Abstract or Summary:
Several recent studies have attempted to lay the groundwork for a simple, low level computer security system involving cognitive and word association passwords. The present research investigated the effect of three different word association formulation techniques on recall rates when participants were presented with their original cue words after a 2 week retention interval. Differences in the susceptibility of word associations to being guessed by someone the participant knew well were also examined. In the ‘response only’ group (n = 24), respondents were provided with a list of 20 cue words and were required to generate an associated response for each cue. In the ‘cues and responses’ group (n = 25), respondents generated both cues and associated responses, while in the ‘theme’ group (n = 24), respondents generated both cue and response words having first decided upon a theme for their word associations. The average recall and guessing rates were 66% and 8%, respectively. No statistically significant differences in recall or guessing rates were found between the three groups. The effect sizes obtained strongly suggest that insufficient statistical power was available to produce significant effects. Nevertheless, it is tentatively concluded that the formulation techniques employed have only a small effect on ease of recall. Future research on the utility of word associations as computer passwords needs to be based in a real computer security setting.