Detection of Digit Preference and Age Misreporting by Using Demographic Techniques

Abstract

Age misreporting and digit preferencel is very common error all over the world. In developing countries this error is more critical than the developed countries (Talib et aI.,2003). The age misreporting and digit preference can be found frequently in population census surveyor other epidemiological and clinical studies. There are many studies in the area of age and sex data. However, this research is unique in the sense as no previous study attempted to evaluate simultaneously the age and sex data of medical outpatients, particularly of Chittagong Medical College (CMC) in Bangladesh and that of the population of Health and Demographic Surveillance System (HDSS) in Matlab. The rationale for using two data sets of the population in this research is to compare the quality of age collected by two sources- one with a long history of proper data record keeping and verification and the other has no such history. There is no proper record of keeping and verification system of age and sex in outpatients of CMC Hospital. However, in Matlab HDSS there is a systematic data management system from its establishment in 1966. Since then the Matlab HDSS has maintained the registration of births, deaths, and migrations, in addition to carrying out periodical censuses. A group of trained staff were involved in data collection and periodic vi sit of all households in the area. This data is then maintained continuously in a database system . It is expected that age and sex data collected by Matlab HDSS would be of better quality than that of the data collected by CMC Hospital as the former collects these data more rigorously and meticulously than the later. This hypothesis is also supported by data. We know that demographic data are usually classified by age and sex, and errors in age reporting are more frequent than in sex reporting. Different demographic techniques were used to evaluate both the data sets. These techniques include Whipple's Index (WI) that indicates age data which show systematic heaping for particular ages such as those ending with ' 0' and ' 5', Myer's Blended index (MI) that measures the extend of digit preference for all the digits, and Age-sex Accuracy Index that determines the accuracy of age reporting. I