Definition
Primary and secondary data
Statistics • Data Classification
Statement
Data are primary when the investigator collects them first-hand for the enquiry in hand, and secondary when they are taken from a record somebody else compiled, whether a government office, a published report or an earlier study. The classification is relative to the user, not to the numbers: the same figures are primary to whoever collected them and secondary to everyone who later reads them off that record.
How to use it
- 1Ask who did the collecting and why: if the phrase is 'obtained from', 'taken from a report' or 'supplied by an office', the data are secondary for this user.
- 2Words like census, survey, interview or experiment conducted by the investigator point to primary data.
- 3Secondary data are cheaper and quicker but carry the definitions, units and errors of whoever compiled them, which must be checked before use.
Worked example
A researcher studying rainfall copies the monthly totals for the last ten years out of the Meteorological Department's yearbook. Are these primary or secondary data?
Step-by-step
- 1The Department recorded the rainfall for its own purposes.
Why: The researcher did no measuring.
- 2The researcher is reusing an existing record, so the data are secondary.
Answer
Secondary data
Watch out
'Secondary' is not a judgement on quality or on how recent the figures are, and it is not the discrete/continuous or qualitative/quantitative distinction, which classifies the variable rather than the route by which the data arrived.