Why So Much Science Doesn't Hold Up and What Quantitative Minds Can Do About It

"It's a difference of opinion that makes a horse race." Generally attributed to Mark Twain, this quote encapsulates the debate over the usefulness of research amid growing concerns in the scientific community about the replicability, reproducibility, and reliability of research findings. This article explores the debate from philosophical, statistical, and methodological perspectives, highlighting the scope of the apparent replication crisis in both natural and social sciences. The authors discuss how over-reliance on p-values and null hypothesis significance testing (NHST) has led to unreliable results and explore alternatives - such as Bayesian methods and effect size estimation. Methodological reforms including open science practices and preregistration are reviewed for their role in improving research credibility. Throughout, the authors demonstrate computational aspects of reproducibility with Python code examples, demonstrating how researchers can use programming tools to ensure their quantitative results are robust and replicable.