The calculation uses the total population
This rate divides marriages by everyone in the population. The group includes children, married adults, and people who are not currently likely to marry for the first time. CDC calls this full-population measure a crude rate.
The rate helps track national marriage events over time. It is not a 0.61% annual chance that a single adult will marry.
The rate has trended lower since 2000
The series begins at 8.2 per 1,000 in 2000, moves mostly downward, reaches 5.1 in the disrupted 2020 year, and rebounds to 6.2 in 2022 and 6.1 in 2023.
A long-term comparison should not use 2020 alone as the baseline because marriage ceremonies were unusually affected by pandemic restrictions and postponements.
Provisional means the number may change
State vital-record systems supply counts to the National Vital Statistics System. Provisional releases prioritize timeliness and can be revised when reporting is completed or corrected.
The page dates the figure and links the original two-page table so readers can inspect notes and later revisions.
The formula is events divided by population
An overall marriage rate takes the number of marriage events recorded during the year, divides by the resident population, and multiplies by 1,000. This is also called a crude rate because the population is not narrowed by age, marital status, or other characteristics. The calculation puts the event count on the same population scale, which makes 2000 and 2023 more comparable than their raw counts alone.
The word crude does not imply poor data. It tells readers that the counted group is broad and that changes in the population mix can affect comparisons. An age-specific or eligibility-based rate would answer a narrower question and require different data.
6.1 per 1,000 is not 6.1 percent
Per-thousand and percent units are easy to confuse. A rate of 6.1 per 1,000 is numerically equal to 0.61 per 100, or 0.61%, when the same total population is used in the calculation. Saying 6.1% would inflate the rate tenfold. The original unit is preferable because it matches the CDC table and avoids a false sense of individual probability.
Even 0.61% should not be framed as the annual chance a single person marries. The counted group contains two people for each marriage event as well as children and already-married adults, and the event count includes ceremonies rather than unique eligible individuals. A person-level probability needs a defined eligible group and follow-up design.
The 2000-to-2023 comparison is descriptive
The rate was 8.2 per 1,000 in 2000 and 6.1 in 2023. That is a decline of 2.1 marriage events per 1,000 residents in the overall series. The endpoints show direction and size under one published measure; they do not by themselves show why the change occurred.
Later first marriage, different age structure, cohabitation, economic conditions, remarriage patterns, and changing relationship choices are possible parts of the broader story. Each explanation needs separate evidence. The rate page should remain a guide to the population used in the time series rather than presenting a guess about cause as if it came from vital records.
Why the 2020 trough needs context
The 5.1 rate in 2020 reflects an unusual year in which ceremonies and administrative operations were disrupted. It is the low point in the cited 2000–2023 table, but it is not a clean picture of underlying long-term intent. Some couples postponed events, changed locations, or altered plans because the normal ceremony process was unavailable.
The rise to 6.0 in 2021 and 6.2 in 2022 is properly described as a rebound in recorded events. Calling it a new marriage boom would require comparison with the earlier trend, age structure, and any backlog of delayed ceremonies. The 6.1 result in both 2019 and 2023 is useful context, though equal overall rates do not mean every underlying pattern was identical.
The marriage count comes from official records
Marriage records are compiled through state and local registration systems and summarized nationally by the CDC's National Center for Health Statistics. This makes the series an official event count rather than a survey asking people whether they married. The records are useful, while notes about reporting completeness and early status remain part of the evidence.
The count reflects ceremonies recorded in the system for the year. It does not identify how couples met, whether it was a first marriage, how long the relationship existed, or whether the couple later divorced. Those questions require linked records, surveys, or longitudinal research beyond the rate table.
The population can change even when behavior does not
Because the counted group is the total population, a change in the number or share of children, older married adults, or age groups most likely to marry can move the overall rate. Imagine two populations with identical age-specific marriage behavior but different age structures: their overall rates can differ simply because one contains more adults in ages where marriage events are common.
That does not make the overall rate useless. It makes it a population-level volume measure. Researchers explaining behavior may calculate age-specific or standardized rates, while journalists describing national event volume can use the overall series as long as the total population used in the calculation stays explicit.
National rates and personal odds are different products
A personal marriage probability would depend on age, current relationship status, prior marriage, location, partnership goals, and a defined period of observation. The national overall rate ignores those details by design. It cannot tell a reader how likely a current relationship is to lead to marriage or how long that process might take.
The same caution applies to dating-pool estimates. Population frequency can describe how common measurable traits are, but it cannot be multiplied by 6.1 per 1,000 to create a marriage forecast. The data sets, counted groups, and events do not align. Use each measure for the question it was built to answer.
State marriage rates require an extra location warning
CDC assigns marriages to the state where the ceremony occurred. In places with destination weddings, the count can include many nonresident couples while the population base is the state's resident population. That can make a state rate look high without showing that its residents marry more often.
The national rate is less exposed to interstate ceremony travel because a wedding moved from one state to another remains inside the U.S. total. International ceremonies, reporting coverage, and other system limits can still matter. Readers comparing states should use the dedicated state methodology and avoid treating the ranking as a measure of relationship culture.
How to quote the rate accurately
The safest short form is: the provisional U.S. marriage rate was 6.1 marriage events per 1,000 total population in 2023, based on 2,041,926 recorded marriages. This sentence includes the status, location, event, unit, counted population, year, and supporting count without implying individual odds.
When space allows, add that the rate was 8.2 in 2000, fell to 5.1 in the disrupted 2020 year, and returned to 6.1 by 2023. Link the CDC table and preserve its notes. Do not call the rate a share of singles, a percent of couples, or evidence that any particular relationship is more or less likely to succeed.
Age-based rates answer a different question
If the goal is to compare marriage behavior among people in their twenties, thirties, or forties, an age-specific rate is more appropriate than the overall national rate. It restricts events and population to the relevant age group, reducing distortion from changes in the number of children or older adults. A standardized rate can go further by applying age-specific rates to a common population structure.
Those alternatives require more detailed event counts and population data than the two-page CDC trend table supplies. They should be labeled as separate analyses, not silently substituted into the 6.1 headline. The overall rate remains the correct answer to the narrow question about the published national series; an explanation of behavior may need a richer measure.
A rate can stay flat while the underlying population changes
The same overall rate in 2019 and 2023 does not mean the same people married at the same ages or that the number of first marriages and remarriages was unchanged. Different combinations of population size, age distribution, and event patterns can yield the same per-1,000 result. A single summary rate compresses all of those paths into one value.
That is why the evidence table should be read as a trend indicator rather than a complete model of U.S. relationships. When a story depends on who is marrying, retrieve age, prior-marital-status, location, and group tables designed for that question. The national rate supplies context but cannot replace analysis of smaller groups.
A reproducible update uses one release from end to end
When CDC publishes a newer year, capture the exact release URL, status, marriage count, rate, notes, and access date before editing the headline. Update the evidence table and short answer together, then check every internal page that repeats the national figure. A mismatched new count and old rate would be more misleading than leaving the prior labeled year in place.
After publication, verify the visible answer, structured data, and citation link on the live page. Search snippets can lag behind the page, so a stale result is not proof the site failed to update. The on-page source receipt should remain the authoritative record until Google recrawls and refreshes its copy.
What full group does each percentage or rate use?
| Year | Rate per 1,000 | How to read it |
|---|---|---|
| 2000 | 8.2 | Start of the cited national series |
| 2019 | 6.1 | Last pre-pandemic year |
| 2020 | 5.1 | Ceremony and reporting disruption |
| 2021 | 6.0 | Initial rebound |
| 2022 | 6.2 | Highest post-2020 point in the cited table |
| 2023 | 6.1 | Provisional; 2,041,926 recorded marriages |
What does “denominator” mean? The denominator is the whole group used to calculate a percentage or rate.
How to read it: The overall rate uses the total resident population. This full-population rate is also called a crude rate. It shows event volume, not the annual probability that an unmarried adult will marry.
Download this table (CSV) ↓