though, that this belief may be false. Based on diary studies, there seems to be
no reliable correlation between a woman’s mood and her menstrual cycle, at
least when large numbers of women are studied (see Ross, 1989). The idea of
reconstruction suggests that women may use this belief about mood and men-
struation to remember inaccurately that their mood had been worse during a
previous menstruation phase than during an intermenstrual phase of the cycle.
Evidence consistent with the reconstruction hypothesis is provided by Ross
(1989). He reports a study in which a group of women was asked to keep
detailed diaries in which they recorded various life events and daily moods.
The women were not told that the study focused on the menstrual cycle. At one
point in the experiment, the women were asked to recall their mood from a day
two weeks earlier. The women were supplied the date and day of the week and
a small portion of their diary entries, including an entry that indicated whether
they were menstruating. For one group, the to-be-recalled day was during the
menstruation phase of their cycle, while for the other group the to-be-recalled
day was during the intermenstrual phase. The actual diary entries for the to-be-
recalled days indicated that the women’s mood was no worse on average dur-
ing the menstrual phase than during the intermenstrual phase. Yet the women
tended to recall that their mood was worse on the menstrual day. Moreover,
the more the women believed in a correlation between menstruation and mood
(as assessed by an attitude survey), the more likely they were to exaggerate
how negative their mood was on the day when they were menstruating.
Confidence and Accuracy
As I suggested earlier, record-keeping theories of human memory may concede
that recollection of the past often involves reconstruction. The record-keeping
theory could claim that a person resorts to reconstruction when the retrieval
process fails to locate the necessary record. The constructionist theory claims
instead that people use a reconstruction strategy every time they reflect on the
past.
The record-keeping theory implies that people should be able to tell the dif-
ference between when they are able to read a record that accurately preserves
the details of the past event, and when they are unable to locate the record and
so must resort to making guesses about the past. People would presumably
have more confidence in their memory for a past event if they are reading
the record than if they are only reconstructing it. Therefore, according to the
record-keeping theory, people’s confidence in the accuracy of their memory
for a past event should be reliably greater when the event is remembered
accurately than when an event is remembered inaccurately.
The constructionist theory, on the other hand, claims that all recollection is
reconstruction. Constructionist theory suggests that confidence and accuracy
may sometimes be related, particularly when people have developed learning
and reconstruction strategies for which they have been provided feedback as to
how well those strategies work. In such cases, people may use their knowledge
about how well a strategy has worked in the past to predict accurately how
well it will work in the future. However, constructionist theory predicts that
confidence will not be strongly related to accuracy when people have had
no opportunity to develop adequate learning and reconstructive strategies, or
Memory 339