A growing empirical literature on how to measure uncertainty has emerged following the 2007-2008 financial crisis. Using a principal component analysis that includes the various measures of uncertainty provided by the literature, we develop a monthly global measure of uncertainty for the United States on the period 1990-2015 and we determine the factors explaining fluctuations in uncertainty. Beyond the general level of uncertainty we also identify a second factor which indicates the nature of the uncertainty, financial or non-financial. We investigate the impact of uncertainty shocks using local projection methods and our general measure of uncertainty. We find a significant negative impact of the general measure on economic activity. However, we obtain a positive and significant effect of the second dimension of uncertainty when the general level of uncertainty is high.