Overview

Minimizing the memory requirements and cognitive load for a user are essential for creating usable systems. Psychological theory forms the base for studying the limits of memory. We describe several of those theories here.

Miller's "chunking" concept describes the capacity of short term memory (1956). This theory states that a person can remember seven plus or minus two items in their short term memory, where an item can be a number, a face, a word, or any unit. By grouping similar items into a collection, called a 'chunk', short term memory is expanded. One can store seven plus or minus two meaningful chunks instead of individual items.

Bartlett (1932, 1958) approaches the issue of memory in a similar fashion by proposing the concept of schemata. He arrived at the concept from studies of memory he conducted in which subjects recalled details of stories that were not actually there. From this he concluded that people must create a mental model or structure that they use as an aide for remembering.

The concept of schemata have been studied much further, including research by Mandler (1984), Rumelhart (1980) ,  Bransford & Franks (1971), and  Quinn & Holland (1987).  One interesting extension on the schemata research into learning addresses novice versus expert performance (Chi et al., 1988). This work suggests that the nature of expertise is primarily due to an expert having a set of schemas that guide perception and problem-solving which novices do not have. Chi argues that the increased performance of experts demonstrates that people develop new schemata in their long term memory through learning. Psychological studies demonstrate this principle by tracking the improvement from inefficient, slow, and frustrating to fast, quick, and efficient. The change in performance occurs because as the learner becomes increasingly familiar with the material, the cognitive characteristics associated with the material are altered so that it can be handled more efficiently by working memory (Soloman).

Sweller's Cognitive Load Theory combine's Miller's work with the schemata theory. Recognizing George Miller's research showing that short term memory is limited in the number of elements it can contain simultaneously, Sweller builds a theory that treats schemata, or combinations of elements, as the cognitive structures that make up an individual's knowledge base. Essentially, schemata become chunks for expanding memory (Sweller, 1988).

Peterson & Peterson (1959) performed psychological studies to determine the duration, or volatility, of short term memory. Their research shows that short term memory is very volatile. In the presence of distractions, users have difficulty remembering even three elements after eighteen seconds. Further research has shown that the duration of short term memory may be much shorter, decaying after about 2 seconds (Marsh et al 1997). Because of this volatility, it is not only important that users have structures to aide in remembering information, but that they are not required to remember it for an extended period of time.

Scope/Application:

Information processing theory has become a general theory of human cognition; the phenomenon of chunking has been verified at all levels of cognitive processing.

Sweller's theories are best applied in the area of instructional design of cognitively complex or technically challenging material. His concentration is on the reasons that people have difficulty learning material of this nature. Cognitive load theory has many implications in the design of learning materials which must, if they are to be effective, keep cognitive load of learners at a minimum during the learning process. While in the past the theory has been applied primarily to technical areas, it is now being applied to more language-based discursive areas.

The theories of the volatility of short term memory are still actively being studied and refined. It is clear, though, from all of these studies, that in the presence of any distractions, subjects short term memory will quickly start to decay.

Example:

Users of command line editors such as emacs or vi are often faced with a distractions that may make it difficult to remember task-specific information. To perform a single command, users will often need to enter two pairs of keystrokes. These key strokes are not provided on the screen, so for non-experts, recalling the proper series of keystrokes is a significant distraction from the task at hand. New users who may need look up commands with which they are unfamiliar will clearly have a difficult time remembering any task information in short term memory.

Reliance on familiar schemata is very visible in interface design. For example, many users have a schema that indicates when a popup dialogue box appears, hitting "Return" or "Enter" will confirm the action, while hitting the "Esc" key cancels the action and dismisses the dialogue box.

Both of the above menus are examples. Since users rely on this, they are able to skip the process of reading each menu, thus avoiding distractions which would interrupt the flow of their work. If this schema is violated, say by making the "enter" key activate the "cancel" option in the window,  may cause the user to inadvertently dismiss a dialogue. If this happens in a critical case, say saving a document on program exit, it could cause serious problems for the user.

Principles:

Memory principles:

1. Short term memory is limited to seven plus or minus two chunks of information.
2. Short term memory is volatile, and users will often forget in the presence of distractions
3. A schema is a mental model that makes it easier for users to recall an item. Schemas can serve as the basis for "chunks" because the provide a meaningful method for grouping information
4. Well developed schemata make it easier to remember items that fit within a schema. Thus, experts with well developed schemata outperform novices.

Design Principles:

1. Design minor messages, alerts, and warnings to be minimally disruptive. For example, if a user tries to select something they are not allowed to, use a small beep or tone instead of a pop up alert menu. This will prevent the user from forgetting the data stored in short term memory.
2. Use familiar structures in application design to conform to schemata with which users are familiar to make learning easier.
3. Design interfaces that provide on-screen references for commands

Applicability to HCI

These theories of memory can be tremendously useful to us in HCI, both in design and testing.

The schema theory tells us that users remember based on models (schemata) that they have developed. As a guide in design, this indicates that users will have an easier time learning to use a piece of software if it relies on familiar design schemata. By implementing new software using these familiar principles, the user will be able to direct more attention to learning the new parts, and not be distracted or cognitively burdened by the more familiar interface functions. Some examples of this type of familiar design schemata include the following:

• Consistency with standards - using common keyboard shortcuts for their common purpose, i.e. ctrl-C for copying.
• Dialogue Boxes with familiar options, layout, and commands
• Familiar positioning of menu bars, placement of specific menus within the bar, and organization of menu items within a menu

Mental models are a natural extension of schemata. For more information, please see Amir Khella's article.

The schema theory is also helpful in testing new designs and interfaces. The theory implies that subjects who are very familiar with computers and use them frequently will lean more quickly than subjects who rarely use them. This emphasizes the point that when conducting studies of HCI tools, it is important to have a thorough understanding of your subject's background. It may significantly impact their performance.

Short term memory is volatile. Because of the 7+/-2 limit Miller describes, a task which involves many steps may disrupt the user's workflow. If the user needs to remember a sequence of five or six steps, this clearly taxes short term memory, and interrupts the actual talk that the user is trying to accomplish. This suggests that interfaces will be more usable if they are designed with these memory limitations in mind.

References:

Bartlett, F.C.,  Remembering: An Experimental and Social Study. Cambridge: Cambridge University Press, 1932.

Bartlett, F.C. Thinking. New York: Basic Books, 1958.

Bransford, J.D. & Franks, J.J., "The abstraction of linguistic ideas". Cognitive Psychology, 2, 331-350, 1971.

Chi, M., Glaser, R. & Farr, M. The Nature of Expertise. Hillsdale, NJ: Erlbaum, 1988.

Johnson, P. Human computer interaction, psychology, task analysis and software engineering, 1992.

Marsh, R.L., Sebrechts, M.M.,, Hicks, J.L., & Landau, J.D. Processing strategies and secondary memory in very rapid forgetting. Memory & Cognition, 25, 173-181, 1997.

Mandler, J. Stories, Scripts, and Scenes: Aspects of Schema Theory. Hillsdale, NJ: Erlbaum, 1984.

Peterson,L.R. & Peterson, M.J. Short-term retention of individual verbal items. Journal of Experimental Psychology, 58, 193-198, 1959.

Posner, M.L., & Konick, A.W. On the role of interference in short-term retention. Journal of Experimental Psychology, 72, 221-231, 1956.

Quinn, N. & Holland, D. Cultural Models of Language and Thought. New York: Cambridge University Press, 1987.

Rumelhart, D.E. (1980). "Schemata: The building blocks of cognition. In R.J. Spiro, B.Bruce, & W.F. Brewer (eds.), Theoretical Issues in Reading and Comprehension. Hillsdale, NJ: Erlbaum.

Shneiderman, B. Designing the User Interface: Strategies for Effective Human-Computer Interaction, 3rd Edition, Addison-Wesley, 1998.

Soloman, Howard, Cognitive Load Theory (J. Sweller), http://tip.psychology.org/sweller.html (September 2002).

Sweller, J., "Cognitive load during problem solving: Effects on learning", Cognitive Science, 12, 257-285 (1988).

Sweller, J., Instructional Design in Technical Areas, Camberwell, Victoria, Australia: Australian Council for Educational Research, 1999.