Gesine Dreisbach and Karl-Heinz Bäuml from Regensburg University first instilled new habits in their participants by presenting them with German words and training them over many trials to make the same response to each word - a left-handed key-press for half of them, a right-hand response for the remainder.
Later, participants had to categorise the same words by gender, with key-presses again used to make the categorisations. Crucially, half the words called for the same key-presses as had been trained for those words earlier, whereas the others required a key press that was the opposite to the earlier training. Reaction time differences between matched and mismatched trials tell us how much the earlier learned habit interfered with the current task (a similar philosophy to the well-known Stroop test).
Half the participants were run normally through this process and showed interference in the gender task - making key responses that were against the grain of the earlier training slowed them down, as you'd expect. The other half of the participants, once they'd completed the initial training, were confronted with an apparent computer crash and an apologetic experimenter told them to forget all about what they'd done so far. This group weren't held back by habits on the later task: in fact, interference from the earlier training was totally eliminated.
Click on the link below to read the full article
Source material from British Psychological Society