Biostatistics postgraduate personal statement example

Personal statement example

My interest in biostatistics began with a rather unglamorous piece of coursework: cleaning a messy dataset of patient follow-up times in which almost a third of the observations were censored. I had expected the statistics to be the difficult part, but the harder question was what the missing information meant. Had those patients recovered, moved away, or simply stopped attending? That ambiguity, and the fact that the choice of assumption changed the conclusion, convinced me that medical data demands more care than any other kind I had handled.

I took that interest into my final-year project, where I compared parametric and semi-parametric approaches to survival analysis using an openly available trial dataset. Fitting Cox proportional hazards models was straightforward; checking whether the proportionality assumption actually held was not. Plotting Schoenfeld residuals showed a clear time trend for one covariate, and I spent several weeks reading about time-varying coefficients and stratified models before settling on an approach I could justify. My supervisor pushed me to write the limitations section as carefully as the results, which I now think was the most useful instruction I received in three years. Alongside this I built up my R skills, moving from single scripts to organised, commented code with reproducible outputs, because I had learned the hard way how difficult it is to retrace an analysis six weeks later.

Over one term I volunteered two afternoons a week with a clinical audit team at a hospital trust, mostly taking notes in meetings, tidying spreadsheets of extracted case-note data and checking entries against agreed definitions. I had no analytical responsibility and made no clinical judgements, but observing the process was valuable in ways my degree could not replicate. I saw how much effort went into agreeing what counted as an eligible case before anyone calculated anything, and how readily a small inconsistency in coding could distort a proportion. I also watched a statistician explain a confidence interval to a group of clinicians without once using the word "significant", which struck me as a skill worth learning deliberately.

My part-time work as a data assistant for a housing association has given me ordinary but relevant discipline. I produce monthly summaries of repair response times and tenancy turnover, which means dealing with duplicated records, inconsistent dates and colleagues who need a clear answer rather than a caveat-laden paragraph. Writing short, honest summaries for non-specialists has improved my communication more than any presentation module. Coaching a junior netball team on Saturdays has similarly taught me to explain the same idea three different ways until it lands.

My reading has been steady rather than extensive. Rothman's Epidemiology: An Introduction helped me understand confounding as a substantive problem rather than a regression technicality, and I have been working through applied material on multiple imputation, since missing data now seems to me the central practical difficulty in health research. I am keen to study clinical trial design formally, particularly randomisation, sample size calculation and the handling of interim analyses, and to strengthen my understanding of Bayesian methods and hierarchical models for clustered data.

In the longer term I would like to work as a statistician within a clinical trials unit or a public health team, contributing to studies rather than only analysing them afterwards. Postgraduate study in biostatistics is the necessary step: I have the mathematical foundation and the habits of careful, documented work, and I want the specialist training to apply them where the conclusions matter to patients.