PhD Student in Psychology Konstantinos Voudouris, worked with The Office for National Statistics (ONS) to survey the enduring prevalence of Covid-19 across different groups of the population. He found the ONS to be like an ‘academia outside of academia’, and gained valuable advanced computing skills and experience of working in a large team. Konstantinos reported back to us on his experience:
“The Office for National Statistics is the gold standard for statistical and scientific rigour, and for transmitting the output of this to a non-specialist audience. These are skills that are particularly important to Early Career Researchers. Statistical knowledge is particularly important today, in light of the Replication Crisis which has struck the biological sciences. This has been diagnosed as resulting in part from poor application of statistical methodology.
Further, the ability to translate scientific results into a format understandable by non-specialist audiences, government, and policymakers is particularly important given the recent distrust of ‘science’ and ‘experts’ in the media. All of this, married with the fact that I could work on COVID-19 data, perhaps the most important data generated by the UK for several decades, meant that this project was ideal for me.
My fellowship project focused on Enduring Prevalence of SARS-CoV-2 (COVID-19) in UK subregions during the second wave of the COVID-19 Pandemic. Enduring Prevalence is when the proportion of the population of a subregion testing positive for SARS-CoV-2 (the pathogen that causes COVID-19) is significantly higher than the proportion of the population in surrounding areas testing positive. This higher prevalence is enduring if it stays high for an extended period of time, or shows a slower decline relative to other areas. Studies using Test & Trace data have shown that Enduring Prevalence can be related to higher levels of deprivation, larger Asian/Asian-British populations, larger proportions of the population working in manufacturing, construction, food processing, and transport/logistics, higher population densities, and greater levels of deprivation. However, Test and Trace data is biased in that individuals must volunteer to get tested, so it has reduced ability to detect asymptomatic cases and cases in individuals who do not want to test for other reasons (such as fears about self-isolation and its impact).
The Office for National Statistics, in collaboration with Oxford University and scholars at Manchester University, run the COVID-19 Infection Survey (CIS), which randomly samples the UK population, providing a less biased sample. We aimed to see to what extent the results found with Test and Trace data could be replicated using CIS data. This project was successful.
The ONS environment was very friendly, and I was settled in within a few days. Most people have a PhD or at least a masters, so in terms of academic background, we were similar. However, in terms of extra-academic backgrounds, the ONS and the civil service is very diverse and inclusive, which was really refreshing in contrast to the often-homogenous Cambridge environment. I was provided with bespoke training in the R programming language, Generalised Additive Models, Integrated Nested Laplace Approximation, parallel computing, code benchmarking, Reproducible Analytical Pipelines, data visualization, Bayesian methods, and statistical interpretation.
For the ONS, the aims were to explore to what extent we could answer questions about enduring prevalence using CIS data. For me, the aim was to gain advanced statistical and computational knowledge for my PhD research, to understand how the ONS translates detailed scientific analysis into a format digestible by a non-specialist audience and how they inform policy decisions, and to gain an insight into work in the civil service and whether this could be a viable career option for me post-PhD.
The fellowship met all of these aims. The ONS were pleased with my work, and I received praise from all levels of management, including an official recognition for my work on subregional analysis. I also gained important knowledge and experience which will be invaluable to me, and I saw that the ONS is like an ‘academia outside of academia’, meaning it could be a viable career option for me in the future.
I produced a 50-page internal report on Enduring Prevalence using CIS data, laying the foundation for future work on the topic by the ONS and academic partners. I presented this to over 150 people from the ONS, government, and academia, and received praise for my work. I also contributed to the weekly bulletin subregional analysis, and received recognition for my work from the ONS and the UK Statistics Authority.
My knowledge of R and machine learning methods was used. However, my research was quite different from my project focus, so my academic knowledge was not used greatly. Beyond my academic knowledge I brought my work ethic. They did not expect me to be able to complete this project in only 3 months. I was hoping that they would integrate me into the workforce, so I could get a sense for working in a large team in a fast-paced environment in the civil service. They did do this, it was great.
The added value this fellowship brought me were primarily a better statistical and scientific intuition, advanced scientific computing skills, and presentation skills (I presented my results to over 150 people from academia, government, and the civil service). I also made great contacts in the ONS and found everyone to be really friendly.
The placement will make an important impact on my research, giving me the advanced skills to produce rigorous and robust research using modern statistical methods. This will also help in my future career either in academia, industry, or the civil service. I gained statistical and advanced computing expertise, presentation skills, experience of working in a large team of over 300 people, and time management within the remits of ordinary business hours.
I was initially nervous about applying for this fellowship , since I did not have research interests directly related and because it meant time away from my PhD after only 1 year. However, I was willing to learn quickly and so the fellowship was a great experience and one that was definitely timely. I can use the experiences gained at the ONS throughout the remainder of my PhD, so my advice would be to pinpoint exactly what you want to get from an fellowship and then just go for it!”
I would definitely recommend an fellowship . It’s a good break from the PhD and means you can gain valuable experience and expertise that will mark you out from other candidates in the post-PhD job market.
Konstantinos Voudouris’s placement with ONS took place between July and October 2021
Recent Comments