Search results

This paper proposes a benefit of the doubt (BoD) approach to construct and analyse teacher effectiveness scores (i.e. SET scores).

The BoD approach is related to data envelopment analysis (DEA), a linear programming tool for evaluating the relative efficiency performance of a set of similar units (e.g. firms, departments, individuals, etc.) who use (possibly multiple) inputs to produce (possibly multiple) outputs in operation environments typically characterised by no reliable information on the prices of inputs and/or no (exact) knowledge about the “functional form” of the production or cost function.

A major appeal of BoD is the flexibility in the construction of the SET scores. In particular, BoD puts teacher performances into a relative perspective to be evaluated optimally, thereby accounting for different values and interpretations that teachers attach to “good teaching”. Furthermore, if available, stakeholder opinion can be easily incorporated into the evaluation. The swift identification of teachers' relative strengths and weaknesses is another advantage. The advantages show in the application.

Several issues specific to the construction and analysis of SET scores have remained largely untouched, and hence unresolved, in the literature. One particular blank area is the weighting and aggregation of SETs into teacher performance scores. This paper contributes to the literature in that it presents a methodology that addresses this issue. To illustrate the usefulness of BoD for teacher evaluations with summative intentions and/or formative purposes, the methodology is applied to SET data collected at a university college in Brussels (Belgium).

The purpose of this paper is to propose a non‐parametric methodology to construct composite scores of citizen satisfaction with local police services. More precisely, the paper advocates a custom made version of the popular Data Envelopment Analysis approach, also referred to as the “Benefit‐of‐the‐Doubt” model. The key advantage of this approach is that it weights the citizen satisfaction rates with the multiple local police functions and tasks into the composite score in an endogenous manner, thereby allowing for different values and interpretations of “good local policing” among police services. The methodology is illustrated with citizen satisfaction data on a sample of Belgian local police services.

The paper develops a multidimensional measure of local police effectiveness based on citizen satisfaction measures. It uses a non‐parametric evaluation methodology related to the popular DEA‐model. The paper looks for strengths and weaknesses in the performances of local police services both at the micro level (per local police service) and at the macro level (region).

With an average overall satisfaction score of 91.94 per cent, it seems safe to say that the majority of the citizens are generally satisfied with local police services. The BoD‐model identifies per local police service the basic functionalities that citizens rated relatively highly and poorly. Results show that urbanization per se does not drive the satisfaction scores of the participating local police services. Of much more importance are the regional disparities. Participating local police zones in Flanders receive higher satisfaction scores than those in the Walloon Region (the two big regions in Belgium). One of the findings is that the BoD‐model offers both conceptual and practical advantages in the evaluations of local police services. BoD grants each police service the benefit‐of‐the‐doubt in the effectiveness evaluations. The BoD‐model also identifies the factors of strength and weakness that explain the citizen satisfaction scores.

To the authors’ knowledge, this is the first time that multidimensional scores of citizen satisfaction are used in the effectiveness evaluations of local police services. The paper uses a methodology that accounts for the own particular circumstances of the local police services (by assigning in an endogenous manner weights in the evaluations).