Source monitoring helps determine the origin of a memory. Qualitative characteristics such as perceptual details, affective information and cognitive operations are retrieved. Details are weighed and if possible, a source judgment will be made. This can happen quickly or slowly, depending on the situation.

This study examines "the degree to which source monitoring requires access to recollective details." According to some researchers, because recollective details are made up of qualitative characteristics, recollective details must be required for source monitoring.

But some researchers think partial information can contribute to source monitoring and that successful source monitoring can be aided by either very weak info or actual recollective details. In other words, a lot of help (recollection) or very little help (familiarity) may be used to make the decision about the source of a memory.

Rather than using ROC curves, these researchers used remember/know judgments to look at the phenomenological experiences of the subjects during source monitoring. Therefore, if a subject indicated remember, that subject was reporting recollection, however if a subject indicated only knowing, that subject was reporting familiarity.

Two previous studies have used remember/know judgments to look at what is needed for successful source monitoring and the studies produced mixed results. Why? Perhaps subjects were able to use strong and weak sources against each other to make decisions. Therefore, these researchers used sources that were equally detectible in experiment 1 (these sources yielded equal old-new detection rates). If subjects replied know to correctly attributed source judgments, than it is possible to successfully source monitor without recollective details.

Experiment 1

Method

One hundred twenty medium frequency words were used. The participants read 40 words off of a computer screen, listened to 40 words read aloud by the experimenter and 40 new words were presented later at test time. The words were randomly ordered so that they could be read, heard or new for each participant. The subjects read and heard the first 80 words at a rate of 1 every 3 seconds. The subjects were informed to learn the words for a memory test. During the memory test, the subjects were to state whether the word was seen, heard or new. If they reported the word as seen or heard, they were asked to indicate remember or know.

Results

Items were source monitored equally well whether they had been seen or heard by subjects. The experimenters went on to look at whether subjects remembered or knew items that the subjects correctly attributed to a source. There were no differences in subjects choosing remember or know for the seen words or heard words. Also, if subjects chose a new word as old and made a source judgment about that word, there were no differences in whether or not they chose remember or know, showing that recollective details do not guarantee correct source attributions.

Experiment 2

Method

This experiment was identical to experiment 1 except that the heard words were anagrams instead. This made the seen words and anagram words different instead of the same. In other words, whereas the words in experiment 1 were equally likely to be attributed to the correct source, the words used here were either generated (anagrams) or only seen, making them unequally likely to be attributed to a correct source. It was expected that the generated words would be easier to attribute correctly to a source because they would be more deeply encoded.

Results

As expected, correct attribution to the source was higher for the generated words. Furthermore, the subjects reported more remember responses for these words than know responses. But if the subject had only seen the word, the subjects tended to report a know response instead of a remember response. The experimenters think this is due to the fact that both words were actually seen and not just heard as half the studied words were in experiment 1, therefore vaguer details were able to be used to attribute the word to a source. If a subject made a mistake and labeled a word old when it was new, they were more likely to say that the word was seen and not generated, and the subject was more likely to say know than remember.

General Discussion

The researchers believe that these results show that familiarity is useful in source monitoring and that vague information can still result in a correct source attribution. They think that early theories on source monitoring have been overshadowed by research from the 1990’s, including Jacoby’s research. The researchers try to convince readers that older theories about source monitoring are still viable and may explain the findings of this report.