SAFETYLIT WEEKLY UPDATE

We compile citations and summaries of about 400 new articles every week.
RSS Feed

HELP: Tutorials | FAQ
CONTACT US: Contact info

Search Results

Journal Article

Citation

Urbaniok F. Fortschr. Neurol. Psychiatr. 2004; 72(5): 260-269.

Vernacular Title

Validitat von Risikokalkulationen bei Straftatern -- Kritik an einer methodischen

Affiliation

Psychiatrisch-Psychologischer Dienst Justizvollzug Kanton Zürich/Schweiz. frank.urbaniok@ji.zh.ch

Copyright

(Copyright © 2004, Georg Thieme Verlag)

DOI

10.1055/s-2004-818395

PMID

15136947

Abstract

Catalogs of multidimensional criteria have become established in continental Europe. In contrast, due to the limited reliability of clinical prognoses, procedures such as "Actuarial Risk Assessments" have been established in Anglo-Saxon countries as a means of assessing prognoses. The basic assumptions of "Actuarial Risk Assessments" (principally a statistical methodology) show a remarkable discrepancy compared with the fundamental ideas of the European experts (clinical assessment procedures), although this has not to date been discussed in depth in the literature. Apart from this discrepancy, however, certain basic assumptions are internationally regarded as generally accepted methodological cornerstones of prognostic theory. Thus, such a generally unchallenged "essential" methodological principle as the "Two by Two Table" has been mentioned for many years in a dichotomous sense in numerous textbooks or basic lecture on prognostic theory. Far-reaching conclusions concerning the validity of prognostic statements are made on the basis of this methodological axiom. Thus, the putative error rate for prognostic assessments is derived from so-called "false positive prognoses", and, among other things, the unjustified committal of offenders wrongly judged to be dangerous is criticised. Surprisingly, there is no discussion in the considerable literature currently available, that it is methodologically highly questionable to use the Two by Two Table that is adequate for dichotomous decision-making, without using further differentiation with respect to probability factors, as is undoubtedly the case for risk calculations. The article discusses the above mentioned problems and gives possible approaches for solutions.


Language: de

NEW SEARCH


All SafetyLit records are available for automatic download to Zotero & Mendeley
Print