DEMOGRAPHV@ Volume 13, Number 4 November 1976 ABORTION AND FAMILY·BUILDING MODELS: FERTILITY LIMITATION IN HUNGARY Kathleen Ford Family Growth Survey Branch, Division of Vital Statistics, National Center for Health Statistics, 5600 Fisher's Lane, Rockville, Maryland 20782 A bstract-This paper demonstrates how family-building models can be used for checking the quality of abortion data. A family-building model, called FERMODA, is presented and used to check whether the available data on abortion for Hungary are consistent with Hungarian data on other aspects of family building. Model cohorts are generated, using in FERMODA the proportions of women married and the birth and abortion rates of two selected Hungarian cohorts. The estimates obtained from the simulated cohorts concerning use of contraception are compared with the information available on contraception, and, on the basis of this comparison, a judgment is made about the quality of abortion data for Hungary. An increasing number of countries have liberalized their abortion laws over the past two and a half decades, and data are becoming available in several of these countries on the number and character- istics of women obtaining abortions. Be- cause completeness of reporting varies widely, the accuracy of abortion data is often doubtful. This paper demonstrates how family-building models can be used for checking the quality of such data. A family-building model, called FER- MODA, is presented, encompassing pro- portions of women married, proportions using contraception, birth rates, and abor- tion rates. This model is used to check whether the available data on abortion for Hungary are consistent with Hungarian data on other aspects of family building. Model cohorts are generated, using in FERMODA the proportions of women married and the birth and abortion rates of two selected Hungarian cohorts. The estimates concerning use of contraceptives obtained from the simulated cohorts are compared with the information available on contraceptive practice for Hungary, and, on the basis of this comparison, a judgment is made about the quality of abortion data for Hungary. THE MODEL Family-building models are concerned with the series of conceptions and out- comes of pregnancy that constitute the reproductive history of a woman. Models of varying complexity have been con- structed, but four major variables are common to all: (a) fecundability, i.e., the probability that a fecund woman will con- ceive in a given month; (b) outcome of pregnancy, including live birth, stillbirth, miscarriage, and induced abortion; (c) length of gestation period; and (d) the period of anovulation following a termi- nation of pregnancy (Potter and Sakoda, 1966, p. 451). The models constructed using these four variables fall into two groups, sto- chastic and deterministic. Stochastic mod- els are those in which a random variation in the number of events is included. The occurrence and timing of an event are de- termined by probability distributions and a randomizing mechanism. Ridley and Sheps (1966), Holmberg (1970, 1972), Jac- quard (1967), and others have constructed models of this type. These models are quite flexible, although their use requires a relatively large number of calculations. 495