Laboratory studies of analogical reasoning have shown that subjects are mos
tly influenced by superficial similarity in the retrieval of source analogs
. However, real-world investigations have demonstrated that people generate
analogies using deep structural features. We conducted three experiments t
o determine why laboratory and real-world studies have yielded different re
sults. In the first two experiments, we used a "production paradigm" in whi
ch subjects were asked to generate sources for a given target. Results show
that the majority of the analogies that were generated displayed low level
s of superficial similarity with the target problem. Moreover, most of the
analogies were based on complex underlying structures. The third experiment
used a "reception paradigm" methodology. The subjects had to retrieve pred
etermined sources instead of generate their own. In this case, retrieval wa
s largely constrained by surface similarity. We conclude that people san us
e structural relations when given an appropriate task and that this ability
has been underestimated in previous research on analogy.