Having a sound safety climate is seen by many as one of the predominant factors determining the safety performance (Hudson, 2007). Since the Baker-panel investigation into the cultural aspects of the Texas city refinery disaster showed how a poor safety culture can contribute to an explosion, this relation has become almost'self evident' (Baker, 2007). However, changing the safety climate is a very complex matter: it is not always clear which aspect of an improvement program has a positive effect on actual safety statistics (Parker, Mearns and Hudson, 2004). Generating a feeling of'being in a safe environment' may be one of the contributing factors: "Perception survey results showing that employees do not feel safe at work, for example, are a leading indicator that an injury may occur in the future" (Parker, 2006). The reverse is not necessarily true: feeling safe is no indicator at all, that injuries will not happen in the future. This study will describe one of the popular ways of giving employees a feeling of safety and its effects on safety statistics: a strong leadership setting so called SMART goals about safety performance, preferably'zero accidents'. SMART is an acronym with an unclear history, although some attribute the term to Drucker (1954). Although many versions of SMART exist, most commonly the S stand for Specific, the M for Measurable, the A for Attainable, the R for Relevant and the T for Time-bound. This paper will explore the evidence supporting a relation between the safety climate of a firm and workers' self-reports of injury, non lost-time injury reports as recorded by the Workplace Safety & Insurance Board in Canada and lost-time injury rates as recorded by the Workplace Safety & Insurance Board, non-unionized or unionized status, the role SMART goal setting might play in improving safety performance, and self-reports of working safely. 2 SPE 127152