Research Methods and Data Collection for Computing Projects

Alex: Today we're covering research methods and data collection for computing projects. Sam, what's the first thing a student needs to understand about research?

Sam: The distinction between primary and secondary research. Primary research involves collecting original data directly from sources, through surveys, interviews, observations, or experiments. Secondary research involves analysing and synthesising existing published work: academic papers, industry reports, technical documentation. Both are valuable, and most good projects use a combination.

Alex: For a computing project, when would you use primary research versus secondary?

Sam: Secondary research is essential for understanding the current state of knowledge in your area, the existing solutions to the problem you're addressing, and the theoretical frameworks that apply. You'd do secondary research first to understand what's already known before deciding what new data you need to collect. Primary research is then used to address the specific questions that the existing literature doesn't answer, or to apply established knowledge to your specific context.

Alex: What are the main primary research methods for computing projects?

Sam: Surveys are widely used because they're relatively efficient for gathering data from a moderate number of participants. You can use them to gather opinions, preferences, usage patterns, and requirements. Interviews allow you to explore topics in much greater depth with a smaller number of participants, uncovering insights and nuances that a survey wouldn't reveal. Observations involve watching users interact with a system or perform a task, which can reveal issues that neither survey responses nor interviews would capture.

Alex: What about experiments?

Sam: Experiments are particularly appropriate in computing when you're comparing different approaches or measuring the performance of a system. You might compare the query response times of two database designs, or measure the accuracy of two different classification algorithms on the same dataset. Experiments need careful design to ensure the results are valid and that any differences observed are genuinely attributable to the variable being tested.

Alex: You mentioned validity and reliability. Can you explain those?

Sam: Validity is about whether your method actually measures what you think it measures. If you're trying to measure user satisfaction with a system but your survey questions are ambiguous or leading, your results won't validly represent satisfaction. Reliability is about consistency: if you repeated the same research in similar conditions, would you get similar results? A reliable measurement method gives consistent results. Both are essential for trustworthy research.

Alex: What ethical considerations apply to research involving human participants?

Sam: Participants must give informed consent: they need to understand what they're participating in and agree to it freely. They should have the right to withdraw at any time without consequence. Their anonymity and confidentiality should be protected where possible. Data should be stored securely and used only for the purposes described to participants. In an academic context, research involving human participants usually needs ethical approval from the institution.

Alex: Great practical grounding. Thanks Sam. Next we look at exploring business requirements and stakeholder needs.