Key Concepts of Information Processing Models

Why This Matters

Information processing models are the backbone of cognitive psychology. They explain how your mind takes in raw sensory data and transforms it into memories, decisions, and actions. On exams, you'll be tested on your ability to distinguish between competing models, explain their mechanisms, and apply them to real-world scenarios like learning, problem-solving, and cognitive errors. These models connect to broader themes of memory systems, attention, cognitive development, and individual differences in intelligence.

Don't just memorize the names and components of each model. Know what problem each model was designed to solve, how it differs from alternatives, and when you'd use it to explain a cognitive phenomenon. The strongest exam responses compare models directly, showing you understand not just what each model claims, but why psychologists developed different frameworks in the first place.

---

Sequential Stage Models

These classic models propose that information moves through distinct stages in a fixed order. The key mechanism is serial processing: each stage must complete before the next begins.

Atkinson-Shiffrin Model (Multi-Store Model)

Proposed in 1968, this was one of the first comprehensive models of memory and remains a foundational reference point for everything that came after.

• Three sequential stores: sensory memory, short-term memory (STM), and long-term memory (LTM). Information flows in one direction through the system.
• Attention gates the transfer from sensory memory to STM. Without focused attention, sensory information decays within milliseconds to seconds.
• Rehearsal is essential for LTM encoding. The model predicts that more repetition equals stronger memories, a claim later models challenged significantly.
• One limitation worth knowing: the original model treats STM as a single, passive holding area. It also implies that information flows strictly one way, but retrieval from LTM clearly feeds back into STM during tasks like problem-solving.

Sensory Memory Model

• Ultra-brief storage of raw sensory input. Iconic memory (visual) lasts roughly 250ms; echoic memory (auditory) lasts roughly 3–4 seconds.
• High capacity, rapid decay. You capture almost everything in the sensory field but lose it immediately without attention.
• Sperling's partial report procedure (1960) demonstrated that sensory memory holds far more than we can report. Participants could recall any cued row of a letter grid if prompted immediately, but performance dropped sharply after a brief delay. This supported the existence of a large-capacity but fleeting sensory buffer.

---

Active Processing Models

These models reject passive storage in favor of dynamic manipulation of information. The core insight: memory isn't a filing cabinet, it's a workspace.

Baddeley's Working Memory Model

Baddeley and Hitch (1974) proposed this model specifically because the Atkinson-Shiffrin STM couldn't explain how people do multiple cognitive tasks at once.

• Four components replace STM: the central executive (attentional control and coordination), phonological loop (verbal and acoustic information), visuospatial sketchpad (images and spatial layouts), and episodic buffer (integrates information from the other components and LTM). The episodic buffer was added later, in 2000.
• Parallel processing of different information types. You can hold a phone number (phonological loop) while navigating a room (visuospatial sketchpad) simultaneously, because these are separate subsystems.
• Central executive limitations explain why multitasking fails when two tasks compete for the same subsystem. Try holding a conversation while writing an email: both demand the phonological loop, so performance on one or both suffers.

Levels of Processing Model (Craik and Lockhart)

Craik and Lockhart (1972) shifted the focus away from where information is stored and toward how it's encoded.

• Depth determines durability. Semantic processing (thinking about meaning) creates stronger, more retrievable memories than structural processing (noticing appearance) or phonemic processing (focusing on sound).
• Elaboration and distinctiveness enhance encoding. Connecting new information to existing knowledge builds more retrieval pathways, making recall easier.
• Challenges the rehearsal assumption. Maintenance rehearsal (rote repetition, like repeating a phone number) is far less effective than elaborative rehearsal (making meaningful connections). This directly contradicts Atkinson-Shiffrin's emphasis on repetition as the path to LTM.
• A common criticism: "depth" is hard to measure independently of memory performance, which risks circular reasoning (deeper processing leads to better memory, and we know processing was deeper because memory was better).

---

Network and Connectionist Models

These models draw inspiration from neural architecture, emphasizing distributed representation and simultaneous activation rather than discrete storage locations.

Parallel Distributed Processing Model (PDP)

Also called connectionism, PDP models (Rumelhart & McClelland, 1986) represent a fundamentally different way of thinking about how knowledge is stored.

• Simultaneous activation across interconnected nodes. No single location "holds" a memory. Instead, patterns of activation across the network represent knowledge, similar to how no single neuron stores a concept.
• Learning through weight adjustment. Connections between nodes strengthen or weaken based on experience, mimicking synaptic plasticity. This is how the network "learns" over time.
• Graceful degradation. Partial damage doesn't destroy memories completely because information is distributed across many connections. This explains why brain injuries often cause partial rather than total memory loss.

---

Long-Term Memory Organization

Understanding how information is organized in LTM is essential for explaining amnesia cases, skill learning, and the distinction between knowing that and knowing how.

Long-Term Memory Systems (Declarative and Procedural)

This taxonomy, developed largely by Squire and others, divides LTM based on what kind of knowledge is stored and how it's accessed.

• Declarative (explicit) memory divides into episodic (personal events with spatiotemporal context, like your 10th birthday party) and semantic (general facts without personal context, like knowing that Paris is the capital of France).
• Procedural (implicit) memory includes motor skills, habits, and conditioned responses. These are preserved in amnesia patients like H.M. (Henry Molaison), who could learn new motor tasks but had no conscious memory of the training sessions.
• Different neural substrates. Declarative memory depends on the hippocampus and medial temporal lobe. Procedural memory involves the basal ganglia and cerebellum. This anatomical dissociation is why damage to one system can leave the other intact.

---

Dual-System Models

These frameworks propose that cognition operates through two qualitatively different processing modes, explaining both efficient intuition and systematic reasoning.

Dual-Process Theory

Popularized by Kahneman (2011), though built on decades of research by multiple psychologists, this theory distinguishes two modes of thinking.

• System 1 is fast, automatic, and effortless. Pattern recognition, emotional reactions, and heuristics operate here. You use System 1 when you read a familiar word or recognize a friend's face.
• System 2 is slow, deliberate, and effortful. Logical analysis, complex calculations, and conscious reasoning require this system. You use System 2 when you solve $17 \times 24$ in your head.
• Cognitive biases emerge when System 1 dominates inappropriately. The availability heuristic (judging probability by how easily examples come to mind) and confirmation bias (seeking evidence that supports existing beliefs) reflect System 1 shortcuts overriding System 2 analysis.

PASS Theory of Intelligence

Das, Naglieri, and Kirby developed PASS theory as an alternative to traditional IQ models, grounding intelligence in specific cognitive operations.

• Four processes define intelligence: Planning (goal-setting, strategy selection, monitoring progress), Attention (selective and sustained focus), Simultaneous processing (integrating information holistically, like understanding a diagram), and Successive processing (handling sequential information, like following multi-step directions).
• Executive functions coordinate these processes. Deficits in any single component produce distinct cognitive profiles, which has practical implications for diagnosing learning difficulties.
• Challenges $g$-factor models. Intelligence isn't a single unitary ability but a combination of separable cognitive processes with different neural bases.

---

Cognitive Efficiency Models

These models address resource limitations and how the mind manages competing demands. They're critical for understanding learning, expertise development, and cognitive overload.

Cognitive Load Theory

Developed by John Sweller in the late 1980s, this theory applies directly to instructional design and explains why some teaching methods work better than others.

• Three types of load: intrinsic (inherent complexity of the material itself), extraneous (load caused by poor presentation or unnecessary information), and germane (productive mental effort directed toward building schemas and understanding).
• Working memory bottleneck. When total cognitive load exceeds working memory capacity, learning breaks down. Effective instructional design should minimize extraneous load so that more capacity is available for germane processing.
• Expertise reversal effect. Instructional supports that help novices (like worked examples or diagrams with labels) can actually harm experts by adding unnecessary processing load. Experts already have schemas that make those supports redundant.

Executive Function Model

• Three core components: working memory (holding and manipulating information), cognitive flexibility (shifting between tasks or mental sets), and inhibitory control (suppressing automatic or prepotent responses). These three components interact constantly during complex tasks.
• Prefrontal cortex dependent. Executive functions develop throughout childhood and adolescence as the prefrontal cortex matures, which explains age-related differences in self-regulation, planning, and impulse control.
• Predicts real-world outcomes. Executive function measured in early childhood correlates with later academic achievement, health behaviors, and financial stability, making it one of the strongest cognitive predictors of life outcomes.

---

Quick Reference Table

---

Self-Check Questions

1. Which two models most directly challenge the Atkinson-Shiffrin assumption that rehearsal is the primary mechanism for long-term encoding? Explain what alternative mechanism each proposes.

2. A patient with hippocampal damage can still learn to ride a bicycle but cannot remember meeting their therapist yesterday. Which memory systems are preserved versus impaired, and which model best explains this dissociation?

3. Compare and contrast Baddeley's Working Memory Model with the original short-term memory component of Atkinson-Shiffrin. What phenomena can Baddeley's model explain that the earlier model cannot?

4. A student who highlights entire textbook pages performs worse on exams than a student who writes summary notes. Which model provides the strongest explanation, and what specific mechanism would you cite?

5. How would you use Dual-Process Theory to explain why experienced doctors sometimes make diagnostic errors that medical students catch? Which system is operating in each case?