4000 math 5day

You are going to analyse the data from the 1988 National Survey of Fertility and Contraceptive Prevalence in China. The factors which are commonly found to be associated with contraceptive use in China are previous use, previous failure, age, residence, education, occupation, ethnic, region, motivation, period (after 1984, or before 1984). You are going to explore how these factors affect fertility behaviour in China.

Since 1979, China has changed very much, to assume the impacts of the above factors being constants doesn’t seem appropriate, for example, the impact of region in 1979 would be much larger than 1988 as the difference between regions had been reduced since 1979. Also to explore how the impacts change with the time itself is of interest as it may reveal how the society and policy change. Varying coefficient models are a useful statistical tool to explore this kind of change, the estimation methodology proposed by Fan and Zhang (1999) may be used in this project.

In addition to varying coefficient models, this project also involves censored data analysis. Typically, there are two approaches to analyse censored data, one is hazard function based approach, for this approach, Cox proportional hazard model, see Cox, 1972, together with partial likelihood estimation procedure would be a very important tool, which may be used in this project. Another one is unbiased transformation based regression approach. From modelling point of view, regression approach may be more flexible than Cox model. You may use either the Cox model or unbiased transformation based regression approach in this project.

This project would involve either R or SAS programming.




Use the order calculator below and get started! Contact our live support team for any assistance or inquiry.