This is the de-identified data set used to conduct the analyses of our study "A Use-Case Specific Framework for Designing Representative Vignettes (RepVig) and Evaluating Triage Accuracy of Laypeople and Symptom-Assessment Applications" (https://doi.org/10.1101/2024.04.02.24305193). The data comprises the answers to cases given by laypeople, symptom-assessment applications, and large language models and the corresponding solutions for each case. The cases were developed in the study with a focus on external validity. The dataset contains three datafiles: collected data for laypeople, for symptom-assessment applications, and for large language models.