Testing "variance in levels of alertness" can be problematic as the vary awareness one is in "a study" tends to elevate one's state of alertness; thus diluting or confounding results. If both groups were blind to the fact they are in a study, a more accurate (i.e., real life effect) might be ascertained. Also, the placebo control may well be known, at various levels, to most of the test subjects. Of course, small sample size and self report weaken results. However, "computer tasks to test levels of memory, attentiveness and vigilance" yield subjective data. Still, alertness affects all there above, and alertness can present more like a light switch (on or off) than a dimmer switch (varying gradations).