Inhibition of Return (IOR) is a mechanism whereby the attentional system favors novel locations by inhibiting already scanned ones. In spatial attention tasks, it commonly occurs when the interval between cue onset and target onset is longer than 300 ms. The positive difference between reactions in the valid condition and those in the invalid one shows that responses to target stimuli are slower following a valid cue than responses to target stimuli following an invalid cue. IOR is a very robust phenomenon at the group mean level; however, this study demonstrates that its standard error of measurement is extremely high, which seriously challenges any attempt to interpret an individual score as representing the characteristics of a subject's attention system. Furthermore, this reliability problem might diminish the likelihood of finding differences between groups and conditions. The study shows that these problems may be partially corrected by employing the back-to-center paradigm.