Quantitative research turns a question into measurements that can be analysed numerically. It can describe how long tasks take, estimate the proportion of people choosing an option, or compare outcomes across conditions. Its value depends on what the numbers represent and how they were produced.
It is not defined by having a large sample, and a large sample does not automatically make a finding reliable or representative. The design must support the conclusion being drawn.
Define the quantity before collecting it
The Australian Bureau of Statistics’ guide to data types distinguishes numerical quantities from categorical descriptions. In research practice, categories can also be counted and analysed quantitatively, provided their meaning is clear.
An illustrative study of successful setup needs a definition of success. Reaching the final screen, configuring the account correctly, and being able to complete a first task are different outcomes. Decide which one serves the question and how it will be observed.
Plan the analysis alongside the measure. Time-on-task data may be skewed, repeated observations from one person are not independent participants, and missing responses may differ systematically from completed ones. These details affect what a summary or comparison means.
Distinguish describing the sample from estimating beyond it
Descriptive statistics summarise the observed data. Inference uses a design and assumptions to reason about a wider population or process, often expressing uncertainty through a confidence interval or other appropriate measure.
Recruitment matters to that step. A voluntary survey of enthusiastic customers can produce precise-looking percentages while missing less engaged users. More responses from the same selection process do not necessarily resolve sampling bias.
Choose sample size according to the desired precision, effect size, variability, and design, rather than a universal minimum. A study intended to detect a small difference may require substantially more information than one intended to describe a broad pattern.
Interpret the result in the context of the decision
A statistically significant difference may be too small to matter practically, while an uncertain estimate may still identify a question worth investigating. Report the size and uncertainty of the result rather than reducing it to whether a threshold was crossed.
Qualitative research can help explain how people interpreted a task or why an unexpected pattern deserves attention. The methods overlap in what they can investigate; quantitative research can explore, and qualitative work can evaluate. Choose the combination that gives the decision adequate evidence, with the limitations of each contribution kept visible.