The Erasmus+ funded project Diagnostic Tool in Mathematics (DiToM) addresses the need for national and international diagnostic instruments to assess mathematical key skills in the field of arithmetic and algebra which are necessary to acquire competencies that build on them. DiToM is a project that started in January 2023 as part of an Erasmus+ Cooperation Partnership (DAAD; KA220-HED-1194D587). The project consortium consists of the following seven institutions: Bielefeld University (Germany) as the coordinating partner, University of Paris-Est Créteil (France), Linnaeus University (Sweden), Free University of Bozen (Italy), University of the Aegean (Greece), University of Rijeka (Croatia) and the Spanish Federation of Mathematics Teachers (Spain).

Initial Situation:
National surveys in Germany reveal a growing number of students failing to meet minimum standards in mathematics. Similar trends and stagnating progress are also reflected in several European countries, as seen in international PISA results. These difficulties often stem from an inadequate understanding of foundational mathematical concepts, which in turn leads to persistent struggles with more advanced topics.

Background:
Screening procedures play a central role in preventing learning difficulties in lower secondary mathematics education. They allow for the early and reliable identification of individual risks for potential challenges with current and future mathematical content. Many existing diagnostic tools are either summative in nature or designed as time-consuming individual assessments. Screenings for optimised resource allocation are playing an increasingly central role for optimal individual support.

Objective:
This paper presents initial results from a preliminary study within the DiToM project aimed at developing and evaluating a screening tool to be used at the end of grade 6 or the beginning of grade 7.

Method:
The study focuses on the development and evaluation of a screening process using a sample of 69 seventh-grade students from Germany. A set of domain-specific predictors of mathematical performance in arithmetic and algebra was used for analysis.

Results:
Using a ROC analysis that incorporated a curriculum-based reference test and teachers’ assessments of students at risk, a predictive model was identified that demonstrates strong predictive accuracy (AUC > .80).

Discussion:
The results provide valuable insights into the application of a valid group screening method for the early identification of students who may encounter difficulties with current and future mathematics content in lower secondary education.

Keywords: Screening, ROC analysis, mathematical key skills, DiToM.