Data Collection, Sources and Acquisition Methods
Primary versus secondary data
Primary data is collected first-hand for the specific purpose at hand - surveys, interviews, observations, sensors, experiments and system logs generated by the organisation. Secondary data already exists and was collected for another purpose - government open data, published reports, census statistics, social media archives and third-party datasets. Primary data gives control over method and relevance but is costly and slow; secondary data is cheap and fast but may be outdated, differently defined or collected with unknown rigour.
Methods and sampling
Candidates should recognise quantitative methods (surveys with closed questions, sensor readings, transaction logs) versus qualitative methods (interviews, focus groups, open-ended responses). Sampling matters: a census measures everyone; a random sample gives every member an equal chance; stratified sampling divides the population into groups and samples each; convenience sampling uses whoever is easiest to reach and is the most biased. Sample size affects reliability - small samples produce unstable conclusions.
Traps
A common trap is labelling data as primary or secondary based on who holds it rather than who collected it and why. Data purchased from a market-research firm is secondary to your organisation because you did not collect it for your purpose. Another trap is sampling bias: a survey distributed only through a social media page reaches a self-selecting group, not the general population. Also watch for response bias, where wording of questions nudges answers.
Worked example
A school wants to know how students feel about the canteen. Surveying only students who visit the canteen at lunch is convenience sampling and excludes those who avoid it - the results will overstate satisfaction. A better design draws a random sample from the full enrolment list and includes an option for students who never use the canteen.