Those who had been asked to recall a bad deed were about 60% more likely to respond with cleansing-related words like "wash", "shower" and "soap" instead of alternatives such as "wish", "shaker" or "stop".
[1] In another experiment, experimenters were able to reduce choice-supportive bias by having subjects engage in forms of self-cleaning.
[2] The effect is apparently localized enough that those who had been asked to lie verbally preferred an oral cleaning product and those asked to lie in writing preferred a hand cleaning product over the other kind of cleanser and other control items.
[3] Other researchers have been unable to replicate the basic effect using larger samples.
[6][better source needed] A meta-analysis of 15 studies examining the relationship between primes related to moral threat and cleansing preferences found a small effect, with no significant relationship evident across 11 studies conducted by researchers other than the original ones.