It Came from Beneath the Sea! Report on Loch Ness Monster Sightings Transforms Anecdotes into Data

A recent report by Dr Charles Paxton, from the University of St Andrews’ Centre for Research into Ecological and Environmental Modelling, explores how a database of Loch Ness Monster sightings has been used to teach students about the importance of a critical approach to data.

The database is composed of over 1800 anecdotal reports of ‘Nessie’ up to 2016, both from public documents and unpublished records, from as far back as 1934. Dr Paxton worked in collaboration with Adrian Shine from the Loch Ness Monster Project and Dr Valentin Popov from the School of Mathematics and Statistics in amassing the collection. Shine described his role in the project as gathering “report data from my years at the Loch Ness Centre,” especially reports from the Loch Ness Phenomenon Investigation Bureau — his organisation’s predecessor. While this data has been sorted into a digital format, Shine added that “much archival material is to be deposited at the Highland Archive Centre in Inverness.”

When using this database for teaching, Dr Paxton first asked students if they considered anecdotes to be a form of data. Only nine of his twenty-nine master’s students saw this to be true, while twenty-two of the seventy-one undergraduates thought the same. The next step was to prompt students to debate what a suitable statistical population of Loch Ness Monster Reports would look like. Dr Paxton vouches for the importance of this approach in analysing data, stating that it is “an important, but often overlooked, step”. After this discussion, the conclusion was reached that the reports must involve a witness, data collector, and data analyst.

The multifaceted nature of the data is reflected in the method of cataloguing it; Dr Paxton stated that he recorded “what is described, when reports happen, [and] where they happen.” Understanding that each report represents these different steps of formation allows the assessment of issues of independence and bias.

Firstly, the database invites analysts to scrutinise the independence of sightings. In his publication, Dr Paxton explained that alleged witnesses confer amongst each other, describe the same sighting many times, and subsequently that one report may be made by multiple witnesses. Therefore, assuming that all 1800 reports are entirely independent could be pseudoreplication — the wrongful estimation of data as independent. This is an example of how the Loch Ness Monster database has been used to stimulate a close interrogation of appropriate statistical populations.

Secondly, the exercise described in Dr Paxton’s report confronts biases in anecdotal data. An exploratory approach to the data exposes patterns in the reports which can highlight the bias of anecdotes. For instance, most sightings were reported in the summertime, specifically the midmorning and afternoon. While this may suggest the best times for a sighting to those who believe in the existence of this mythical monster, this trend in the data could also reflect the behaviours of witnesses. A further instance of bias in the database is that the clearest reports are more likely to survive and therefore be a part of the statistical population. Recognising that this characteristic of the data again provides a valuable reflection on the usefulness of a dataset.

Dr Paxton emphasised the importance of recognising biases within data, demonstrating the significance of this exercise for his students: “Any conclusions from biased data could be completely invalid.” He continued, “It would be like ascertaining the popular interest in golf by taking a sample of passersby from North Street, St Andrews.”

However, this does not discount the significance of anecdotes as data. When asked about the importance of anecdotal data in other fields, Dr Paxton affirmed that “understanding the value and limits of anecdotes and eyewitness testimony is vital to jurisprudence and the study of history.” Shine also advocated for the worth of anecdotes: “The latest project is a clear demonstration that anecdotes can be beaten into data.”

Image by Wikimedia Commons