Estimates of short-term memory capacity vary – from about 3 or 4 elements (i.e., words, digits, or letters) to about 9 elements: a commonly cited capacity is 7±2 elements. In contrast, long-term memory indefinitely stores a seemingly unlimited amount of information.
Short-term memory can be described as the capacity (or capacities) for holding a small amount of information in mind in an active, readily available state. The information held in short-term memory may be:
A classical model of memory developed in the 1960s assumed that all memories pass from a short-term to a long-term store after a small period of time. This model is referred to as the "modal model" and has been most famously detailed by Atkinson and Shiffrin. The exact mechanisms by which this transfer takes place, whether all or only some memories are retained permanently, and indeed the existence of a genuine distinction between the two stores, remain controversial topics among experts.
One form of evidence, cited in favor of the separate existence of a short-term store comes from anterograde amnesia, the inability to learn new facts and episodes. Patients with this form of amnesia, typically caused by damage to the hippocampus, have intact ability to retain small amounts of information over short time scales (up to 30 seconds) but are dramatically impaired in their ability to form longer-term memories (a famous example is patient HM). This is interpreted as showing that the short-term store is spared from amnesia.
Other evidence comes from experimental studies showing that some manipulations (e.g., a distractor task, such as repeatedly subtracting a single-digit number from a larger number following learning) impair memory for the 3 to 5 most recently learned words of a list (presumably still held in short-term memory), while leaving recall for words from earlier in the list (presumably stored in long-term memory) unaffected; other manipulations (e.g., semantic similarity of the words) affect only memory for earlier list words, but do not affect memory for the last few words in a list. These results show that different factors affect short term recall (disruption of rehearsal) and long-term recall (semantic similarity). Together, these findings show that long-term memory and short-term memory can vary independently of each other.
Not all researchers agree that short-term and long-term memory are separate systems. Some theoriests propose that memory is unitary over all time scales, from milliseconds to years. Support for the unitary memory hypothesis comes from the fact that it has been difficult to demarcate a clear boundary between short-term and long-term memory. For instance, Tarnow shows that the recall probability vs. latency curve is a straight line from 6 to 600 seconds, with the probability of failure to recall only saturating after 600 seconds . If there were really two different memory stores operating in this time frame, one could expect a discontinuity in this curve. Other research has shown that the detailed pattern of recall errors looks remarkably similar for recall of a list immediately after learning (presumably from short-term memory) and recall after 24 hours (necessarily from long-term memory)
The relationship between short-term memory and working memory is differently described by various theorists, but it is generally acknowledged that the two concepts are distinct. Working memory is a theoretical framework that refers to structures and processes used for temporarily storing and manipulating information. As such, working memory might also be referred to as working attention. Short-term memory generally refers in a theory-neutral manner to the short term storage of information. Thus while there are short-term memory components to working memory models, the concept of short-term memory is distinct from these more hypothetical concepts. Within Baddeley's influential 1986 model of working memory there are two short-term storage mechanisms: the phonological loop and the visuospatial sketchpad. Most of the research referred to here involves the phonological loop, because most of the work done on short-term memory has used verbal material. In recent years, however, there has been a surge in research on visual short term memory, and also increasing work on spatial short term memory
The most important characteristic of a short-term store is, clearly, that it is short-term — that is, it retains information for a limited amount of time only. Most definitions of short-term memory limit the duration of storage to less than a minute; no more than about 30 seconds, and in some models as little as 2 seconds. Memory that exceeds short-term memory duration limits is known as long-term memory.
The limited duration of short-term memory immediately suggests that its contents spontaneously decay over time. The decay assumption is part of many theories of short-term memory, most notably Baddeley's model of working memory. The decay assumption is usually paired with the idea of rapid covert rehearsal: In order to overcome the limitation of short-term memory, and retain information for longer, information must be periodically repeated, or rehearsed — either by articulating it out loud, or by mentally simulating such articulation. In this way, the information will re-enter the short-term store and be retained for a further period.
Several researchers, however, dispute that spontaneous decay plays any significant role in forgetting over the short term , and the evidence is far from conclusive.
Authors doubting that decay causes forgetting from short-term memory often offer as an alternative some form of interference: When several elements (such as digits, words, or pictures) are held in short term memory simultaneously, their representations compete with each other for recall, or degrade each other. Thereby, new content gradually pushes out older content, unless the older content is actively protected against interference by rehearsal or by directing attention to it.
Whatever the cause or causes of forgetting over the short term may be, there is consensus that it severely limits the amount of new information that we can retain over brief periods of time. This limit is referred to as the finite capacity of short-term memory. The capacity of short-term memory is often called "memory span", in reference to a common procedure of measuring it: The experimenter presents lists of items (e.g., digits or words) of increasing length; the tested person's span is determined as the longest list length that she or he can recall correctly in the given order on at least half of all trials.
Prior to the creation of current memory models, George Miller argued that human short-term memory has a forward memory span of approximately seven items plus or minus two. More recent research has shown that this magical number seven is roughly accurate for college students recalling lists of digits, but memory span varies widely with populations tested and with material. For example, the ability to recall words in order depends on a number of characteristics of these words: fewer words can be recalled when the words have longer spoken duration; this is known as the word-length effect, or when their speech sounds are similar to each other; this is called the phonological similarity effect. More words can be recalled when the words are highly familiar and/or occur frequently in the language; recall performance is also better when all of the words in a list are taken from a single semantic category (such as sports) than when the words are taken from different categories.
Though the average person may only retain about 7±2 different units in his or her short term memory, chunking can greatly increase a person's recall ability. Through putting each unit into a meaningful word or phrase, a person's recall ability can skyrocket through practice. For example, in recalling a phone number, the person usually chunks the digits into three groups: first, the area code (such as 814), then a three-digit chunk (123) and lastly a four-digit chunk (4567). This method of remembering phone numbers is far more effective than attempting to remember a string of 10 digits. In one testing session, an All-American cross-country runner was able to recall a string of 73 digits after hearing them only once by chunking them into different running times (e.g. the first four numbers were 1518, a three-mile time.)