Sensitivity of Hyperparameters to Prior distribution: Exploratory study of Discontinuity of Family planning in Uganda

Objective: To investigate the sensitivity of hyperparameters to prior distribution.

Methodology: Three choices of hyperparameters were used to investigate the sensitivity of the prior distribution to the inverse Gamma distribution using Uganda Demographic Health Survey data on Family Planning. The Bayesian framework based on Markov Chain Monte Carlo (MCMC) simulation techniques from full conditional of nonlinear, linear and random effects were used for estimation of the unknown posterior distribution.

Findings: The three choices of hyperparameters are less sensitive to variations. Similar results were obtained in the different choices except in private place of delivery where two variations in hyperparameters were insignificant while one was significant at 95% CI.