Etiologic Studies’ Methods Design

  • O. S. Miettinen
  • I. Karp


Once the study’s objects design has defined what the investigators are concerned to shed light on – the values of what parameters of Nature in what domain, conditionally on what extraneous determinants on the outcome phenomenon’s rate of occurrence – they can turn their attention to the study’s methods design. In it they define how they aim to ‘measure’ those parameters’ values – ultimately by means of fitting the logistic counterpart of the objects-defining log-linear model for the outcome’s occurrence to the data (the realizations of Y, X1, X2, …) on the study’s case and base series. The resulting empirical rate-ratio function translates into the result for the etiogenetic proportion of interest.

The methods design of an etiogenetic study proceeds in two distinct stages. In the first stage, the objects-defining occurrence relation that was designed in reference to Nature in the abstract is adapted to the actual study, translated into the operationalized model. The independent variates in the logistic model to be fitted are defined in reference to this counterpart of the theoretical model. The second stage of the methods design defines the means by which empirical content of that form is judged to be best obtained. While the essence of the study has an a-priori definition, many particulars of it remain to be specified/designed.

This second stage of the study’s methods design is governed by the imperative of validity assurance, descriptively, together with the desideratum of efficiency optimization. Both of these, in turn, are governed by principles of the studies’ methods design.

Apart from efficiency, a determinant of the study’s informativeness is its size; but this is not subject to principles-guided optimization, with one exception: when a violation of principle is detected in the study’s objects design or in its methods design in respects other than study size (number of datapoints in the two series), the optimal – indeed, correct – size of the study is zero.


Study Base Source Population Object Parameter Base Series Source Base 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Copyright information

© Springer Science+Business Media Dordrecht 2012

Authors and Affiliations

  • O. S. Miettinen
    • 1
    • 2
  • I. Karp
    • 3
  1. 1.McGill University Université de MontréalMontréalCanada
  2. 2.Cornell UniversityNew YorkUSA
  3. 3.Université de MontréalMontréalCanada

Personalised recommendations