This paper explores safety climate as a phenomenon. Comparative analyses of the survey designed and distributed among European universities with a study published in 2013 demonstrating that safety climate studies targeting respondents with supervisory positions are not representative and create informational and perception gaps. Higher laboratory safety perception, management commitment and awareness were typical for respondents with supervisory positions. In the second part, hypotheses developed during literature research were tested. Laboratory safety and safety rules perception strongly depend on the management's commitment. The latter is strongly related to the time and quality of the respondents' safety training. We propose some modifications: inclusion of safety behaviour into individual performance assessment, formalisation of safety, reprioritisation of teaching and research objectives. We assume that mentioned improvements will ease the safety leadership in academia.