GlossaryStatistics and interpretation
Year-over-year comparison
A year-over-year comparison uses the corresponding prior-year period. Learn how matching dates, definitions, and coverage affect the result.
A year-over-year comparison compares a measure with the corresponding period in the previous year. It can compare one month, one quarter, a full year, or matching year-to-date periods. The dates, definitions, coverage, and counting rules need to be compatible for the difference to describe a meaningful change.
Use corresponding periods
Comparing this July with last July is different from comparing July with June. The former can help reduce the influence of recurring calendar patterns, although it does not remove every seasonal or calendar effect.
The Bureau of Labor Statistics distinguishes over-the-year changes from changes between consecutive months and from changes in annual averages. Those are separate comparisons even when they use the same underlying series. BLS percent-change guidance.
Example
Hypothetical example: A service records 450 eligible calls from January 1 through March 31 this year and 400 during the same dates last year. The year-over-year increase is 50 calls, or 12.5%.
Comparing the 450 calls with the previous full year's 1,700 calls would not be a matching-period comparison. It would compare three months with twelve. Similarly, a current partial March should not silently be compared with all of last March.
Common confusion
Year-over-year does not mean “since January.” A July year-over-year figure might compare July alone with the previous July. A year-to-date year-over-year figure would compare January through the stated July cutoff with the matching prior-year span.
The term also does not establish a trend. One pair of periods can be influenced by an unusual event, delayed reporting, or an exceptionally low starting value. Look at a longer series before describing a sustained direction.
Check what changed besides the number
Use the same date basis, such as occurrence date or report date. Confirm that both periods include the same agencies, geographic area, categories, and record statuses. A revised historical total should be identified rather than compared with an older unrevised extract without explanation.
Calendar alignment can matter. Leap years, numbers of weekdays, holidays, and shifting event dates can affect the comparison even when month names match. Do not assume that dividing by the same population resolves these differences.
A clear label states the exact periods and whether counts are provisional. Explain known breaks in collection or classification alongside the figures, so the reader can distinguish a measured difference from a change in the reporting system.