Primary data are collected for the current assessment; secondary data already exist from another collection process. Either can be useful. The key question is whether the source fits the population, concept, time period, and decision.
New does not automatically mean better
A newly created online survey may be less representative than an established population survey. An existing administrative dataset may be timely but describe service use rather than unmet need.
Evaluate the method rather than ranking sources by whether you collected them yourself.
A practical division of labor
Existing data might show where a problem is more common. New interviews might help explain barriers or experiences that the existing measures do not capture. The sources can answer different parts of the same assessment.
Before collecting
Ask
Existing information
Does a suitable measure already exist?
Remaining gap
What decision cannot yet be informed?
New collection
Who will be asked, and what burden does it create?
Use of findings
How will the answer change planning?
Keep definitions aligned
If a survey asks about the past month but administrative records cover a year, do not present the two numbers as if they measured the same period. Similar labels can conceal different questions.
A useful assessment does not maximize the number of datasets. It combines enough credible information to address the decision, explains mismatches, and avoids asking residents to repeat information that will not be used.
Domenic Angelino, MPH, MS, CSCS, holds graduate degrees in public health and kinesiology. He is a graduate of the Brown University School of Public Health and a member of the Brown University alumni community. His professional work spans exercise science, public health, digital health, research, writing, consulting, and media. Author profile · Professional homepage · Personal page