Can We All Get Along? Integration of Classical Statistics and Modern Data Science

Chong Ho Yu, Ph.D., Psychology

Experts in data science predict that the size of digital data will double every two years. However, big data may present challenges to traditional data analysts, who are accustomed to use smaller and structured data sets. Further, lack of replicability is one of the major challenges in many research projects. In 2015, after replicating 100 psychological studies, Open Science Collaboration found that a large portion of the replicated results were not as strong as what were reported in the original studies, in terms of significance and magnitude. As a remedy to these shortcomings, some analysts suggested replacing the classical methods with the modern approach. Nonetheless, some traditionalists defended the merits of the classic method, such as the power of causal inferences in experimentation. Do researchers have to choose either traditional statistics or data science? This presentation aims to discuss how traditional statistical procedures and data science can work hand in hand.