The affectively valenced scripts used by S. Tiffany (1990) suggest that different scripts produce relatively equivalent levels of cue reactivity, although it is unclear if these laboratory findings generalize to clinical samples. In this study, cessation-motivated smokers were tested 7 days before they tried to quit smoking and were exposed to 3 audiotaped scripts that depicted different affectively valenced situations (neutral, positive, or negative). The latter 2 scripts also contained smoking cues. The findings using a clinical sample differed considerably from those using analogue laboratory samples across affective, cognitive, and physiological response measures. Reactivity to these standardized scripts failed to predict treatment outcome through a 30-day follow-up. The use of affectively valenced scripts beyond a laboratory sample is questioned.