Effect Inference from Two-Group Data with Sampling Bias
This addresses a statistical inference problem for researchers analyzing biased two-group data, offering a more reliable alternative to standard methods.
The paper tackles the problem of unreliable inference when comparing populations with sampling-biased data, developing a method that controls false positives under moderate bias levels and demonstrating it on synthetic and real biomarker data.
In many applications, different populations are compared using data that are sampled in a biased manner. Under sampling biases, standard methods that estimate the difference between the population means yield unreliable inferences. Here we develop an inference method that is resilient to sampling biases and is able to control the false positive errors under moderate bias levels in contrast to the standard approach. We demonstrate the method using synthetic and real biomarker data.