Despite massive and increasing public spendig in training for the unemployed, little is known about how job seekers decide to enroll in a training program. Decisions related to job training might be undermined by informational gaps, especially about program costs,enrollment procedures,and expectations of reemployment chances. This paper reports the results of a low-cost intervention aimed at testing for the existence of misinformation about training costs and returns, and its impacton enrollment. Partnering with the FrenchPublic Employment Services and the largest training provider in France, we sent 50,000 emails advertising training opportunities to job seekers in four regions of France in late summer 2016. We randomly added short messages on training costs, registration procedures, and training returns to the basic email template. We find that receiving a nemail with a message emphasizing training returns in terms of employment more than doubles the likelihood that job seekers call back the training center. However, callback rates are low in absolut evalue (less than one percent) and we detect no impact on enrollment one to six months after the intervention. We provide suggestive evidence that the effects on callbacks are driven by increasing salience of basici nformation about training rather than by belief updating.