Sensitivity analysis with simulated data errors : synthetic extinct generations method

• Main Author: Ruzvidzo, Martin
• Other Authors: Dorrington, Rob
• Format: Dissertation
• Language: English
• Published: University of Cape Town 2014
• Subjects: Demography
• Online Access: http://hdl.handle.net/11427/5889

Includes abstract. === Includes bibliographical references (leaves 59-61). === This study develops the key components of the Synthetic Extinct Generations (SEG+delta) method in trying to answer research questions raised by Dorrington and Timeaus (2008) and in the process investigates the different sets of combinations of the key components of the SEG+delta method when applied to the 23 error scenarios used by Hill and Choi (2004). In addition, the study determines the pattern of estimates of 4sfi15 per set of combinations, the combination that results in the best estimate of 45q15 per scenario and per combination of scenarios and the best combination that result in best estimate of 4sfi15 across all 23 error scenarios. The current study assesses the errors in age reported in censuses by comparing the weighted average of the ages of seven countries in the sub-Saharan African region to the age error pattern used by Hill and Choi (2004). These findings suggest that there is no significant difference (except of the zig-zag pattern in age errors at older ages in the Hill and Choi scenario) in general pattern of age errors of the sub-Saharan African region and the pattern used by Hill and Choi (2004).