Set honest limits on claims

Judge whether a dataset can support a description, comparison, prediction, or causal claim. Use its variables, row unit, timing, coverage, and measurement limits to decide what it can and cannot honestly say.