Several statistical issues that arise in the construction and interpretation of measures of uncertainty from forecast surveys that include probability questions are considered, with application to the Bank of England Survey of External Forecasters. Substantial heterogeneity of individual forecasters' uncertainty is found, together with significant persistence in their relative uncertainty, which is a new finding in professional forecast surveys. It is an individual characteristic akin to the individual optimism and pessimism already established in the literature on point forecasts; the latter is also found in the current dataset, now in a bivariate sense with respect to joint inflation and output growth point forecasts. Whether disagreement among point forecasts is a useful indicator of uncertainty is shown to depend on the underlying macroeconomic environment. Copyright © 2014 John Wiley & Sons, Ltd.