Abstract: Traditionally studies on fertility desire have a hypothesis that people have only one optimal fertility preference,ignoring the existence of double-peak fertility preference that could lead to bias on estimation and projection of fertility preference.This paper theoretically demonstrates the occurrence of double-peak preference and the reasons for bias in fertility prefernece survey under the assumption of a single peak preference.A Probit model is uesed to analyze a survey results,testing the existence of double-peak fertility preference.Combining theory and statistical results,this paper gives a new analysis of the issues,such as "DINK" family,gender differences in fertility preference,and differences of fertility preference between only and non-only children.Finally this paper proposes a new explanation of the process of fertility decline in the double-peak preference perspective.

Keywords: Fertility Preference, Double-peak Preference, Probit model

摘要： 传统的生育意愿调查研究都以单峰偏好即人们只有一个最优生育偏好为假设,忽视了生育意愿存在双峰偏好的可能,因而有可能对生育意愿做出错误的估计和预测。 本文从理论上论证了生育意愿出现双峰偏好的原因以及在单峰偏好假设下生育意愿调查的偏误;利用Probit模型对问卷调查结果进行了分析,检验了生育意愿双峰偏好的存在性;结合理论与统计结果对"丁克"家庭、生育意愿的性别差异、独生子女与非独生子女间的生育意愿差别等问题作了新的分析,得到了一些有意义的结果;并从生育意愿双峰偏好的视角对生育率下降过程提出了一个新的可能性解释。

关键词: 生育意愿, 双峰偏好, Probit模型