Backward masking is also a term found in psychological experiments that deal with visual processing. It is a phenomenon that occurs when items already stored in visual memory are interfered with (or even obscured by) newly added stimulus; they are masked. The term was later "appropriated" by the religious right to describe hidden messages -- chosen because there were already mentions of it in scholarly journals. It was hoped that the right's snake-oil arguments would appear (upon brief examination) to have scientific reasoning behind them.

Backward masking was shown in experiments by Averbach & Coriell (1961/1973). They used a tachistoscope -- a device capable of displaying a given frame or slide for any amount of time (all the way down to milliseconds) -- to flash sequences of letters for the subject. In the control condition, the letters were flashed for 50 ms, a bar was flashed under one of them for 50 ms afterward, and the subjects were asked to report the underlined letter. In the experimental conditions, an empty or filled circle that would have surrounded the letters was flashed instead of a bar, and the same thing was asked. Results from the circle conditions were dramatically worse than those from the bar; it seemed that the circle had erased memory of the given letter.

Recent findings by Vince DiLollo (1999) suggest that if the subject isn't looking for an identifier mark, the backward masking effect doesn't occur. That is, if a set of letters is shown followed by a mask, but the mask isn't mentioned by the experimenter, recall of the letters is unaffected. This is important because it shows backward masking to be a cognitive effect (top down) rather than one based on visual processing alone (bottom up).