| Abstract: | Becker (1981) presents some theory about related measures of fertility. He SUMMARY compares his theoretical predictions with observed relationships found in a set of data collected in Bangladesh. In general, he finds good agreement. In this paper, we reanalyse the data using Bayesian methods. In particular, we use Gibbs sampling to fit trigonometric regression models with autocorrelated errors. The results are generally in agreement with Becker's. However, evidence from one of the autocorrelation parameters and a residual analysis casts some doubt on whether the basic cosine model which is assumed fits the data well. |
