Social statistics postgraduate personal statement example

Personal statement example

My interest in social statistics grew out of a piece of work I nearly avoided. In my second year I chose a quantitative methods module mainly because it fitted my timetable around shifts, and I expected to find it dry. Instead I discovered that I liked the discipline of having to say exactly what I was measuring. By the end of the module I was the person in the seminar asking why we had dropped the cases with missing income data, and what that might do to our conclusions.

My dissertation took this further. Using an existing large-scale household survey, I looked at the relationship between housing tenure and self-reported wellbeing among adults under 35. Learning to handle a complex dataset, with weights and derived variables, taught me more than any reading list: I spent a fortnight simply working out how tenure had been recoded between waves. My regression models showed the expected association, but the interesting part was how much of it thinned once I added measures of financial strain. Writing that up honestly, rather than overstating what I had found, was the most useful thing I did as an undergraduate. It also showed me the limits of my training. I could run and interpret a linear model, but I could not properly handle the clustered structure of the data, and my understanding of measurement error was intuitive rather than technical.

Since graduating I have worked as an administrator at a community advice centre dealing with debt, benefits and housing enquiries. Part of my role is maintaining the case record system, and I volunteered to produce a quarterly summary for the management committee. This is modest work, but it has been an education in how social data is actually made. Enquiry types are recorded inconsistently by different volunteers, postcodes are sometimes missing, and one person's "housing" is another's "benefits". I rewrote the categories with the advice team and now produce simple tables and charts of enquiry volumes by type and month. When the committee wanted to claim a rise in housing problems, I had to explain that our recording had changed halfway through the year. Being the person who says "we cannot tell that from this" is not always popular, but it matters.

Alongside this I have been teaching myself R using free online materials and a borrowed library copy of an introductory text on statistical learning. My small ongoing project is with published neighbourhood-level statistics for my local authority: I have been mapping indicators of deprivation against the location of our advice centre's users, partly to practise joining datasets and partly out of genuine curiosity about who does not reach us. It is slow, and I have restarted the code twice, but I now understand why data cleaning takes up most of the time in real analysis.

I want formal training in survey design, sampling, multilevel modelling and causal inference, because I keep meeting questions I can pose but not answer. I am particularly interested in longitudinal data, and in how far apparent effects of housing circumstances reflect selection into those circumstances. I am aware that postgraduate study will stretch my mathematics, and I have been working through calculus and matrix algebra exercises to prepare rather than hoping to catch up later.

In the longer term I would like to work in research for a charity, local government or a research institute, producing analysis that is careful about its own uncertainty. I have seen how easily a number becomes a claim, and I would rather be trained properly than continue improvising.

Review this personal statement

Latest reviews