About me (View CV)

My PhD journey is driven by a curiosity about how we can enhance research replicability. My work is dedicated to improving the evaluation and planning of close replications, as well as assessing the credibility of psychological findings. As an aspiring early-career researcher, my ultimate goal is to pioneer methods for guiding science to make our research more robust and reproducible.

My doctoral work introduces a new replication success classification framework for multisite direct replication studies (e.g., the ManyLab replication projects) that goes beyond the traditional binary distinction of replication success and failure by distinguishing between true and false replication successes and failures. My work reveals how bias (p-hacking and publication bias) and study-level factors (e.g., statistical power) dynamically shape the replication outcomes. I developed a Shiny app to visualize this framework.

I also work on methodological issues in meta-analytic methods, with a particular focus on the meta-analysis of proportions. My comprehensive tutorial in this area has become a widely referenced resource in fields such as epidemiology, medicine, public health, and more. I collaborate with researchers from diverse disiplines to conduct systematic reviews and meta-analyses. I have developed a research tool for systematic review called CoviPanel. It is a ChatGPT-assisted browser extension built for facilitating title and abstract screening in Covidence.

My expertise encompasses a broad spectrum of statistical techniques and tools, including Monte Carlo simulations, high-performance computing, Shiny App development, Bayesian analysis, mixed-effects modeling, among others. With nearly a decade of experience in statistical programming using R, I am passionate about educating the public in R programming for the greater good. You can view my video tutorial to see a glimpse of my teaching approach in R, available here.

One of the coolest computational skills I’ve picked up for research is high-performance computing (HPC). I run large-scale simulations on HPC servers in the university or on cloud-based platforms, like AWS EC2. Using HPC, simulations that would’ve taken months on a regular computer now wrap up in just a week or even days! This kind of power opens the door to more ambitious, complex, and exciting research questions.