The Philosophy of Information Quality by Luciano Floridi & Phyllis Illari

Author:Luciano Floridi & Phyllis Illari
Language: eng
Format: epub
Publisher: Springer International Publishing, Cham

9.5 Underdetermination of Evidential Significance

The primary use of QATs is to estimate the quality of information from particular medical studies, and the primary use of such information is to estimate the strength (if any) of causal relations. The relata in these purported causal relations are, of course, the medical intervention under investigation and the change in value of one or more parameters of a group of subjects. The best available QATs appropriate to a given domain differ substantially in the weight assigned to various methodological properties (Sect. 9.2), and thus generate discordant estimates of information quality when applied to the same information (Sect. 9.4). The differences between the best available QATs are fundamentally arbitrary. Although I assume that there must be a unique value (if at all) to the strength of purported causal relations in the domains in which these tools are employed, the low inter-tool reliability of QATs—together with the fundamentally arbitrary differences of their content—suggests that, in such domains and for such relations, there is no uniquely correct estimate of the quality of information. This is an instance of the general problem I call the underdetermination of evidential significance.

Disagreement regarding the quality of information in particular scientific domains has been frequently documented with historical case studies. One virtue of examining the disagreement generated by the use of QATs is that such disagreements occur in highly controlled settings, are quantifiable using measures such as the κ statistic, and are about subjects of great importance. Such disagreements do not necessarily represent shortcomings on the part of the disagreeing scientists, and nor do such disagreements necessarily suggest a crude relativism. Two scientists who disagree about the quality of information from a particular study can both be rational because their differing assessments of the quality of that information can be due to their different weightings of fine-grained features of the methods which generated the information. This explains (at least in part) the low inter-rater and inter-tool reliability of QATs.

Concluding that there is no uniquely correct determination of the quality of information by appealing to the poor inter-rater and inter-tool reliability of QATs is not merely an argument from disagreement. If it were, then the standard objection would simply note that the mere fact of disagreement about a particular subject does not imply that there is no correct or uniquely best view on the subject. Although different QATs disagree about the quality of information from a particular study, this does not imply that there is no true or best view regarding the quality of information from this particular trial—goes the standard objection—since the best QATs might agree with each other about the quality of information from this trial, and even more ambitiously, agreement or disagreement among QATs would be irrelevant if we just took into account the quality of information from this particular trial by the uniquely best QAT. The burden that this objection faces is the identification of the single best QAT or at least the set of good ones (and then

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.