Busy Mum Finally Secures her PhD
After completing her master’s degree more than 40 years ago, Dr Gill Hendry took on studies for a PhD in Statistics at the prompting of fellow tennis player and UKZN academic, Professor Delia North.
North, who is Dean and Head of the School of Mathematics, Statistics and Computer Science at UKZN, interacted with Hendry at tennis events over a period spanning more than 10 years.
Spurred on by North, Hendry eventually decided to embark on the final phase of her academic journey.
Hendry originally graduated in the 1970s with a Bachelor of Science degree majoring in Applied Mathematics and Mathematical Statistics from the University of Witwatersrand (Wits). ‘I continued with my Honours degree in Operations Research and was extremely fortunate to study under Paul Fatti, one of South Africa’s foremost statisticians,’ said Hendry.
‘After graduating, I joined the lecturing staff in the Department of Applied Mathematics, Computer Science and Mathematical Statistics at Wits and, while lecturing, completed my masters under Paul Fatti and Michael Sears – now a crime writer!’ said Hendry.
With marriage to Keith Hendry and the arrival of two children, Neil and Liesl, Hendry continued teaching Mathematics at secondary level, and later lectured on data analysis for postgraduate students at the Durban University of Technology and UKZN. ‘Once again, I was able to explore the pleasures of statistics,’ she said. ‘I pursued my doctoral studies only after my two children graduated from university.’
Hendry’s doctoral thesis was titled: “The Management of Missing Categorical Data: Comparison of Multiple Imputation and Subset Correspondences Analysis”.
In 2004, Hendry had investigated the relationships in a set of asthma severity data gathered specifically for a study on the effects of air pollution on the respiratory health of children in the South Durban basin.
Hendry soon realised that a challenge to this data set was the missingness present. (In Statistics, missing data, or missing values, occur when no data value is stored for the variable in an observation. Missing data are a common occurrence and can have a significant effect on the conclusions that can be drawn from the data.) Hendry decided to focus her attention on ways to analyse categorical data that suffers from missingness. Two methods (Multiple Imputation and the Subset Correspondence Analysis) were studied and their methodologies and results compared.
Multiple imputation is a relatively modern method for handling missing data. ‘The practical challenge in the application of multiple imputation, that was previously undocumented, was the identification of interactions needed for the imputation model. On the one hand, the data was needed to identify relevant interactions; on the other hand, the interactions are needed to impute the data. This dilemma was explored and a possible solution presented.’
Subset correspondence analysis is also a relatively new method. Dr Hendry stated: ‘Although applications to subsets of data have been published, its use on data with missingness was not well documented. Apart from applying this method to the asthma data, I showed how interactions could be included in an analysis with subset correspondence analysis. I further examined the effect that different missingness mechanisms have on subset correspondence analysis.’
Hendry’s study also identified the relationships between asthma severity and various environmental, behavioral, socio-economic and genetic factors.
Hendry plans to continue her research in the missing data field which she identified during her PhD studies. ‘I hope that I can use my knowledge from my work so that others can benefit from my experience in this field,’ she said.
‘While there were times of frustration, the excitement of achieving small steps in the process far outweighed the negatives. I was extremely lucky that I had the support of my family and friends and it was rather special to be doing postgraduate studies at the same time as both my children,’ said Dr Hendry.