For women going on to have their babies, relatively few considered abortion as an alternative. Among those women who did not use contraceptive techniques, 91-1% indicated they did not consider having an abortion; among those women who did use contraceptives, 86-4% indicated they did not consider having an abortion.

As reported previously by Smith et al. (1971) and Steinhoff, Smith & Diamond (1972), the reasons most often given by a woman for deciding to have an abortion were: 'I am not married', ‘I cannot afford the child financially', 'I already have enough children', 'A child would interfere with my education', and 'I am not able to cope with a child at this time'.

The reasons vary with contraceptive usage and this is related to age and marital status. For example, older or married women were more likely to have been using contraceptive techniques and indicated they chose abortion because they could not cope with a child or considered they had enough children already.

The most frequently given reasons for carrying the pregnancy to term were, as might be expected, related to whether the baby was planned and whether contraceptive techniques were used.

Quite often when the woman found she had an unplanned pregnancy, she and her husband decided to have the child anyway. Disapproval of abortion was given as a reason for having the baby by about 13% of the women with unplanned pregnancies who had not used contraception. Similarly, of women who had used contraception, 10% said they did not approve of abortion so were having the baby.

The decision of a pregnant woman to have an abortion or carry to term is often thought to be related to marital status, age and religion. We have considered these factors along with contraceptive use (Text-fig. 3).

Marital status was seen to relate with contraceptive use to predict the decision regarding pregnancy termination. Non-married contraceptors were more likely to have an abortion than were married contraceptors; non-married women, in general, were more likely to have an abortion than were married women. This is in sharp contrast to national findings under illegal conditions in the late 1950s (Gebhard et al., 1958) but is in keeping with recent findings in New York and California (Berkov & Sklar, 1972; Duffy, 1971; Pakter & Nelson, 1971; Russell, 1969; Tietze & Lewit, 1972). It must nevertheless be emphasized that a large percentage of women, regardless of their marital status, chose to carry the pregnancy to term (Text-fig. 3).

This relationship of marriage to choice of abortion was, as might be expected, related to age. The older married woman more often chose an abortion than did the younger. While in absolute numbers more women under 25 years had abortions, a much greater proportion of pregnant women 25 years and older chose abortion than did younger women (Text-fig. 4).

Religion, too, was related to how a woman terminated her pregnancy. Regardless of marital status, Catholics consistently chose abortion less frequently than did Protestants or those women who professed no religion (Text-fig. 5). Buddhists, too,

Text-fig. 3. Choice of abortion related to marital status at conception and birth

Text-fig. 4. Choice of abortion related to marital status and age at conception.

chose abortion less frequently than Protestants but this was due to a greater perbentage of their pregnancies being planned.

With marital status, we have related a simple multiple choice question as to how each woman viewed the prospect of having a child and her choice of pregnancy outcome. Predictably, almost all women who expressed 'I would be unhappy to have this baby' decided to have an abortion. Less predictable is the finding that a number of women chose abortion despite their saying 'I would be happy to have this baby'. Least predictable was the finding that the overwhelming majority of women

Text-fig. 5. Choice of abortion related to marital status at conception and religion.

who claimed 'I would be neither happy nor unhappy to have this baby' chose abortion. A woman's so-called neutral statement regarding her desire to have a baby when she is pregnant was thus generally seen not to be neutral.

Finally, we correlated some early steps in the decision-making process with the final pregnancy outcome. Of those women not using contraceptives who claimed 'Sex was unexpected', 25% had their pregnancies terminated; so, too, did 26% of those women who claimed 'I thought I was in a safe period'. These two groups of women who were non-contraceptors and who did not plan to become pregnant comprise 68% of those who chose abortion. Of those who indicated that their pregnancy was the result of a birth control failure, 31% chose abortion.

We have attempted to show, by using a fairly complete population of pregnant women (conception cohort), a decision-making process relating how sexual behaviour, birth control use and pregnancy outcome could be monitored under conditions where abortion is legal and access to the total population was possible. Starting with the expectancy component of coitus, we traced pregnancy planning, birth control use and pregnancy outcome. The analysis was basically sequential but it is obvious that the decision-making process cannot be fully considered as a linear progression of events. Coitus for the majority of women was anticipated and yet pregnancy was not generally planned for. Regrettably, two-thirds of the women who had not planned their pregnancies had not used a birth control method to prevent such consequences. Of these women, as of the one-third that did attempt to prevent conception, only one out of three women elected to have an abortion. Marital status, age and religion were seen to affect the decision as to whether or not to have an abortion, but we also found that a woman's acceptance of her sexuality and attendant use of contraception related to these factors and correlated with how she decided to terminate her pregnancy.

This use of conception cohorts and the analysis of the decision-making process is believed to provide a model that is applicable elsewhere. Different independent variables may be used as filters at any stage to provide analyses of the decision-making process for any group under study. For example, one could monitor the change in dependent variables such as coitus or birth control usage, while using as indepen

lent variables religion, socio-economic status or age. Other models for similar anaysis have been proposed (e.g. Friedman, 1972). The present model seems advanageous primarily because it provides a built-in comparative control group of women who planned their pregnancy and indeed chose to deliver. This type of analysis of choice deals with a selected population from among a larger one available to us (Diamond et al., 1972). We consider that this gives a better reflection of the actual parameters attendant on the decision to have an abortion or carry to term, particularly if used repeatedly so that controlled comparisons can be made over a period of time.

Acknowledgments

This work was supported mainly by the National Institutes of Health Grant NICHHD HD-05853 and in part by the Ford Foundation and the East-West Center Population Institute. We would also like to acknowledge the help of Mr Kay H. Hoke, Mrs Alice M. Beechert and Mr Mark Takaguchi.

This is Report No. 5 of the Hawaii Pregnancy, Birth Control and Abortion Study.

References

BERKOV, B. & SKLAR, J. (1972) The Impact of Legalized Abortion on Fertility in California. Preliminary Paper No. 1, Results of Current Research in Demography, International Population and Urban Research, University of California, Berkeley.

DIAMOND, M., PALMORE, J.A., SMITH, R.G. & STEINHOFF, P.G. (1972) Abortion by request in Hawaii: indications from and for legal reform. Family Planning Perspectives (in press). DUFFY, E.A. (1971) The Effect of Changes in the State Abortion Laws. US Government Printing Office, Washington.

FRIEDMAN, H.L. (1972) An Approach to Psychosocial Research in Fertility Behavior. Working Paper No. 3, presented to the International Conference on Contraception, Abortion, Sterilization, and Prostaglandins, Paris.

GEBHARD, P.H., POMEROY, W.B., MARTIN, C.E. & CHRISTENSON, C.V. (1958) Pregnancy, Birth, and Abortion. Hoeber-Harper, New York.

PAKTER, J. & NELSON, F. (1971) Abortion in New York City: the first nine months. Family Planning Perspectives, 3, 5.

RUSSELL, J. (1969) Therapeutic abortions in California. Am. J. Obstet. Gynec. 105, 757.

SMITH, R.G., STEINHOFF, P.G., DIAMOND, M. & BROWN, N. (1971) Abortion in Hawaii: the first 124 days. Am. J. publ. Hlth, 61, 530.

STEINHOFF, P.G., SMITH, R.G. & DIAMOND, M. (1972) The characteristics and motivations of women receiving abortions. Sociological Symposium (in press).

TIETZE, C. & LEWIT, S. (1972) Joint Program for the Study of Abortion (JPSA): early medical complications of legal abortions. Stud. Fam. Plann. 3, 97.