Overview 

Between 2017 and 2019, Dr George Tzougas carried out three research projects in collaboration with six undergraduate interns. The projects ran between June and August, with each research intern completing 100 hours of paid work.

Target audience

Third year undergraduate students

Details

The research projects ran from Monday 8th June – Friday 31st August and during that time the undergraduate research assistants completed 100 hours of paid work. In particular, in 2017 I was very enthusiastic to get involved in the undergraduate research internship scheme which was launched by LSE LIFE and LSE Teaching & Learning Centre. The research paper that the students and I produced in 2017 recently appeared in the internationally recognised European Actuarial Journal, was presented in the Statistics Department in November 2017 and at the 10th International Conference on Computational and Methodological Statistics which was hosted by the University of London in December 2017. The main purpose of this paper was to propose an EM scheme that reduces the computational burden for ML estimation in the Negative Binomial-Inverse Gaussian (NBIG) regression model. The NBIG regression model extended the commonly used specification that assumes that the number of claims is distributed according to a mixed Poisson regression model, which was widely accepted for insurance ratemaking. Furthermore, the project which I did together with two students in 2018, was presented in the Statistics Department in November 2018 and at the 11th International Conference on Computational and Methodological Statistics which was hosted by the University of Pisa in December 2018. The second paper was concerned with presenting the Exponential-Lognormal regression model as a competitive alternative to the Pareto, or Exponential-Inverse Gamma, regression model that has been used in a wide range of areas, including insurance ratemaking. This was the first time that the Exponential-Lognormal regression model was used in a statistical or actuarial context. The main contribution of the study was that we illustrated how maximum likelihood (ML) estimation of the Exponential-Lognormal regression model, which does not have a density in closed form, can be accomplished relatively easily via an Expectation Maximization (EM) type algorithm.

Regarding the application process, students who were interested in applying for one of the two available posts were asked to send a CV and cover letter explaining why they were interested and how their current studies / courses made them suitable for the internship.  Indeed, it was very encouraging that we received a high volume of applications. Specifically, we received more than 70 applications from students who were interested in participating in these projects. Due to the high number of applications, a selection procedure was carried out by members of the Department of Statistics, designed so as to identify students more motivated and better prepared for a satisfactory research internship.

The research paper that the Department of Statistics Undergraduate students Wei Li Hoon and Jun Ming Lim, along with Dr Tzougas produced in 2017, appeared in the internationally recognised European Actuarial Journal.  It was also presented at the 10th International Conference on Computational and Methodological Statistics (December 2017).  The 2018 project (with Woo Hee Yik and Muhammad Waqar Mustaqeem) was presented at the 11th International Conference on Computational and Methodological Statistics, hosted by the University of Pisa (December 2018).  The project's findings will be published in Annals of Actuarial Science. 

The most recent research project was presented in the Department of Statistics in November 2019 and was presented at the 12th International Conference on Computational and Methodological Statistics (December 2019). Dr Yunxiao Chen supervised this project, in collaboration with Ms Sarah Hagart, Head of Management Information in the LSE Planning Division, on grade inflation at the LSE.

Reflections on experiences

“Previously, I would normally take science as face value, rarely questioning the premise. As I approach researching new ideas, I believe I am better equipped to analyse the research rather than blindly believing it. Additionally, knowing how to sieve through a huge amount of information by applying a disciplined approach is extremely beneficial, even in my current job as a actuarial consultant researching new topics and techniques.”  - Intern in 2017

“It was a novel opportunity for a fresh graduate to work on cutting edge problems in the practical world, and have a chance to get published as a co-author at such an early stage of my educational career.”  - Intern in 2018

"Especially noteworthy is the outstanding collegiality and support that exists among the Department of Statistics staff’s members..." - Dr George Tzougas

"To claim that this experience has been one of the highlights of my time in LSE is indeed no exaggeration." - Jeria Kua

Impact

The research internship scheme is an excellent opportunity for 3rd year students to learn new statistical tools, build research skills, learn how to write a research article and provide them with the ability to think in a critical manner by making formal inferences on the basis of real statistical data and other fundamentals upon which they will draw in their professional careers. Finally, especially noteworthy is the outstanding collegiality and support that exists among the Department of Statistics staff’s members, which resulted in bringing both projects into fruition in a timely manner and hence place our Department into an elite group of few Departments in which students can gain a first-hand experience of research by publishing their work in prestigious academic journals. Finally, it is worth mentioning that students return to give a presentation to other students.  This gives other students an insight into the research process and demonstrates how skills they have developed during the programme can be applied.

Next Steps

The Department of Statistics values this activity very highly and there was a plan to do a summer research project with the Academic Registrar’s Division which would involve analysis of LSE student data, but this has been put on hold due to the coronavirus pandemic.