Robust test statistics for data sets with missing correlation information

Not all experiments publish their results with a description of the correlations between data points. This makes it difficult to do hypothesis tests or model fits that data, since just assuming no correlation can lead an overestimation underestimation resulting uncertainties. work presents robust test statistics be used datasets missing information. They are exact in case and either guaranteed conservative---i.e., uncertainty is never underestimated---in presence correlations, they also degenerate perfect