Psychological research, especially empirical research, should be based on theoretical thinking, especially theoretical derivation. Research questions need to be raised for theoretical derivation. Such research problems can be directional and operational. Directional problems have insight and guiding significance, and can usually lead to operational problems. It is the operational problem that corresponds to the specific research, including empirical test.

Research by the speech psychologists Peterson and his wife shows that people will quickly forget the information they receive. After 9 seconds, they will forget about 30%, and after 15 seconds, they can only remember 10%. Based on this, the researchers put forward the directional question: why does short-term memory forget so quickly? If the memory disappears in more than ten seconds or even a few seconds, where is the memory? Or what happened to the forgetting process?

There are two possibilities for the forgetting process of short-term memory. First, memory traces fade and even disappear with the passage of time. Second, information is excluded from each other, and memory is disturbed and unavailable. If it is the first case, then time is the reason for forgetting. In the second case, interference is the main cause of forgetting. Time factor also plays a role in long-term memory, needless to say. If it is an interference factor, and if the capacity of short-term memory is very small, there are two cases: one is retroactive interference, that is, new information comes in and replaces existing information; The second is proactive interference, that is, the existing information confuses the new information, and the new information cannot come in effectively.

Is time or interference at work? From the perspective of experimental psychology, the solution is to keep one constant and change the other. Therefore, researchers can put forward the operational problem - changing the presentation rate. If the word list is presented at different rates, it means that under the condition of slow presentation, the items in the word list have more time decay when the recall test is carried out. Therefore, if the forgetting of short-term memory depends on time, the memory performance under the condition of slow presentation of word lists is worse.

Many researchers use this idea for experiments. For example, in the study, Bennett Murdock presented 19 subjects with a word list composed of 20 unrelated words at the rate of one or two per second. After presentation, give the subjects 1.5 minutes to write as many words as they can remember, regardless of order. After 5 to 10 seconds, present a new word list and repeat the process.

Modoc drew the experimental results in the rectangular coordinate system, the abscissa represents the items (the first item in the word list is on the far left and the last item is on the far right), and the ordinate represents the memory score. Connect the points representing the memory performance of each item in order to form a curve. Since there are two presentation rates, there are two curves.

For both curves, the memory scores of the items at the beginning of the word list are very high, and then decrease. For the last 7 or 8 items, the memory scores rise again - the score of the last item is close to 100%, and the score of the penultimate item is about 90%. In particular, for each curve, the slope at the corresponding position is the same, that is, the forgetting situation under the two presentation rates is the same.

Why did this happen? Modoc believes that when subjects are asked to recall words, their strategy is to recall items in short-term memory first. Because short-term memory can only maintain the last few items, the recall performance of the last few items is better. In other words, the subjects started from the end of the word list and recalled backwards as soon as possible. This idea is easy to test, because subjects are allowed to write words in any order. Therefore, the experimental data contains the order information of subjects' recall of words. The results show that, indeed, the first words written by the subjects are often the last words in the word list.

It is speculated that the rise of the right tail of each curve reflects short-term memory. Can this idea be tested? Certainly. In other experiments, two conditions were set. One condition was to allow subjects to write down the words they remembered at the end of the word list presentation. Another condition is to insert a 20 second delay at the end of the word list to prevent the subjects from repeating the last few items - just use the countdown task of 3 created by Mr. and Mrs. Peterson. If this procedure eliminates the short-term memory, and if the short-term memory corresponds to the last rising part of each curve, the last rising part will disappear under the condition of 20 second delay.

This is the result of the experiment. When the recall is delayed for 20 seconds, the rise of the right tail of the curve disappears. At the same time, the rising part on the left of the curve was not affected, because that part reflected the long-term memory of the items in front of the word list.

So far, we can answer the previous question: is time or interference at work in the forgetting of short-term memory?

It is assumed that the forgetting of short-term memory depends on time. If 20 items are presented at the rate of 1 per second, all word lists are presented. When testing, 1 second has passed for the last item, 2 seconds for the penultimate item, and so on. If it is presented at the rate of 2 seconds and 1 item, the time of all items will be doubled, the presentation of all word lists will be completed, the last item has passed 2 seconds, the penultimate item has passed 4 seconds, and so on. That is, under the condition of slow rendering, each item has twice the time to forget. Therefore, if forgetting depends on time, the performance of slow condition will be worse. However, the experimental results show that this does not happen. For the items at the back of the word list, short-term memory is dominant, and there is no difference between the two presentation rates.

Similar studies also show that the forgetting of short-term memory is mainly due to the interference of other information rather than the disappearance of new information. This reflects the limitations of how much information people can keep, not how long they can keep.

In short, the exploration of forgetting mechanism of short-term memory typically shows the characteristics of psychological research: Based on existing research, carry out theoretical thinking and theoretical derivation, put forward directional and operational problems, design experiments, carry out fine empirical tests, obtain new opinions, and constantly promote the development of relevant fields.