Meta-Research
Questionable Research Practices & Correcting Bias
Research is fundamentally a decision-making process. At any stage of the research lifecycle, common design and analysis decisions can lead to suboptimal research conditions, sometimes deemed Questionable Research Practices (QRPs). Through simulated meta-analyses, we demonstrate how QRPs at the individual study level impact the accuracy of meta-analytic effect sizes, and the extent to which current bias detection techniques are (or are not) successful in correcting for these biases.
Network Replicability & Generalizability
Use of network analysis has increased in recent years. However, in light of observed inconsistencies, there is debate as to whether network properties can be expected to replicate across samples. Using a resampling approach, we explore the individual and joint effects of different sources of variability (i.e. sampling variability, scale variability, measurement error) in order to provide an overview of the (in)consistencies in network properties that can be expected to arise under varying conditions.
Predictors of Study Replicability
In the wake of the replication crisis, researchers have sought to better understand which factors of an individual study may or may not predict whether it will replicate. By applying a large-scale reanalysis of replication data sets (using k-fold cross-validation techniques), we showed that model fit positively correlated with replication success.
Improving Statistical Misconceptions
When it comes to psychological research, p-values continue to dominate the published literature. This widespread misuse of p-values (or null hypothesis significance testing) have led some to advocate for the abandonment of these statistics altogether. Via online training, we demonstrate that instructional materials can improve statistical literacy and alleviate statistical misconceptions amid online users.
Survey Research
Survey Development & Validation
Survey research can be a key tool to gain insights about individuals' beliefs, behaviors and attitudes, and hinges on the development of reliable and valid items. During my doctoral studies, I taught a graduate-level statistics course with a focus on survey construction, improvement, and validation.