Choosing a research method that fits the question you have
Qualitative or quantitative is the wrong first question. Start with what kind of claim you need to make, and the method usually chooses itself.
Name the claim first
Are you describing something, comparing two things, explaining why something happens, or predicting what happens next? Descriptive claims tolerate survey data. Causal claims need a design that rules out the obvious alternatives. Predictive claims need enough history to test against. When the claim is vague, every method looks defensible, which is exactly the problem.
Match the sample to the claim
A small purposive sample can carry a rich explanatory claim and cannot carry a population estimate. A large panel can carry an estimate and will not tell you why people answered as they did. Mixed designs are not a compromise; they are a way of using each sample for the claim it can actually support.
Write the analysis plan before collecting data
State the tests, the thresholds and the handling of missing data in advance. This is standard practice in registered studies and it is just as useful in commercial work. It protects you from the temptation to search the dataset until something reaches significance, and it makes the writing-up stage considerably faster.
More Reading
How to scope a feasibility study before you spend anything
Most feasibility work fails at the scoping stage, not the analysis stage. Here is the sequence we use to define a question that evidence can actually answer.
18 August 20266 min read
ReadWhat makes a business plan fundable rather than merely complete
Lenders and investors read plans in a particular order and look for a particular kind of honesty. Here is what they check first.
8 July 20267 min read
ReadHow to read a data analysis you did not run yourself
You do not need to be a statistician to interview a results section. Five questions will surface most of the problems.
16 June 20264 min read
Read