This section addresses the theoretical differences and similarities amongst
the three approaches. Furthermore, the differences and similarities in the
implementation of each approach are explored.
7.3.1 Theoretical Underpinnings of DBDM, Afl, and DT
To answer our first question (What are the similarities and differences in the
theoretical underpinnings of DBDM, AfL, and DT?), we compared the underlying learning
theories of these approaches and their goals
Table 7.1 shows that DBDM, AfL, and DT are underpinned by elements from
different learning theories. Consequently, the goals of the three approaches differ
substantially; each approach aims to promote learning through different mechanisms, which
results in different expectations of the roles of teachers, students, and other actors in the
learning, assessment, and feedback process. These expectations sometimes contradict each
other; for example, in traditional views on DBDM, the responsibility for the assessment
process is primarily in the teacher‘s hands, whereas in AfL, the teacher and students share this
responsibility, e.g., in the form of self- and peer assessment (Stobart, 2008; Wiliam, 2011).
However, recent literature on DBDM shows a shift towards shared teacher-student
responsibility for assessment (Schildkamp et al., 2013).
We observed that all three approaches converge, following a view of learning based on
the principles of meta-cognitivism, social cultural theory, and (social) constructivism. Central
to these learning theories is the acknowledgement that most knowledge and skills are
inextricably connected to the context in which they are taught. Nevertheless, as described in
the next section, this convergence does not affect the implementation of each assessment
approach to the same degree.
7.3.2 Implementation of DBDM, AfL, and DT
To answer our second question (What are the consequences of these similarities and
differences for implementing DBDM, AfL, and DT in educational practice?), we compared
the aggregation levels, assessment methods, and feedback loops of the three approaches
(Table 7.2). Figure 7.1 shows the overlapping levels of the decisions in the three approaches.
In DBDM, data are aggregated at the school level to make decisions with regard to improving
the school‘s quality (formative evaluation), as well as to judge the latter (summative
evaluation). Additionally, data are used at the class and student levels to adjust instruction to
meet the students‘ needs (formative assessment). The latter overlaps with AfL. DT solely
focuses on assessment and instructional decisions at the student level. Because the three
approaches aim to promote learning at different aggregation levels, they can complement each
Figure 7.1. Overlapping levels of the decisions in the three approaches.
The diversion in the goals of DBDM, AfL, and DT is associated with a large variety of
the use of assessment methods. For example, AfL uses classroom conversations, while
DBDM and DT often employ standardised tests. The different choices in the use of
assessment methods are primarily associated with the nature of the data, and the purposes and
stakes regarding the use of these data. In DBDM, most data are quantitative in nature;
especially at the school level, high-quality data are needed as the stakes are often higher. In
contrast, most data are qualitative in AfL because they mainly aim to provide immediate
information, which informs decisions on how to direct learning processes. These are lowstakes
decisions; if the adaptations in the learning environment do not produce the intended
effects, this will become quickly clear from subsequent assessments, whereupon the
adaptation strategy can be changed. Thus, the adaptation process is flexible. In DT, finegrained,
quantitative data are usually gathered and translated into qualitative statements on
which teachers can take immediate actions. Although DT uses quantitative data similar to
DBDM, the quality requirements are different from those of DBDM. The stakes associated
Data-Based Decision Making, Assessment for Learning, and Diagnostic Testing in Formative
with DT decisions are usually lower because the decisions concern individuals, and the
feedback loops are shorter.
With respect to feedback mechanisms, we found the use of feedback loops in all three
approaches. However, because the approaches aim at formative assessment and formative
evaluation at different levels, these feedback loops also take place at various levels and
frequencies. In DBDM, the retroactive feedback loops that occur at the school level are spread
out over time. In AfL, continuous dialogues and feedback loops are essential, which results in
short, frequently interactive, and sometimes retroactive or proactive feedback loops.
Regarding DT, the length of feedback loops should match the student‘s learning curve for the
subject matter that is the assessment‘s objective.