viva-self-check-guide

Preparing for Your Viva: How to Evidence-Check Your Own Thesis Before the Examiners Do

The weeks before a thesis defence are strange. The writing is finished, the document is submitted, and yet the most demanding reader has not yet engaged with it. Examiners will probe your argument, question your methods and, increasingly, check your citations. The best preparation is to read your thesis the way they will, systematically and sceptically. This guide describes a practical self-examination process and explains how an evidence-traceable research agent such as Yila AI can help you verify claims and rehearse answers without replacing your own understanding.

Why self-examination works

Examiners usually read a thesis with three questions in mind. Does the candidate understand the field? Is the original contribution genuine and significant? Are the methods and evidence sound enough to support the conclusions? If you answer those questions about your own work before the viva, you will not be surprised by most of what is asked.

Step 1: Write a two-page summary from memory

Without opening the thesis, write a two-page summary: the problem, the gap, your research questions, your methods, your main findings and your contribution. Then compare it with the thesis. Where your summary is vague or differs from the text, you have found a weak point. Examiners often begin with "Can you summarise your thesis in a few minutes?" and this exercise prepares you directly.

Step 2: Map every major claim to its evidence

Go through each chapter and list the major claims, especially in the introduction, discussion and conclusion. For each claim, note the evidence that supports it: your own results, a cited source, or reasoning. Claims that rest on a single source or on reasoning alone are natural targets for questions.

This is where an agent can save time. Yila can read your thesis and the sources you cite, and for a given claim it can locate the passage in the cited paper that supports it, linking directly to that passage. When a citation does not support the claim as written, you find out before the examiner does, and you can prepare an honest answer or an erratum.

Step 3: Audit a sample of citations

Pick at least twenty citations at random and check each against the source. Does the paper exist with the details given? Does it support the specific sentence? Is the claim represented fairly? Most theses contain a few misattributions, usually innocent ones such as citing a review instead of the original study. Knowing about them in advance lets you address them calmly. If you used AI writing tools at any point, extend this audit to every reference; fabricated citations are a serious problem and examiners are alert to them.

Step 4: Interrogate your methods

For each empirical chapter, write down the three most likely methodological criticisms. Typical ones include sample size and power, selection bias, measurement validity, choice of statistical model, handling of missing data and generalisability. For each criticism, prepare a short answer: why you made the choice, what alternatives you considered, and what the limitation means for interpretation.

If your analysis was done in code, rerun it from scratch before the viva. Confirm that every number in the thesis matches the output. Examiners occasionally ask for a specific figure to be explained in detail, and discovering a discrepancy in the room is unpleasant. Yila returns reusable analysis code alongside results, which makes this kind of rerun straightforward if you used it during your project.

Step 5: Position your contribution against recent work

Theses take years, and the literature moves. Search for papers published since you finished your literature review, especially any that address the same gap. Examiners may ask how your work relates to a study that appeared last month. You do not need to have cited it, but you should be able to discuss it.

An agent that searches the literature and summarises new papers with source links is useful here. Ask for recent work on your specific question and read the linked passages to judge relevance. Prepare one or two sentences on how each relevant new paper compares with your findings.

Step 6: Rehearse the hard questions

Draft a list of the twenty hardest questions you can imagine and answer them aloud. Include:

• What is the single most important contribution of this thesis?
• What would you do differently if you started again?
• What is the weakest part of your argument?
• How do your findings generalise beyond your sample or setting?
• Which result surprised you, and why?
• What are the practical implications?
• What is the next study you would run?

Ask a colleague or supervisor to run a mock viva using your list and their own questions. The point is not to memorise answers but to practise thinking aloud under pressure.

Step 7: Prepare a corrections list

Keep a list of typos, unclear passages and minor errors you find while rereading. Bringing a short, organised corrections list to the viva shows care and often reduces the corrections examiners request. Do not try to hide errors; acknowledging them confidently is far better than being caught out.

Using AI responsibly in preparation

There is a difference between using AI to check your work and using it to understand your work for you. Use an agent to locate evidence, flag possible misattributions, find recent papers and rerun analyses. Do not use it to generate answers you have not thought through. Examiners can tell when a candidate is reciting rather than reasoning.

The value of a tool like Yila in this context is traceability: every claim it makes points back to a source passage you can read. That supports exactly the kind of verification a viva demands, while leaving the judgement and the explanation to you.

On the day

Bring a clean copy of the thesis with tabs marking key sections, your corrections list and a notebook. Listen to each question fully before answering. It is acceptable to pause, to ask for clarification and to say that you do not know something, followed by how you would find out. Examiners generally want you to pass; they are testing whether you are an independent researcher, not whether you are infallible.

Summary

Self-examination before a viva means summarising from memory, mapping claims to evidence, auditing citations, interrogating methods, positioning against recent work and rehearsing hard questions. An evidence-traceable research agent such as Yila AI at https://yila.ai can accelerate the verification steps while keeping every link to the source visible. The goal is to walk into the room knowing your thesis as well as, or better than, the examiners who have just read it.