Age patterns of marriage

Abstract

In different populations there is a common curve describing first-marriage frequency (first marriages per woman) as a function of age for each cohort. To fit the variety of patterns of human nuptiality it suffices to choose the age that serves as origin for a standard curve of first-marriage frequency, and to choose appropriate horizontal and vertical scales for the curve. The prevalence of a standard form for first-marriage frequency implies that the proportion ever-married in any cohort also rises along a standard curve, subject to choice of origin (the earliest age of first marriage), vertical scale (the proportion ever-marrying by the end of life), and horizontal scale (the pace at which the proportion ever-married increases with age). A mathematical expression (a double exponential) is found to fit the risk offirst marriage (among those who ever marry), and some of the implications of uniform features of nuptiality in different populations are discussed.

The ideas summarized in this paper have emerged over a period of several years in conversations between the author and others at the Office of Population Research, especially Ivan Lakos and Etienne van de Walle. Professor Lakos, Chairman of the Department of Economics at Antioch College, devoted the academic year of 1968–69 to research and study in demography at Princeton. Among his projects was the estimation of age-and-marital-status distributions from incomplete data, and it was in connection with his work that the existence of a standard pattern of nuptiality was first noted. Dr. van de Walle's estimations of age-and-marital-status distributions in nineteenth-century France helped develop some of the techniques described below.

The ideas summarized in this paper have emerged over a period of several years in conversations between the author and others at the Office of Population Research, especially Ivan Lakos and Etienne van de Walle. Professor Lakos, Chairman of the Department of Economics at Antioch College, devoted the academic year of 1968–69 to research and study in demography at Princeton. Among his projects was the estimation of age-and-marital-status distributions from incomplete data, and it was in connection with his work that the existence of a standard pattern of nuptiality was first noted. Dr. van de Walle's estimations of age-and-marital-status distributions in nineteenth-century France helped develop some of the techniques described below.

Notes

The ideas summarized in this paper have emerged over a period of several years in conversations between the author and others at the Office of Population Research, especially Ivan Lakos and Etienne van de Walle. Professor Lakos, Chairman of the Department of Economics at Antioch College, devoted the academic year of 1968–69 to research and study in demography at Princeton. Among his projects was the estimation of age-and-marital-status distributions from incomplete data, and it was in connection with his work that the existence of a standard pattern of nuptiality was first noted. Dr. van de Walle's estimations of age-and-marital-status distributions in nineteenth-century France helped develop some of the techniques described below.