Etiologic Studies’ Objects Design

Abstract

One of Immanuel Kant’s most seminal teachings was the idea that a scientist does not learn by simply observing, ‘reading’ Nature; that (s)he learns by reading into Nature the terms in which (s)he thinks about Nature. It is in these terms that (s)he, in a study, ‘interrogates’ Nature – seeks tentative answers to questions (s)he poses.

In respect to population-level etiogenetic research this means that, preparatory to a study, the researchers think, deeply, about the rate of occurrence of a health phenomenon – the event of a cancer becoming clinically manifest, say – in a particular domain – initially defined by a range of age alone, perhaps. And they think about the phenomenon’s rate of occurrence in causal relation to something in the people’s past – their diets’ antioxidant content, perhaps.

With such a sketchy point of departure, the investigators need to proceed to refine their idea, in that example the particular meanings they elect to associate with the generic terms in: the (rate of) occurrence of ‘cancer’ in causal relation to ‘histories’ in respect to their diets’ ‘antioxidant’ content, this in a particular ‘domain’ of the occurrence. They need to decide whether their thinking in this is specific to a particular type of cancer and a particular type of antioxidant, or whether it is generic in one or both of these respects. They need to specify their thinking in respect to the range of time for the etiogenetic role of dietary antioxidants, retrospectively as of the outcome event (its occurrence/non-occurrence). And they need to decide whether the domain of the study is to be age-restricted and one of no previous clinically manifest cancer, for example.

Upon decisions like these, the investigators go on to design a statistical model (log-linear) for the outcome’s occurrence in the defined domain, with determinants in the form of statistical variates (Xs). The parameters of Nature constituting the objects of study are imbedded in this model, the design of this model amounting to the study’s objects design.

A future with examples like this is envisioned in this chapter. As of now, a study’s objects design is not even in the common vocabulary of epidemiological research.