RT Article
T1 On the use of inferential statistics on administrative police data
JF The crime data handbook
SP 197
OP 210
A1 Verlaan, Tim
A2 Langton, Samuel
LA English
YR 2024
UL https://krimdok.uni-tuebingen.de/Record/1919247084
AB This chapter aims to spark a conversation about the use of inferential statistics when using administrative police data. Through a (mini) scoping review of eight recent issues of journals within the field of criminology and crime science, we observe that all but one of the examined studies employ inferential statistics without meeting the requirements for generalization. We identify a number of negative consequences of this, namely: wrongful suggestions of generalizability of the observed difference/effects; undervaluing of the observed difference/effects; and obscuring of observed difference/effects. We hypothesize three possible drivers of misuse of inferential statistics: uncritical application of statistical methods (aided by easy-to-use statistical packages); the need for greater generalizability and impact of research findings; and the need for normative judgements on observed differences. In response to these three identified contributing factors, we call for: a general caution of inferential approaches and subsequent revaluing of descriptive statistics when working with administrative police data, and the opening up of a debate around (the need for) making normative judgements. Furthermore, we echo the call of Gibbs et al – who diagnose similar problems in educational research – for academic (crime) journals to ‘adopt or revise standards to make the interpretation of inferential statistics and p-values, such as defining the intended inferred population, a required part of reporting in analytic methods sections’.
NO Literaturverzeichnis: Seite 209-210
SN 9781529232042
K1 Police
K1 Data Records
K1 Statistics
K1 Inference
K1 Descriptive
K1 Error
K1 Measurement
K1 Research
K1 Methodology