Text as Data final project.
Alı́ J. Limón, Felipe N. Ducau
New York University
In this project we analyze if it is possible to determine if a scientific publication contains within its text some information about its replicability. Our hypothesis is that there is sufficient information embedded in the body of a published paper to recognize if the reported results are reproducible or not. Our question is, do scientist write in a different way when they are presenting flawed or false results? The way the results are presented, the language used and the level of obfuscation when presenting results might lead us to determine the reproducibility of the research.
Results and description are in the file TAD_Final_Project.pdf