is most useful before the proposal is approved, and the same support can continue into the analysis and a final review if the topic goes ahead.
Research Planning & Study Design
Data access, measures, sample size and the analysis plan tested against the shortlisted topic, so the proposal rests on what the project can deliver.
Quantitative Data Analysis
Once the topic is approved, the models it calls for are run on your data, checked with diagnostics and reported as estimates with their intervals.
Qualitative & Mixed Methods Analysis
For topics built on interviews, documents or two strands, an analysis plan and a coding framework suited to the chosen method.
Independent Statistical Review
The statistical section of a draft proposal is read for measurable variables, a realistic sample and claims that match the design.
FREQUENTLY ASKED QUESTIONS
Questions About Choosing a Dissertation Topic
A feasibility check, an advisor's existing project, the limits of statistical advice and qualitative topics each get a short answer.
What does a statistical feasibility check look at?
It looks at whether the data, the measures and the sample can support the question, and whether the design can carry the claim the topic implies. In practice that means confirming access to the data, checking that the key variables exist, estimating the precision the likely sample would give and listing the analyses the topic would require. It does not judge whether the topic is original or important.
Can my topic come from my advisor's existing project?
It can, and joining an existing project often settles access, measures and sample size at once, but the question still has to be one you can call your own and answer with the data being collected. Agree early who controls the data, what you may publish from it and whether the variables your question needs are recorded in a usable form.
Will you choose or approve my topic?
No. The topic, the research question and the direction of the work belong to you and your committee. A statistical review can show whether a topic you have shortlisted is feasible with the data you describe and what the analysis would involve, but it does not select the topic or replace your advisor’s approval.
Is feasibility different for a qualitative topic?
Yes, because the questions concern access and sufficiency and not statistical power. The check covers whether participants or documents can be reached, whether the material will be rich enough for the question and whether the chosen method fits the aim. The amount of material is justified by the depth and variety needed and not by a power calculation.
How do you protect my research data and confidentiality?
Your topic ideas, proposal drafts and any data you share are handled confidentially, as are all project details. 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
Test a Shortlisted Topic
Before You Commit
Two or three shortlisted topics, each with a line on where its data would come from, give a feasibility check something concrete to work on while the choice between them is still open.
No Commitment Required
Strictly Confidential
Fixed Project Fee
In-House PhD-Level Statisticians
Same-Day Response
Serving Researchers Since 2014