| Course | PS 350 Cognitive Psychology |
|---|---|
| Unit | Unit 3 |
| Paper type | Working memory paper |
| Length | About 1,020 words, 6 pages |
| Format | APA 7 student paper |
| School | Herzing University |
| Program | BS in Psychology |
| Updated | October 2026 |
Free sample paper for PS 350 Unit 3
Separate Stores or One Focus of Attention? Baddeley and Hitch's Multicomponent Model and Cowan's Embedded-Processes View Compared on What Each Predicts Under Interference and Added Load
Student Name
Psychology, Herzing University
PS 350: Cognitive Psychology
Instructor Name
September 22, 2026
Separate Stores or One Focus of Attention? Baddeley and Hitch's Multicomponent Model and Cowan's Embedded-Processes View Compared on What Each Predicts Under Interference and Added Load
The Term at Its Technical Meaning
In everyday speech, working memory is often used to mean short-term memory, the ability to hold a phone number for a few seconds. In cognitive psychology the term means more. Working memory refers to the system that holds information in an active state while it is being used, for reasoning, comprehension or calculation. The difference matters: a short-term store only holds, while working memory holds and works. This paper uses the term in that technical sense throughout and compares two influential models on what each predicts when a task is made harder by interference or by added load.
The Multicomponent Model
Baddeley and Hitch (1974) proposed that working memory consists of several components. A phonological loop holds speech-based information and keeps it active through inner rehearsal. A visuospatial sketchpad holds visual and spatial information. A central executive, with limited capacity, directs attention and coordinates the other components. Baddeley (2000) later added a fourth component, the episodic buffer, a capacity-limited system that binds material from the other parts, and from long-term memory, into single coherent episodes. The model was built from experiments in which participants performed two tasks at once, and its central claim is that the components draw on partly separate resources.
The Embedded-Processes View
Cowan (2001) proposed a different organization. On his view, working memory is not a set of separate stores but the portion of long-term memory that is currently activated, with a smaller subset of that activated material held in the focus of attention. The focus has a strict capacity limit. Reviewing a wide range of studies designed to prevent rehearsal and grouping, Cowan argued that the focus holds about four chunks, meaningful units, give or take one, rather than the larger number often cited from older research. Domain-specific effects arise, in this view, from interference among similar items within activated memory rather than from separate stores.
What Each Predicts Under Interference
The models diverge most clearly when a second task is added.
Same-domain versus cross-domain interference
Suppose participants must remember a list of letters while performing a second task. The multicomponent model predicts that a verbal second task, such as repeating a word aloud, which occupies the phonological loop, will impair letter memory much more than a spatial second task, such as tapping a pattern, which occupies the sketchpad. Baddeley and Hitch (1974) reported dual-task results of this kind, in which interference depended on whether the two tasks drew on the same component. The embedded-processes view also predicts more interference from similar material, because similar items interfere with each other in activated memory, so both models expect the same direction of effect.
Load that exceeds the focus of attention
The models differ on what happens when the total load is large but spread across domains. The multicomponent model predicts that verbal and visual material can be held largely in parallel, so a person can maintain a verbal list and a spatial array with relatively little cost, provided the central executive is not overloaded. The embedded-processes view predicts a general limit: because both kinds of material must compete for the same focus of attention when they need to be actively held, combining them should still produce a cost, with total capacity near the four-chunk limit.
Preventing rehearsal and grouping
When rehearsal is blocked, the multicomponent model predicts that verbal memory will fall because the phonological loop can no longer refresh its contents. The embedded-processes view predicts that what remains reflects the focus of attention, which is why Cowan used tasks that prevent rehearsal and chunking to estimate its capacity.
Why the Difference Matters Outside the Laboratory
The disagreement can sound technical, but it has practical consequences for how tasks should be designed. Consider a cashier who must hold a customer's spoken order in mind while scanning items and reading prices on a screen. On the multicomponent model, the spoken order occupies the phonological loop and the visual scanning occupies the sketchpad, so the two can proceed with modest interference, and the main risk comes from anything else that also needs the loop, such as a second customer speaking. On the embedded-processes view, both the order and the prices compete for a single, small focus of attention whenever they must be actively held, so the cost of combining them is larger, and the best design reduces how many items must be held at once, for example by displaying the order on the screen. The two models therefore point to somewhat different fixes for the same problem. Knowing which model a recommendation rests on, and what each predicts, is part of applying cognitive research responsibly rather than borrowing its vocabulary.
Reporting a Capacity Limit
Capacity limits should be reported with their unit and their conditions. Cowan's estimate of about four is a number of chunks, under conditions that limit rehearsal and grouping; it is not a claim that people can remember only four letters in any situation. With chunking, a person can hold far more raw items. A number reported without these qualifications is easily misread.
Which Model the Evidence Favors
Neither model has displaced the other. The multicomponent model explains the robust finding that same-domain interference is greater than cross-domain interference and has been elaborated over decades. The embedded-processes view explains the general capacity limit and the close link between working memory and attention. Baddeley's addition of the episodic buffer, a general store linked to long-term memory, moves the multicomponent model somewhat toward Cowan's emphasis on integration. A fair conclusion is that the models agree on much of the data and differ mainly in how they divide the system.
A Note on Personal Span Tests
Online span tests report a single score under unknown conditions, often allowing rehearsal and chunking. Such a score cannot test either model and does not belong in the paper. Published studies, with their controlled conditions and many participants, are the only appropriate evidence for comparing the models.
References
Baddeley, A. (2000). The episodic buffer: A new component of working memory? Trends in Cognitive Sciences, 4(11), 417-423. https://doi.org/10.1016/S1364-6613(00)01538-2
Baddeley, A. D., & Hitch, G. (1974). Working memory. In G. H. Bower (Ed.), The psychology of learning and motivation (Vol. 8, pp. 47-89). Academic Press. https://doi.org/10.1016/S0079-7421(08)60452-1
Cowan, N. (2001). The magical number 4 in short-term memory: A reconsideration of mental storage capacity. Behavioral and Brain Sciences, 24(1), 87-114. https://doi.org/10.1017/S0140525X01003922
PS 350 Unit 3 instructions, in plain terms
The third PS 350 unit asks students to treat working memory as more than short-term storage and to compare models on what they predict. Expect to define the term at its technical meaning, attribute each model to the researchers who proposed it, describe each in its own terms and compare their predictions about performance under interference or added load. Two models are usually enough for a focused comparison. Capacity limits can be reported as numbers only with their units and conditions, since a figure such as four applies to chunks under specific conditions. Your own score on an online span test cannot test either model and does not belong in the paper.
Inside the PS 350 Unit 3 example
The sample begins by distinguishing working memory from short-term storage. The multicomponent model is set out with its components and the later episodic buffer, then Cowan's view with its activated memory and limited focus. The comparison is organized by manipulation under run-in headings: same-domain versus cross-domain interference, load spread across domains, and blocked rehearsal, with convergent and divergent predictions marked. A cashier example shows how the models point to different design fixes. A section explains how to report capacity limits, another weighs which model the evidence favors, and a short note excludes personal span tests. Each manipulation is described before the predictions are compared.
Reading the PS 350 Unit 3 rubric
Herzing's rubric for the working memory paper usually allots points to the technical definition, accurate attribution of models, a prediction-based comparison, careful reporting of capacity and a reasoned conclusion, plus APA. The definition earns credit when it separates holding from working. Attribution scores when each model and its later revisions are credited. Comparison is often the largest criterion, rewarding papers organized by manipulation rather than by model. Capacity earns points when numbers come with units and conditions. Conclusions are credited when they note where the models converge on the same data and differ mainly in how they divide the system. Showing where the models agree is often credited as much as showing where they differ.
PS 350 Unit 3 help before you submit
Working memory papers lose points when they use the term as a synonym for short-term memory. Another frequent weakness is describing each model separately without comparing predictions. Students often report capacity as seven items, or four items, without units or conditions. Some present the multicomponent model in its 1974 form without the episodic buffer. Others declare one model correct when both explain much of the evidence. Including personal span test results adds nothing. Overlooking the practical implications of the models, such as how tasks should be designed, misses a chance to show understanding. Ignoring the role of chunking, which changes what counts as one unit, is a related slip. Omitting dual-task evidence removes the main data the comparison rests on.
Get a PS 350 Unit 3 example written to your instructions
Put the Unit 3 instructions and your PS 350 rubric in front of us, with the models your section covers. We write a custom example that separates the terms at the start, derives each model's predictions and brings the evidence against them. First custom sample free, back in 24 to 48 hours.
More PS 350 papers and psychology samples
- PS 350 Unit 1: Cognitive methods paper
- PS 350 Unit 2: Attention paper
- PS 350 Unit 4: Long-term memory and encoding paper
- PS 350 Unit 5: Memory distortion paper
- PS 350 Unit 6: Language or knowledge paper
- PS 350 Unit 7: Problem solving and decision paper
- PS 350 Unit 8: Applied cognition paper
- PS 303 Unit 3: Attitudes and persuasion paper
- PS 323 Unit 3: Cognitive and ability differences paper
- PS 360 Unit 3: Anxiety disorders case
- PS 101 Unit 3: Biological basis short paper
PS 350 Unit 3 questions, answered
How many models should the comparison cover?
Two handled properly is usually plenty, and your instructions may name them. Each model you add has to be given its components, its predictions and the evidence bearing on them, so a third account often means three thin descriptions instead of one real comparison. Where the choice is yours, pick accounts your readings actually cover in detail.
Do I need to report capacity limits as a number?
Report whatever the source reports, with the source named, and resist rounding it into a slogan. Published estimates differ by material, by procedure and by how the limit is defined, so a figure quoted without those conditions says less than it appears to. If your readings give a range rather than a value, write the range.
Can I report my own performance on a span test I tried online?
Interesting to try, not usable as evidence. An unsupervised web version has no controlled procedure, no comparison condition and no documented scoring, so it cannot support a claim in a paper built on experimental method. If your classroom runs a demonstration, your instructions will say how the data may be used, and that is the version to rely on.
Where can I find a free PS 350 Unit 3 sample paper?
Read the whole PS 350 Unit 3 working memory paper here without charge, with both models, the prediction comparison and margin notes. Comparing different models of working memory? Email the instructions and a prediction-based comparison is prepared for you free.
Do I need to report capacity limits as a number in PS 350 Unit 3?
Only with the unit and conditions. The sample reports Cowan's estimate of about four chunks under conditions that prevent rehearsal and grouping, not a fixed number of items.