The Scientific Approach Developing a Model

Is there something special about the way scientists think If we could break down a typical modern scientist s thought processes, we could organize them into an approach called the scientific method. This approach is not a stepwise 

Observations. These are the facts that our ideas must explain. Observation is basic to scientific thinking. The most useful observations are quantitative because they can be compared and allow trends to be seen. Pieces of quantitative information are data. When the same observation is made by many investigators in many situations with no clear exceptions, it is summarized, often in mathematical terms, and called a natural law. 

Hypothesis. Whether derived from actual observation or from a spark of intuition, a hypothesis is a proposal made to explain an observation. A valid hypothesis need not be correct, but it must be testable. Thus, a hypothesis is often the reason for performing an experiment. If the hypothesis is inconsistent with the experimental results, it must be revised or discarded. 

Experiment. An experiment is a clear set of procedural steps that tests a hypothesis. Often, hypothesis leads to experiment, which leads to revised 

---

In the system presented here, ontologies provide the base for a formal description and a classification of information resources. Ontologies have been implemented in order to describe the relation of scientific documents (publications, reports), similar to the ScholOnto -approach developed by Motta et al. (2000). Additional ontologies describe concepts of scientific research that have to do with the development of mathematical models (Weiten et al., 2002 and Weiten et al., 2002b). These allow a linking of different resources based on common concepts (e.g. model parameters). 

The development of models incorporating biomarker assays to predict the effects of chemicals upon parameters related to r has obvious attractions from a scientific point of view and is preferable, in theory, to the crude use of ecotoxicity data currently employed in procedures for environmental risk assessment. However, the development of this approach would involve considerable investment in research, and might prove too complex and costly to be widely employed in environmental risk assessment. 

---