Statistical modelling of demographic trends and changes in a London suburb during the 19th century

Crouch, Susan A. L. (2007) Statistical modelling of demographic trends and changes in a London suburb during the 19th century. (MSc(R) thesis), Kingston University, .

Abstract

This thesis examines closely the baptism, burial, marriage and census data of Kingston upon Thames, provided by the Centre for Local History Studies within Kingston University. This thesis discusses demographic trends and changes in Kingston over the nineteenth century. Generally, the population of Kingston grew rapidly over this period, with the largest increase occurring between the ages of ten and nineteen. This growth can be explained partly by better record keeping, improvements in health and hygiene and migration into the area. Baptism records showed continual increases in baptisms and marriage records supported this. The burial records showed an increase in age of death, probably connected to better general living conditions and improved sanitation. After an initial investigation of the data, statistical modelling techniques are applied to the marriage records data. Firstly, multiple linear regression was performed to analyse the factors affecting age at marriage. The results showed that a bridegroom's previous marital status, his social class, his bride's father's social class and the church were significant factors in predicting his age at marriage. Analysis also showed that a bride's previous marital status, her social class and her groom's father's social class were important factors in predicting her age at marriage. Secondly a conditional probability distribution has been produced by the method of Maximum Entropy to predict a bride or groom's age at marriage given prior knowledge of certain factors, such as social class; the prediction so obtained is compared with actual ages at marriage. Comparison is made between the results of the two statistical methods and finds that the previous marital status of the party whose age is of interest is found to be the most significant factor by both methods.

Actions (Repository Editors)

Item Control Page Item Control Page