Describir: Large-scale assessment of the effect of popularity on the reliability of research.