Developing models to predict outcomes is essential in the biomedical sciences. A new article by Sreeram Ramagopalan (F Hoffmann-La Roche, Basel, Switzerland) and colleagues describes the steps to build a simple, predictive model. The team demonstrated the utility of their results by applying them to the popular show Love Island, providing a way for Love Island hopefuls to predict their likelihood of winning the show.
The article is published as a Christmas special in the open access, biomedical and biotechnology journal, Future Science OA.
Love Island originally aired in the UK in 2015, and has since become popular across the globe. In the show, contestants live on-camera, in an isolated villa. To maintain their ‘survival’ they form relationships with other contestants – be that love or friendship – and other contestants and the public vote as to whether they stay on the show. Relationships are made at the beginning of the show – based on first impressions – but may change over time, in ‘recouplings’. The show ends with the couple with the most votes winning (in the UK version) a cash prize of £50,000.
Post-show earnings have been estimated to have a lifetime added value greater than obtaining a degree from top universities, and the show has seen participation from medical and scientific professionals, although none have thus far made the top three.
However, attempting to win the show was not the driving factor for the study. “Whilst it was of interest to see if there was anything that could actually predict success on Love Island, the primary purpose of this analysis was to describe the steps to build predictive models in general,” noted Ramagopalan. “These predictive models are used in healthcare generally, yet many people are unclear as to how they are derived.”
As such, depending on the data you use, the method used – here resulting in a model named DO-BITS – can be used to both predict your likelihood of winning Love Island and your likelihood of having a heart attack.
“Using data on all contestants participating in the UK version of the television show Love Island we tried to analyze what predicts success on the show (i.e., finishing in the top 3),” Ramagopalan explained. “We discovered that having a four-letter first name, being a brunette and working as a tradesman was linked to doing well. In addition, behaviour on the show, including the number of relationships you have and ‘doing bits’ also helped. Whilst this is based on information from previous shows, it could help people decide whether or not they want to apply to participate on Love Island.”
The brief history of Love Island meant only a small sample size was available, meaning the method needs to be validated.
The article “Derivation of DO-BITS, a score for predicting success on Love Island: a retrospective, observational cohort study (doi:10.2144/fsoa-2020-0168)” is available free to view in Future Science OA.
About Future Science OA
Launched in March 2015, Future Science OA is a peer-reviewed, open access, biomedicine and biotechnology journal from the Future Science Group. Its articles present the latest research, putting results into context for the future, all published using a CC-BY license. Alongside research, the journal features review articles, editorials and perspectives, providing readers with a leading source of commentary and analysis.
About Future Science Group
Founded in 2001, Future Science Group (FSG) is a progressive publisher focused on breakthrough medical, biotechnological, and scientific research. FSG’s portfolio includes two imprints, Future Science and Future Medicine. In addition to this core publishing business, FSG develops specialist eCommunities. Key journals and sites include Bioanalysis Zone, Epigenomics, Future Oncology and the award-winning Regenerative Medicine.
The aim of FSG is to service the advancement of clinical practice and drug research by enhancing the efficiency of communications among clinicians, researchers and decision-makers, and by providing innovative solutions to their information needs. This is achieved through a customer-centric approach, use of new technologies, products that deliver value-for-money and uncompromisingly high standards.