Since the original experiments on choice blindness for faces (Johansson et al 2005), many studies have extended the phenomenon to other domains but few have focused on attaining a deeper understanding of the phenomenon itself. We here report on an experiment which closely follows the original study (albeit in computerized form and with a Singaporean population), with additional elements aimed at explicating the causal structure and temporal dynamics of the underlying processes. Contrary to intuition, we find that subjects who notice and immediately report a “manipulation” trial still miss about half of subsequent manipulation trials. Detection is also found to be highly sensitive to the form and timing of the opportunities to report. We also investigate the effects of choice manipulation on attractiveness ratings, finding that manipulations do modulate subsequent attractiveness ratings but that this effect falls below significance after a 2-week interval.