Purpose: This study aims to investigate the role of two structural factors – threat level depicted on fear messages and warning size – as well as two contextual factors – repeated exposure and type of packs – on pictorial and threatening tobacco warnings’ effectiveness.Design/methodology/approachA two (warning threat level: moderate vs high) × two (coverage: 40 vs 75 per cent) × two (packaging type: plain vs branded) within-subjects experiment was carried out. Subjects were exposed three times to pictorial and threatening tobacco warnings. Both self-report and psychophysiological measurements of emotion were used.FindingsResults indicate that threat level is the most effective structural factor to influence smokers’ reactions, while warning size has very low impact. Furthermore, emotional arousal, fear and disgust, as well as attitude toward tobacco brand, decrease after the second exposure to pictorial and threatening tobacco warnings, but stay stable at the third exposure. However, there is no effect of repetition on the emotional valence component, arousal-subjective component, on intention of quitting or of reducing cigarette consumption. Finally, there is a negative effect of plain packs on attitude toward tobacco brand over repeated exposures, but there is no effect of the type of packs on smokers’ emotions and intentions.Social implicationsUseful marketing social guidance, which might help government decision-makers increase the effectiveness of smoking reduction measures, is offered.Originality/valueFor the first time in this context, psychophysiological and self-report measurements were combined to measure smokers’ reactions toward pictorial and threatening tobacco warnings in a repeated exposure study.