Intended for healthcare professionals

Rapid response to:

On The Case

The case of the disappearing teaspoons: longitudinal cohort study of the displacement of teaspoons in an Australian research institute

BMJ 2005; 331 doi: https://doi.org/10.1136/bmj.331.7531.1498 (Published 22 December 2005) Cite this as: BMJ 2005;331:1498

Rapid Response:

Toward a Bayesian model of teaspoon location

The inspired research of Lim et. al. (2005) and consequent rapid
responses highlight that many factors are indeed relevant when it comes to
estimating teaspoon location. Due to lack of omnipresence, observation of
a population of teaspoons will tend to be partial, and as such suggests
the need for a probabilistic model of teaspoon location.

As such, the development of a factor analysis type model may be a
sensible progression. In this approach there are a set of locations, each
of which can contain a certain number of teaspoons, and a set of causal
factors that influence teaspoon location. There is uncertainty about
specific teaspoon locations, for example direct observation of the number
of teaspoons that have migrated to a remote island is impractical.
Therefore the number of teaspoons in each location should be represented
by a discrete-valued distribution. If we make a direct measurement of
number of teaspoons in a tea making area, then this distribution will be
highly peaked on that particular observation.

This model, from the Bayesian perspective, allows the use of prior
knowledge. For example we can formalise our prior belief in each factor
and teaspoon location by used of distributions. Evidence from other fields
would suggest the use of a zero-peaked prior, as many possible factors
such as quantum effects are a lot less likely than others, such as
anthropogenic effects.

Examples of these causal factors that have previously been mentioned
are gender, quantum, and socioeconomic constraints. Future quantitative
teaspoon studies will aid our statistical inference into how each
hypothesised factor influences teaspoon population density in different
locations.

Competing interests:
None declared

Competing interests: No competing interests

18 January 2006
Benjamin T Vincent
Postdoctoral Researcher
Department of Experimental Psychology, University of Bristol, BS8 1TN