We determine the events that cause large shocks in volatility of the DJIA index over the period 1928-2010, using intervention analysis and conditional heteroscedasticity model. We use a moving subsample window to take into account the periods with high or low volatility, allowing thus to identify large shocks as extraordinary movements perceived by the investors. We show that these large shocks can be associated with particular events (financial crashes, elections, wars, monetary policies, . . . ). We show that some shocks are not identified as extraordinary movements due to their occurring during high volatility episodes, especially the 1929-1934, 1937-1938 and 2008-2010 periods.