Causal models

The need for a causal model to guide data collection is emphasized. This makes the connection between the nature of the error and the PIFs in the situation. 

There is considerable overlap between the processes of data collechon and interpretation as discussed in earlier sections of this chapter. The nature of the data collected will be strongly influenced by the assumed relationship between the observable characteristics of errors and their underlying causes. Similarly, the interpretation process will also be driven by the causal model. 

A specific example of a causal model is the root cause tree described in Section 6.8.4 and Figure 6.8. This is a very elaborate model which includes several levels of detail for both equipment and human causes of incidents. The root causes tree is a generic causal model, and may require tailoring for application to specific plants and processes (e.g., in the offshore sector) where other error causes may need to be considered. 

Microarray data cannot be analyzed by purely brute force techniques to generate a causal model of a set of biological processes because the data represents gene expression patterns that are only correlated with temporal processes of interest in the organism. Lander (1999) comments on this problem as follows ... 

Pearl, J. (2000), Causality Models, Reasoning and Inference, Cambridge University Press, New York. 

Computational methods have been applied to determine the connections in systems that are not well-defined by canonical pathways. This is either done by semi-automated and/or curated literature causal modeling [1] or by statistical methods based on large-scale data from expression or proteomic studies (a mostly theoretical approach is given by reference [2] and a more applied approach is in reference [3]). Many methods, including clustering, Bayesian analysis and principal component analysis have been used to find relationships and "fingerprints" in gene expression data [4]. 

Bayesian statistics are applicable to analyzing uncertainty in all phases of a risk assessment. Bayesian or probabilistic induction provides a quantitative way to estimate the plausibility of a proposed causality model (Howson and Urbach 1989), including the causal (conceptual) models central to chemical risk assessment (Newman and Evans 2002). Bayesian inductive methods quantify the plausibility of a conceptual model based on existing data and can accommodate a process of data augmentation (or pooling) until sufficient belief (or disbelief) has been accumulated about the proposed cause-effect model. Once a plausible conceptual model is defined, Bayesian methods can quantify uncertainties in parameter estimation or model predictions (predictive inferences). Relevant methods can be found in numerous textbooks, e.g., Carlin and Louis (2000) and Gelman et al. (1997). 

Central to any risk assessment is a model of causality. At the onset, a conceptual model is needed that identifies a plausible cause-effect relationship linking stressor exposure to some effect. Most ecological risk assessments rely heavily on weight-of-evidence or expert opinion methods to foster plausibility of the causal model. Unfortunately, such methods are prone to considerable error (Lane et al. 1987 Hutchinson and Lane 1989 Lane 1989), and attempts to quantify that error are rare. Although seldom used in risk assessment, Bayesian methods can explicitly quantify the plausibility of a causal model. 

E. J. Sonuga-Barke, Causal models of attention-deficit/hjrperactivity disorder from common simple deficits to multiple developmental pathways. Biol. Psychiatry, 2005, 57,1231-1238. 

In an influential article, Gary Becker and Kevin Murphy (1988) present a model of rational addiction. The model has two main aspects. On the one hand, it offers a simple causal model of the consequences of consuming addictive substances. On the other hand, it offers a standard belief-desire account of how people might choose to engage in such consumption. 

The Becker-Murphy causal model offers a valuable bridge between economics and neurophysiology. This creation of a common language is perhaps their most important contribution to the analysis of addiction The causal model is in fact retained by several writers in this volume who do not share the belief-desire account of addiction. 

Since, in this causal model, the extended wave 0 represents a real physical finite wave with well-defined energy, it seems natural to represent it by a suitable mathematical form. At the time when de Broglie put forth his causal interpretation of quantum mechanics, it was necessary for him to construct a finite wave using the Fourier analysis, namely, the multiplicity of harmonic plane waves, infinite in space and time, summing up and giving origin to a wavepacket. 

Figure 16. Visibility predicted by the two theories dashed line, usual model continuous line, causal model.

The explanation for the absence of interference comes naturally from the causal model of a particle whose undulatory part is described by a finite localized wavelet. In this situation, the limitless spreading of matter wavepackets originates from the fact that in the initial burst, coming from the source, each individual localized quantum particle travels at a different velocity. Therefore, as the time increases the distance among them also increases, as shown in Fig. 24. 

---