Thesis & Dissertation Support
From concept to completion — topic, methodology, analysis, writing and final review.
Read more
Research design, sampling, hypotheses, data collection and the analysis plan.
Chapter 3 is the chapter examiners read most carefully and candidates write most quickly. That imbalance explains a large share of the corrections handed out at viva.
The six segments on the poster are the parts that have to agree with one another: research design, sampling technique, data collection, analysis plan, hypothesis development and interpretation. They are drawn as a wheel because that is how they behave — change one and the others must move. A shift from a cross-sectional to a longitudinal design changes the sampling, the instrument, the analysis and what the hypotheses can claim.
Most methodology chapters we are sent describe what the candidate did. A strong methodology chapter justifies it: why this design and not the alternative, why this sampling frame, why this sample size and by what calculation, why this instrument and what evidence exists for its validity, and why these tests will answer these hypotheses.
That justification is what an examiner is testing. The question is rarely 'what did you do' — it is 'why was that the right thing to do, and what would have happened if you had done otherwise'. A candidate who has written the reasoning down can answer it; one who has only described the procedure cannot.
The whole of Chapter 3, and the decisions that have to be made before it.
The most common answer we hear to 'how did you arrive at 384 respondents' is that a paper the candidate read used that number. It comes from a formula for an infinite population at ninety-five per cent confidence and a five per cent margin — perfectly respectable, but only if that is what your study needs, and only if you can say so.
The right calculation depends on the analysis. A structural equation model has requirements driven by the number of indicators and paths. A moderated regression needs power for an interaction effect, which is typically much larger than for a main effect. A comparison between two groups needs a stated minimum detectable difference. G*Power will do all of these, and it takes twenty minutes.
We run it, show the parameters, and put the output in the chapter. If your achieved sample is below what the calculation requires, that is reported as a limitation rather than ignored — which is a far better position than being asked about it for the first time in the viva.
The objectives decide everything downstream. If they are vague or unanswerable we say so first, because no methodology can rescue them.
The design that can actually answer those objectives, with the alternatives considered and the reason for rejecting them recorded.
Population, frame, technique and a calculated size with the parameters stated.
Adapted or original, piloted, with reliability and validity evidence gathered before the main collection.
Each hypothesis matched to a named test, with the assumptions those tests require noted in advance.
Chapter 3 in your university's format, with every choice justified and the limitations stated honestly.
Writing the analysis plan before collection is the single habit that most improves a thesis. It forces you to check that every hypothesis has a test, that every test has the data it needs, and that the instrument actually captures the variables at the measurement level the test requires. Ordinal data collected on a five-point scale will not support an analysis that assumed an interval measure.
It also protects you from the temptation that arrives when the results come back. If the plan says the hypothesis will be tested by hierarchical regression, and the regression does not reach significance, the honest reporting is that it did not. Deciding the test after seeing the data — trying several and reporting the one that worked — inflates the false positive rate and is detectable by anyone who reads the numbers carefully.
Where an exploratory analysis genuinely is warranted, it is labelled exploratory and reported separately from the confirmatory tests. That distinction costs nothing and buys a great deal of credibility.
| What you receive |
|---|
| Research design justification with alternatives considered |
| Sampling plan and calculated sample size with parameters |
| Instrument, pilot results and validity evidence |
| Hypotheses aligned to objectives |
| A written analysis plan naming each test |
| Ethics and data-handling section |
| The complete Chapter 3 in your university's format |
| A walk-through so you can defend every choice |
Yes, and this is one of the most common reasons people come to us. We read the chapter against your objectives and identify precisely where the reasoning breaks — usually a design that cannot answer an objective, a sample that cannot support the analysis, or an instrument measuring something other than the construct named.
Usually yes, with citation, and sometimes with the author's permission depending on the scale. What you must not do is change items and keep citing the original validation — an adapted scale needs its own reliability and validity evidence, gathered in your pilot.
Thirty to fifty respondents is typical for checking reliability and for finding questions people misread. It is not enough for factor analysis, so if you need to validate a new scale the pilot has to be larger, and we will tell you how much larger.
Yes. Sampling logic, saturation, interview protocol design, coding framework, inter-coder agreement and the audit trail an examiner will ask for — the requirements are different from quantitative work but no less specific.
From concept to completion — topic, methodology, analysis, writing and final review.
Read more
Reliable analysis, accurate results — from raw responses to a defensible results chapter.
Read more
Search, screen, evaluate and synthesise — a review that builds an argument.
Read moreSend your topic, your dataset or one draft chapter. We will tell you honestly what it needs — before you pay anything. The first consultation is free.