Of all the decisions in a dissertation, methodology is one of the few that’s genuinely difficult to undo once you’re underway. Choose the wrong approach, and you may not discover it until you’re deep into analysis, with months of data collection already behind you. That’s why the most expensive methodology mistake isn’t picking the “wrong” one in some abstract sense — it’s picking the one that feels easiest, rather than the one that actually answers your research question.
How this mistake usually happens
Imagine a student wants to understand why employee turnover is high in a particular industry. A survey feels like the obvious choice: quick to distribute, straightforward to analyse, and comfortably familiar from undergraduate study. So that’s what they go with.
Months later, during analysis, the problem becomes clear. The survey has produced numbers, but not the reasoning behind them. Turnover is driven by things employees rarely capture fully on a ten-point scale – specific incidents, layered frustrations, decisions that don’t reduce neatly to a number. What the research question actually needed was interviews, which would have surfaced that reasoning directly. By the time this becomes obvious, the data is already collected, and there usually isn’t time or scope left to start again.
Nothing about this student’s methodological knowledge was wrong. The mistake was choosing the option that felt most manageable at the time, rather than testing it against what the research question actually required.
Why this is so easy to miss early on
At the proposal stage, every methodology can sound reasonable in the abstract. It’s only once you’re holding actual data that the mismatch becomes obvious – and by then, the cost of switching is high. This is precisely why it’s worth deliberately stress-testing your methodology choice before committing to it, rather than assuming that “it’s a recognised method” is the same as “it’s the right method for this question.”
A simple check before you commit
Before finalising your methodology, ask yourself one direct question:
If this data collection goes exactly to plan, will the results actually tell me what I need to know?
Not whether you’ll be able to collect the data easily. Not whether it’s a well-established method. Specifically: imagine receiving a clean, complete dataset back from this exact approach. Would it let you answer your research question, or would you be left with results that are technically correct but somehow beside the point?
If you can picture getting that data back and still being unable to answer your question, that’s the moment to reconsider — before data collection, not after.
A few practical questions to work through
- Does this methodology capture why, or only what? If your question is about reasons, motivations, or meaning, a purely quantitative approach may leave a gap no amount of statistical testing can close.
- Does it match the scale of your question? A methodology designed for broad generalisation won’t necessarily suit a question about a specific, smaller context, and vice versa.
- Would a different methodology answer the same question more directly, even if it’s less familiar to you? Familiarity is a reasonable factor to weigh, but it shouldn’t be the only deciding one.
If you’re not sure your methodology fits
This is one of the costliest mistakes to catch late, and one of the easiest to catch early with an outside perspective. If you’re choosing a methodology now, or have already started and have a nagging doubt about whether it actually fits your question, book a free 15-minute introductory call and we can talk through whether it holds up.