Marsh, R. L., Hicks, J. L., Hancock, T. W., & Munsayac, K. (2002). Investigating the
output monitoring component of event-based prospective memory performance. Memory & Cognition, 30, 302-311.
The main point of this article is that prospective memory can depend on what people believe to be true about their past performance, which is called output monitoring. Thus, retrospective memory is a component of prospective memory performance. One can believe that he or she performed an activity or intention without having actually done so, an error in prospective memory that is not normally studied.
To build on the earlier work of Einstein, Holland, McDaniel, & Guynn, the authors created an event-based prospective memory paradigm that required an output monitoring component.
Experiment1
In experiment 1, participants made pleasantness ratings on a series of words. When they encountered an animal word, they were instructed to press a particular key (the first key) and then make the rating. If they saw the word a second time and remembered pressing the key, they were asked to press another key (the repeat key). This simple condition was accompanied by an elaborated condition where participants were told to say the animal word aloud when they saw it.
Results
There was a 100 and 200 trial lag, which resulted in no differences, and will not be discussed further. Participants generally remember making an event-based response, and this is better for the elaborated participants. When a response was missed, participants behaved as if they had performed the action earlier whether it was or not.
Experiment 2
To reduce the ambiguity of participants’ responses to repeated items, any time a participant pressed the first key, they were asked “Did this word occur before”. This was included to determine if participants were remembering items and knowing that they did not respond to them, or if they just did not remember being presented with the word.
Results
Results indicate that participants pressed the first key because they did not think that they had responded earlier (very few of the first key responses indicated that participants had forgotten the presentation of the target). The authors conclude that with these results, it can be assumed that first key responses generally indicate the forgetting of the prospective response when it was actually made or the correct identification that they forgot to make the response (not just forgetting that the item was presented).
Experiment 3
This experiment was designed to test context effects. In the control condition, participants either made pleasantness ratings OR imageability ratings to the items while performing the prospective memory task. In the experimental condition, participants completed 150 trials with one context, had a break, and completed the next 150 trials with the other context.
Results
The contextual change increased the forgetting of having made a successful response previously, suggesting that a change in context could increase the probability of a repetition error. The authors note that the consistent aspect of the data of all three experiments so far suggest that people might believe their prospective memory performance was better than it really was.
Experiment 4
In this experiment, participants in the experimental condition pressed the first letter of the animals’ names for the first key response, and then pressed the repeat key for every item that was repeated (this was called the distinctive condition). A simple, control condition was also used.
Results
Performance in the simple condition matched performance in the previous experiments. The pressing of different keys, however, did not change performance, indicating that the distinctive information did not aid participants in remembering that a previous action had been completed. However, for repeated items that had been missed originally, correct identifications that this was the first response exceeded claims of erroneous previous performance, and this is a result of a reduced error rate in the distinctive condition. Thus, these results suggest that the retrospective memory component associated with output monitoring for previously missed intentions can be improved.
In conclusion, prospective memory failures happen not only because of an intention that has gone unfulfilled but because of an erroneous belief of fulfilled intentions (thinking you did it when you didn’t, or thinking you didn’t do it when you did). This could be harmful in real-world instances, such as not turning off the iron because you think that you did, or taking your medicine again because you don’t think that you did it.