Falsifiability Vs The Arkansas Board Of Education

Falsifiability ensures the integrity of socially important science based decision making

Task-based assessment is used to support many socially important decision makings that may directly affect public health. How can one deliberately use a method that can be wrong? This is a good question. Allow me to answer this question with facts.

In 1982, McLean sued the Arkansas Board of Education, arguing that the mandatory teaching of creation science in public school is unconstitutional because creation science is not a science. Judge Overton’s ruling in favor of McLean is based on the “essential characteristic of science”, which includes that “its conclusions are tentative, i.e. are not necessarily the final word” and “it is falsifiable”36. Evolution

non-science or pseudo-science2. Popper’s idea is massively influential, and therefore widely debated among philosophers. The main objections, to my understanding, include that the standard is too high and Popper’s implementation of falsifiability is too naïve. For example, Kuhn believes that falsifiability is a characteristic of revolutionary science rather than that of the day-to-day practice39. He also believes the understanding the psychology of research has more practical value than prescribing the logic of discovery39. Lakatos believes that one counterexample falsifies the whole theory is too naïve and that a scientific theory is a program rather than an individual claim3. Lakatos subsequently proposed “progressive research programme”3 to address these issues. To handle counterexamples, he prescribed the “sophisticated falsificationism”, which is used in developing the ESS theory as mentioned above. Nevertheless, there is a consensus that falsifiability is a desirable characteristic of science, for it ensures that science is grounded on evidence rather than

Particularly, the ESS theory suggests that two characteristics are required to fully quantify search: LKE ROC and M*. The theory receives corroborations in the various scenarios tested with one failed testing27,30. The failed testing suggests more than two characteristics are required. Because current theory and evidence suggest at least two characteristics are needed to fully quantify search, and LKE ROC is only one of the two, the use LKE ROC alone may lead to biased evaluations of images to be used in clinical tasks that involve search. Literature findings also provide empirical evidence that the use of LKE and search tasks may result in inconsistent performance rankings40-42. Practically, this conclusion may have significant indications, particularly in the extensive use of LKE model