Scientists interested in climate change detection and attribution ask many questions in their analyses.  Some of these are basic: are observed trends compatible with internal variability? Is there evidence for an externally forced response in observations?  Some are more complex: does the forced response depend on the background climate state?  What is the spatiotemporal structure of climate noise, and is it changing?  Do multiple forcing agents interact with one another?  How do we select the best models to explain observed change?  I will present an adaptable, hierarchical Bayesian framework for detection and attribution and show how it can be adapted to answer a wide range of questions.  I will argue that Bayesian methods best allow analysts to track and understand how observational, structural, and parametric uncertainties contribute to our ability to detect and attribute climate change.