The consumption of ice cream increases, so do the instances of drowning - Essay Example

Conceptual Models By Jimmy An analysis of the ment: “As the consumption of ice cream increases so do the instances of drowning” Correlation exists between two variables when a variation in one is found to correspond to a proportional variation in the other either positively or negatively. To take this further by implying that one therefore causes the other is a task of interpretation. A correlation is not sufficient for causality but it does need to be present for a possibility of a causal relationship to be identified.

This interpretation based on the correlation may be reasonable but it could also very well be spurious. Coincidence is one explanation of the latter, but there could also be other variables that have not been considered. Ice-cream consumption (first variable) and the incidence of drowning (second variable) may show a positive correlation, but the consumption does not cause the incidence. This may be common sense, but this simple example demonstrates the issue of spurious correlation being taken as evidence for implying causality.

In Correlation Theory, the coefficient of correlation simply measures the ‘goodness of fit’ of sample of the data for the two variables i.e. how well they correlate together. It “does not necessarily indicate a direct dependence of the variables” (Spiegel, M).“The problem of interpretation is always very much more difficult to deal with than the statistical manipulations, and for this side of the work there is no substitute for detailed practical acquaintance with every aspect of the problem.

” (Moroney, M)Conclusions based on the correlation findings are only good as the interpretative ability of the analyst and his or her acquaintance with every detail or aspect of the problem. The possibility of other variables affecting the ones being compared should always be considered to account for the relationship being tested if there is any inkling of a spurious correlation. In our example, a third variable of temperature relates to both the existing variables of ice-cream consumption and instances of drowning and better explains the phenomenon.

More ice-cream is consumed in hot weather especially during the summer heat. This tendency to consume more during hot weather is understandable my most people, and can easily be ascertained by examining the sales figures of ice-cream parlours. These are well known to show seasonal fluctuations being higher in summer and lower in winter. As for drowning, this too can be explained as being a more common occurrence in summer simply because more people like to and do spend more time either in or near water.

This can be near lakes, rivers, seasides, and includes drowning from prolonged submersion especially by young children during bathing or whilst playing unsupervised in a tub of water. Hence in our case, temperature positively correlates with the two variables to satisfactorily explain the phenomenon much better than the statement. Its inclusion in the conceptual model is therefore requisite and basic. If this variable is not included, the model would be lacking, distorted and biased.The example of ice-cream and drowning is often given not just to demonstrate the silliness of believing in spurious correlation, but to show that even in many other ‘believable’ cases where two factors indicate a strong correlation, to exercise caution in jumping to a conclusion implying causality.

The strength of the correlation and its significance can also differ, for which statistical tests can be conducted. Furthermore, even if causation is ascertained, we must remember that in many cases the relationship is not necessarily definitive but a probabilistic one. That is, we actually express the likelihood of one effect causing the other.It is therefore important for the decision analyst to identify alternative explanations to see which are viable and which are not. And, depending on the nature of the problem and the importance of the decision, this has to be an ongoing process too.

I propose the following 3 scenarios to explain the phenomenon in question:1. More ice-cream is consumed in summer, but more people spend time in/near water in summer too. Therefore test the variable of temperature against the two existing variables.2. Most ice-cream parlours are located at places near water (e.g. seasides and lakes). Therefore test the variable of location against the two existing variables.3. Many people who go for a swim like to eat ice-cream first. Therefore test the variable of what has been eaten prior to getting into water against the second existing variable.

These have been given in order of what I think are the most likely (first) to the least likely (third) scenario. If these scenarios were tested for validity as alternatives to the original, they would more accurately explain the phenomenon and help to produce a better model. It is not that the statement is technically wrong. It’s just the connection is spurious and cannot therefore be used as a basis for decision-making. For instance, a policy to ban ice-cream on these grounds that it leads to drowning would be ludicrous.

Even considering a ban near pools, lakes, beaches etc. would be unnecessary and not reduce the risk of drowning. Instead, the “well-intentioned policy would have no impact on drowning, because we failed to look at any other variables that might actually be associated with both drowning and ice cream consumption” (Schulze, 2006). Suggesting the reverse (that drowning encourages greater ice-cream consumption) would be even more nonsensical. Also, following this logic, perhaps shark attacks are also caused by eating too much ice-cream!

Really, in such scenarios, there is basically insufficient information to make a causal claim and a need to find other variables or alternative causes, and isolating and contesting these causes. A more complex model could also consider including daylight hours and people’s free time as other related factors besides temperature.ReferencesMoroney, M. J. (1990). Facts From Figures. Penguin Books. Ch. 16, Pg. 303.Schulze, P. A. (2006). Perennial Issues in Child Development. Retrieved 16 Nov. 2008 from http://www3.uakron.edu/schulze/610/lec-issues.htm.Spiegel, M.

Theory and Problems of Statistics. Schaum’s outline series. McGraw Hill, Inc.