A New Method for the Detection of Multipartner Fertility in Social Surveys: The Age Reporting Discontinuity (ARD)

Bryan L. Sykes, University of California, Irvine

Demographers have developed a number of methods to estimate cohort and period fertility. With changes in rates of nuptiality, increased marital dissolution, and growth in cohabitation and recoupling, new methods are required to estimate multiple-partner fertility. Recently a new, exploratory method for the study of reproduction with multiple partners was proposed: The Age Reporting Discontinuity (ARD). Yet, the formal model and its sensitivities were not specified. Using inmate data on the number, ages, and reporting order of children by incarcerated fathers, I show that ARD has an overall accuracy rate of 91.8% when benchmarked to data on ever incarcerated men in the 2002 National Survey of Family Growth (NSFG) Male Supplement. This method is particularly useful for analyses of father absence and family disruption, but general usage may allow for calculating historical estimates of multiple-partner fertility in social surveys that do not inquire about reproduction with several partners.