GlossaryStatistics and interpretation
Baseline
A baseline is a clearly stated reference value or period for measuring change. Learn how its selection affects comparisons and differs from a goal.
A baseline is a specified reference value or period used to judge change. It establishes what later observations are compared with. A baseline can be a single year's total, an average across several periods, or another documented reference; its meaning depends on how it was chosen and calculated.
Name the reference explicitly
A baseline is useful only when readers can identify it. “Above baseline” leaves important questions unanswered unless the report states the dates, measure, geography, and calculation. A multiyear average is different from the lowest year or the year immediately before a program began.
CDC's healthcare-associated infection reporting provides a specialized example: historical baseline data support expected values in its comparison methods. Those clinical methods are not public-safety formulas, but they illustrate that a baseline is defined for a particular measurement system. CDC progress-report FAQs.
Example
Hypothetical example: An agency selects the average annual call count from three complete years as its reference. The totals are 900, 1,000, and 1,100, producing a baseline of 1,000 calls. A later year has 1,200 calls, or 20% above that reference.
Compared only with the immediately preceding year's 1,100 calls, the increase is about 9.1%. Both calculations can be correct. They answer different questions because they use different starting points.
Common confusion
A baseline is not necessarily a goal. The historical workload level does not establish the workload an agency should have. A benchmark may describe an external comparison or desired standard, while a baseline records the chosen point of reference.
It is also not automatically an estimate of what would have happened without a policy change. A before-and-after difference can reflect many influences. A baseline alone cannot isolate a program's effect or establish causation.
Choosing and maintaining a reference
Explain why the chosen period fits the question and whether unusual events affect it. Selecting an unusually quiet year can make later increases look larger; selecting an unusually busy year can make them look smaller. Show enough surrounding history for readers to assess the choice.
Keep definitions and coverage compatible. If boundaries, categories, or reporting participation change, the baseline may need recalculation or a clear break. Distinguish a fixed baseline from one that moves each year. Changing the reference silently can change the reported comparison even when the current observation remains the same.