To provide clients with the best possible treatment, social workers require fallacy recognition skills. Ideally, these skills are already acquired during university education. We conducted an experimental study with a 2 × 2 factorial design (and an additional baseline condition) with worked examples (with/without) and external scripts (with/without) as scaffolds to support social work students' (N=130) fallacy recognition skills during case-based reasoning on the exchange of fallacious arguments of social workers in the context of a web-based learning environment. Students significantly improved their fallacy recognition skills over time. During treatment, worked examples and external scripts yielded positive effects when presented in isolation, but not when combined. However, the effects of the two scaffolds could not be maintained in the posttest. Further systematic research is needed to identify principles for an effective design of additional instructional guidance.