Patalano,
A.L., & Seifert, C.M. (1997). Opportunistic planning: Being reminded of
pending goals. Cognitive Psychology, 34,
1-36.
This study took a look at some variables that may
possibly affect the ability to complete pending goals, based on plans using items
in the environment. Frequently the
proper resources necessary for the completion of a goal are unavailable, and
the completion of the goal must wait until discovery of the needed resources. Early research on this subject led to the
discovery of the tendency for participants to notice different types of
opportunities to fulfill a goal.
Hayes-Roth and Hayes-Roth call this “opportunistic planning.” A model by the authors and a number of
colleagues proposes that the best time to develop any type of inferences
related to planning is when the person involved in he planning process has
maximum cognitive load available. This
time is proposed by the authors to be during an initial planning phase.
This
model will be tested in three experiments.
Overall
Method
The
planning paradigm used in all three experiments is broken down into three
phases.
Goal study
phase – Ps
read the cover story and the list of goals that need to be completed. Some conditions include further plans for
completing these goals.
Reminding
phase – Ps
are presented with of the available objects that may be used to complete the
goals. Their task is to note any goal
from phase one that could be completed using the item next to the listing of
each item.
Recall phase – Ps are asked to free recall
the goals from the study phase in order to separate any type of effect
forgetting the original goals may have.
Experiment
1A
Conditions
In both the No-Memory Control and
No-Plan Control conditions the Ps were not given suggestions for the goal completion
planning. In the No-Memory Control
condition Ps were given a list of the goals so they did not have to commit the
list to memory. The No-Plan control
required the Ps to memorize the list of goals.
In the Given Plan condition, Ps were given a direct statement of an item
found in the environment and how it could be used to complete the goal (e.g.
“If only you had a chair you could stand on it to reach the elastic
band”). Ps in the Guided Plan condition
were required to generate a plan for completing the goal using the given object
(e.g. “Please write down how you could use a chair to achieve the goal of
retrieving an elastic band from a high shelf”). Each subject received a mix of
previously studied (anticipated), items similar in function to items presented
(unanticipated), and nonpresented items.
Results
and Discussion
Ps in the No-Memory condition showed
a higher rate of remindings than did those Ps in the No-Plan condition,
suggesting that keeping the goals in memory may impair recognition of the
opportunity to complete a goal. In the
two planning conditions there was a main effect of cue type for target
remindings, with less target remindings produced in response to the
unanticipated cues. There was also a
main effect for nontarget happenings based on cue type, with less remindings
produced to the anticipated cues. These
findings suggest that Ps were more likely to be reminded of the opportunity to
complete a goal when the anticipated object was presented and were less likely
to produce a nontarget response to the same anticipated cues. There was also a significant interaction for
target remindings between planning condition and cue type. Cue type had more effect in the Guided Plan
condition than the Given Plan condition.
This would suggest that stronger goal-item associations were made in the
Guided Plan condition. No simple effect
was found for unanticipated cues, but there was a significant simple effect for
anticipated cues. This suggests that the interaction is caused mainly by the
increase in the remindings for anticipated cues. Free recall for this
experiment was 8.3 out of 10 goals, on average. The findings are consistent with the Predictive Encoding model,
as associations between goals and plans at encoding increased retrieval.
Experiment
1B
All Ps in this experiment were
assigned to one condition. In this
Given Plan condition Ps were given a specific plan as in 1A, but this time
there was a delay between study and test, in which a distracter task was completed.
Results
and Discussion
Similar to 1A, a higher rate of
target goals was produced to the anticipated cues, and there were significantly
less nontarget responses to anticipated cues.
Across 1A and 1B there was a main effect of cue type, with anticipated
cues showing more target goals. Time
showed marginal significance, with increased time leading to increased target
remindings. Another marginal finding
showed n interaction between cue and delay, with the change in time happening
in response to anticipated cues (increase) and no change in response to
unanticipated cues. Recall was 8.04
goals out of 10 for this experiment.
Experiment
2
This second experiment was conducted
to test how the learned plans were stored.
Do Ps automatically abstract the plan when given a concrete plan and
item? To test this idea Ps were tested
in a condition to the Guided Condition of 1A with some additional goals and
cues. There were now two pairs of cues
for each goal, with each pair of items dealing with a separate abstraction of a
plan to complete the goal (e.g. “broom” and Hockey stick” are long objects used
to reach the elastic band and “chair” and “stack of books” are object used to
increase height). Ps were either given
the same cue object at study and test, the opposite object from the same
abstraction pair, or an item from the opposite pair.
Results
and Discussion
Each of the cue types showed a significant
difference from the other types. The
identical cues showed significantly higher target percentages than the same
abstraction cues, and the same abstraction cues showed significantly higher
target percentages than the different plan cues. No significant differences were found for nontarget percentages.
Post hoc analysis suggested that there might be some variance in remindings accounted
for by an overall difference in similarity ratings between the same abstraction
pairs. These results suggest that Ps
did not always encode the studied plans at an abstract level, rather there
seems to be a higher likelihood of concrete encoding. Free recall for this experiment was 9.70 goals out of 12.
Experiment
3
This experiment followed through
with the logic that an individual would prefer a method of encoding that
minimized the cognitive load used at encoding, while allowing a large range of
opportunistic detection. The proposal
was that a more abstract encoding without the mention of a specific object may
capture this idea. Three conditions
were used for this study, a No-Plan condition, a Given Plan condition and a
Generate Plan condition. These
conditions were the same as 1A except no specific item was mentioned in the
plan conditions (e.g. “use a long, rigid object to swipe down the elastic
band”). The Ps were instructed to construct an abstract plan in the Guided Plan
condition.
Results
and Discussion
The planing conditions showed
significant improvement from unanticipated to anticipated cues, as in 1A. There was a main effect of cue type for
target findings analogous to the same type of effect in 1A. The effect for nontargets was not
significant. No other effects were
significant. Since there was a trend on
a decrease in the target percentage on unanticipated cues for both 1A and 3,
the experimenters pooled the data for the unanticipated cues from both
experiments to test for a difference between planning conditions. The effect of planning condition was
marginally significant for target remindings.
Free recall for this experiment was 7.80 goals out of 10.
General
Discussion
The predictive encoding model is
backed by the findings of the three experiments. Potential limitations include: the minimal amount of information
presented in the study and retrieval phases of the paradigm, the differences in
the strengths of the associations between the goals and objects used in the
plans generated, and the processes involved in the plan making is not addressed
in this study.