Abstract / Summary
"Factors associated with" studies, exploratory multivariable regression analyses that do not begin with a prespecified primary exposure, have recently come under increasing criticism, with some arguing that they are methodologically weak and contribute little to scientific knowledge. Much of this criticism is justified. Many published studies in global mental health (GMH) rely on small samples, inconsistent adjustment strategies, automated variable-selection methods, and interpretations that imply causality where none can be established. These limitations deserve attention and should not be overlooked. However, abandoning exploratory association studies altogether would ignore the realities of research in many low- and middle-income countries and could further marginalize evidence from the Global South. In global mental health, particularly in Sub-Saharan Africa, limited data systems, poorly understood causal pathways, and resource constraints often make longitudinal and experimental studies difficult to conduct. In these settings, exploratory analyses remain an important starting point for identifying patterns, generating hypotheses, and informing future research. Indeed, many important advances in epidemiology have emerged through an iterative process of observation, exploration, and subsequent causal investigation. We argue that the central issue is therefore not the use of exploratory association studies, but how they are designed, analysed, and reported. Rather than discarding them, researchers should clearly present their findings as hypothesis-generating, avoid automated stepwise variable selection, justify adjustment decisions using directed acyclic graphs (DAGs), adhere to established reporting guidelines, and interpret results within their local context. This is not a call to lower methodological standards in the Global South. Instead, it is a call to apply rigorous methods in ways that are appropriate to the realities of under-resourced settings while strengthening local capacity for causal inference. Silencing exploratory research would risk reducing the visibility of evidence from precisely those settings where stronger representation in global mental health research is most needed. Enhancing the quality of exploratory research can help advance global mental health research, particularly in settings where relevant evidence remains limited.