Divorce demography is the study of the demographic factors that impact divorce as a social phenomenon.
Information about divorces can generally be obtained from the census data published by governmental statistics offices.
One measure of divorces is the crude divorce rate, which is the number of divorces per 1,000 population. It can give a general overview of marriage in an area, but it does not take people who cannot marry into account. For example, it would include young children who are clearly not of marriageable age in its sample. A related measure is the refined divorce rate which measures the number of divorces per 1,000 women married to men, so that non-married persons, e.g. young children are left out of the rate.
Another measure of divorces is the divorce to marriage ratio, which is the number of divorces to the number of marriages in a given year (the ratio of the crude divorce rate to the crude marriage rate). For example, if there are 550 divorces and 1,000 marriages in a given year in a given area, the ratio would be one divorce for every two marriages, e.g. a ratio of 0.55 (55%). However, this measurement compares two unlike populations, those who can marry and those who can divorce.
Say there exists a community with 100,000 married couples, and very few people capable of marriage, for reasons such as age. If 1,000 people obtain divorces and 1,000 people get married in the same year, the ratio is one divorce for every marriage, which may lead people to think that the community's relationships are extremely unstable, despite the number of married people not changing. This is also true in reverse: a community with very many people of marriageable age may have 10,000 marriages and 1,000 divorces, leading people to believe that it has very stable relationships.
Furthermore, these two rates are not directly comparable since the marriage rate only examines the current year, while the divorce rate examines the outcomes of marriages for many previous years. This does not equate to the proportion of marriages in a given single-year cohort that will ultimately end in divorce.
Another solution is to conduct a longitudinal study of married batches to assess the divorce rate.