the robber from a lineup. Some subjects were given photographs (not contain-
ing the robber) taken from the scene of the crime, or were asked to think back
through the events from beginning to end while imagining the robbery. These
subjects tended to identify the robber more accurately than subjects not given
any context-reinstating cues.
Other experiments have demonstrated that memory for an event is more
accurate if retrieval takes place in the same physical environment as the one
where the event originally occurred (e.g., Canas & Nelson, 1986; see Smith,
1988 for a review). In one of my favorite studies, subjects who learned a list of
words while scuba diving later recalled more of the words if the recall test took
place while the subjects were again scuba diving than if the recall test took
place on land (Godden & Baddeley, 1975).
It should be noted, though, that the overlap of physical environments is
probably an important determinant of memory when the to-be-learned infor-
mation can be associated with the physical environment (Baddeley, 1982; Fer-
nandez & Glenberg, 1985). An eyewitness may be more likely to remember the
events of an accident, such as a car crashing into a tree, if the eyewitness rec-
ollects at the scene of the accident, than if the eyewitness recollects in the police
station. The tree at the crash site is associated with the accident, so that seeing
thetreeislikelytoactivateinformationthatmaybeusedtoreconstructthe
accident. On the other hand, it is probably not as important that a student take
anexaminthesameroomwhereheorshestudiedfortheexam(Saufley,
Otaka, & Bavaresco, 1985) since academic information would not ordinarily be
associated with the physical elements of a room. Much more important is that
the student understand the academic material, organize the material, and make
the details contained within the material distinctive.
One demonstration that the overlap principle depends more on the similarity
of cognitive processing than on similarity of physical stimuli comes from re-
search on mood. The usual finding is that subjects induced to feel elated or
depressed will more likely and quickly recall past events experienced in the
same mood, than those experienced in a different mood (Snyder & White, 1982;
Teasdale & Fogarty, 1979; see Blaney, 1986, for a review). In experiments con-
ducted by Eich (1995), subjects were placed in a setting (e.g., a laboratory) and
then responded to a list of 16 words designed to prompt memories of past
experiences. The subjects ’mood was also measured. Later, subjects were placed
in either the same physical setting or a different setting, and were induced to
feel either happy or sad. Mood was induced by having subjects listen to either
joyful musical pieces while entertaining elating thoughts or melancholy musi-
cal pieces while entertaining depressing thoughts. The subjects then had to re-
call the 16 prompt words and the events the prompts elicited. Recall was better
when the mood at the time of recall matched the mood experienced when the
prompt words were first presented. Overlap in physical setting, on the other
hand, did not matter to recall.
Problem Solving and the Overlap Principle Another interesting demonstration
that the overlap principle is based on similarity in the way events are pro-
cessed, and not on the mere presence of overlapping stimulus cues, comes from
research on problem solving. A seemingly perplexing finding of this research is
342 R. Kim Guenther