CS 6750 Quiz 4

50 flashcards covering distributed cognition, situated action, activity theory, value-sensitive design, and agile HCI for CS 6750

What You'll Learn

Free flashcards for Georgia Tech CS 6750 Human-Computer Interaction Quiz 4. Covers distributed cognition (Hutchins), situated action (Suchman), activity theory, value-sensitive design, artifacts and politics (Winner), agile HCI, InfoVis, and social change through design. 50 cards for grad HCI students and UX researchers.

Key Topics

  • Distributed cognition: unit of analysis, cognitive offloading, cognitive load
  • Hutchins's cockpit study: speed bugs, external memory artifacts, over-attribution problem
  • Situated action (Suchman): plans vs. improvisation, recognition over recall
  • Activity theory: three-level hierarchy, automatization, Nardi's comparison
  • Artifacts have politics (Winner): inherently political technologies, Robert Moses bridges
  • Value-sensitive design: three investigations, direct vs. indirect stakeholders, cross-cultural values
  • Three goals of HCI: usability, research, and designing for change
  • Designing for social change: bicycle example, fluorescent lightbulb, unintended consequences
  • Agile HCI: live prototyping, AB testing, parallel track method, Boehm and Turner conditions
  • InfoVis and distributed cognition: internal vs. external representations, problem isomorphs, epistemic actions
  • Processor, predictor, and participant models of the user
  • Gulf of execution and gulf of evaluation
  • Participatory design, action research, and design-based research

Looking for more human computer interaction resources? Visit the Explore page to browse related decks or use the Create Your Own Deck flow to customize this set.

How to study this deck

Start with a quick skim of the questions, then launch study mode to flip cards until you can answer each prompt without hesitation. Revisit tricky cards using shuffle or reverse order, and schedule a follow-up review within 48 hours to reinforce retention.

Preview: CS 6750 Quiz 4

Question

What is distributed cognition?

Answer

A theory in cognitive science and HCI proposing that thinking and problem-solving don't happen solely inside one person's head. Instead, cognition is spread across multiple people, tools, and artifacts working together as a system. Example: a surgeon, their instruments, monitors, and an assisting nurse collectively form a cognitive system that performs a task no single member could alone.

Question

How does distributed cognition change the 'unit of analysis' in HCI?

Answer

Traditional cognitive science treats the individual human mind as the unit of analysis. Distributed cognition expands this to the entire system — people, artifacts, and their relationships. Instead of asking 'what does this person know?', we ask 'what does this system know?' Example: a cockpit's pilots, speed cards, and instrument markers together 'remember' landing speeds even though no single pilot could recall them all from memory.

Question

What is cognitive offloading?

Answer

The process of transferring a cognitive task from the human mind onto an artifact or tool, freeing up mental resources for other tasks. Example: writing a grocery list offloads the task of remembering items so your working memory can focus on something else while you shop. In HCI, well-designed interfaces offload cognitive tasks from users to the system.

Question

What is cognitive load, and how do artifacts help manage it?

Answer

Cognitive load refers to the limited amount of information the human mind can actively process at one time. Artifacts help by taking over some of the cognitive work, effectively spreading the load across more resources. Example: a GPS reduces driving cognitive load by handling navigation, freeing attention for traffic, speed, and passengers.

Question

What is an external memory artifact?

Answer

Any tool or artifact that stores information on behalf of a person, so they don't have to hold it in their own memory. Examples: a to-do list, a calendar, sticky notes, email, a checkbook's carbon copy, or a clock. In distributed cognition, these are treated as genuine parts of a cognitive system — not just conveniences.

Question

In Edwin Hutchins's cockpit study ('How a Cockpit Remembers Its Speeds'), how is memory distributed across the system?

Answer

Three levels of memory are distributed across artifacts: (1) Long-term memory = a booklet of speed cards covering all possible flight parameters. (2) Short-term memory = the specific card pulled out and pinned up for the current flight. (3) Working memory = small 'speed bug' markers placed directly on the speedometer for instant visual comparison during descent. No single component — pilot or artifact — holds all this alone.

Question

What are 'speed bugs' in the context of distributed cognition?

Answer

Small movable markers placed on a cockpit speedometer to indicate specific speeds at which the pilots must make configuration changes during landing. They externalize working memory — pilots can visually compare the current speed needle to the bug markers without doing any mental math, reducing cognitive load at a critical, high-stakes moment.

Question

What is the over-attribution problem in distributed cognition?

Answer

The mistake of crediting an individual person with cognitive abilities that actually belong to the broader system they're part of. Example: saying 'the pilot is skilled at remembering landing speeds' when in reality it's the system of pilot + speed cards + instrument markers that holds that knowledge. Isolating the pilot from those tools would reveal they can't do it alone.

Question

What is social cognition, and why is it relevant to interface design?

Answer

Social cognition studies how people accomplish tasks through social connections, and how cognitive processes like memory, perception, and learning are shaped by social relationships. It's relevant to interface design because many modern interfaces are inherently social (social media, multiplayer games, collaborative tools). Designing them well requires understanding how people actually think about and manage social relationships — not just how they use the software.

Question

What is situated action in HCI?

Answer

A theory proposing that human behavior is best understood as improvised responses to immediate situations, rather than the execution of pre-formed plans. From an HCI perspective, this means the task a user performs grows out of their real-time interaction with an interface — it isn't fully defined in advance by the designer. Implication: designers should focus on what resources a user has available at any given moment, not assume they're following a predetermined script.

Question

What did Lucy Suchman argue in 'Plans and Situated Actions'?

Answer

Suchman contrasted two views of human behavior: (1) the plans-first view, where actions follow from pre-formed plans (the dominant assumption in interface design), and (2) the situated-first view, where people simply act in the world and 'plans' are interpretations we assign to those actions retroactively. For designers, this suggests focusing on the immediate screen context — what can a user do and understand right now — rather than assuming they're executing a mental plan.

Question

Why does situated action claim that recognition is easier than recall, and what does this mean for design?

Answer

Memory is highly context-dependent. Recognizing the right option is easier because the context (seeing the option in front of you) cues the memory. Recalling from scratch provides no such cues. Implication for design: present users with options to recognize rather than forcing them to remember commands, values, or steps from memory. Menus, autocomplete, and visual affordances all leverage this principle.

Question

What is activity theory, and what does it add to HCI beyond task analysis?

Answer

Activity theory is a broad framework for understanding human interaction that predates HCI. It expands the unit of analysis from 'the task' to the broader 'activity' — including the user's motivations and reasons for doing the task at all. Example: designing a fitness tracker looks different if the user is training for a marathon (intrinsic motivation) vs. being required to use it by their insurance company (external mandate). The why changes the design.

Question

What are the three levels of the activity theory hierarchy?

Answer

Activities (the highest level — why the user is doing something, their overarching motive) → Actions (what the user consciously does to achieve that goal) → Operators (automatic, subconscious steps that make up actions). Example: 'get healthy' is an activity; 'go for a run' is an action; 'tie your shoes' starts as a conscious action but becomes a subconscious operator once you've done it thousands of times.

Question

How can actions shift up and down the activity theory hierarchy?

Answer

With practice, conscious actions become subconscious operators (automatization). With difficulty or novelty, operators can become conscious again. Example: the first time you drive a car, changing gears is a deliberate, conscious action requiring thought. After experience, it becomes an automatic operator. This is similar to the concept of learning curves in HCI — good design accelerates the move from conscious action to operator.

Question

How does Nardi compare activity theory, situated action, and distributed cognition?

Answer

Activity theory and distributed cognition are both goal-driven and rely on permanent, persistent structures (like long-term memory and stored plans). Situated action de-emphasizes goals, focusing instead on improvisation and treating goals as constructed retroactively. The key difference between activity theory and distributed cognition: activity theory sees people and artifacts as fundamentally different (humans have consciousness); distributed cognition treats them as conceptually equivalent cognitive components.

Question

What does it mean to say 'artifacts have politics'? (Langdon Winner, 1980)

Answer

Winner argued that technological artifacts can embody and enact political qualities — they can reflect or reinforce forms of authority, power, and social order. This doesn't mean a toaster has political opinions. It means that design decisions have social consequences, sometimes intentional and sometimes not. Example: designing a public park with no benches discourages homeless people from staying — the artifact (the park) enacts a social/political preference through its design.

Question

What is the difference between 'inherently political technologies' and 'technical arrangements as social order'? (Winner)

Answer

Inherently political technologies are only compatible with certain political structures by their very nature. Example: nuclear power requires top-down, centralized management due to complexity and safety — it structurally favors authoritarian organization. Technical arrangements as social order are technologies with no inherent political nature, but which are deliberately deployed to achieve political ends. Example: Robert Moses's bridges — bridges are not inherently political, but building them too low for buses was a deliberate political choice.

Question

Describe the Robert Moses parkway bridge example and its HCI relevance.

Answer

Moses was a city planner in 1900s New York who oversaw the construction of parks on Long Island and the parkways leading to them. He intentionally built the parkway bridges too low for buses to pass under, ensuring only people wealthy enough to own cars could access the parks — excluding poor and minority communities. In HCI terms, this illustrates how design decisions that appear neutral can be deliberate tools of social exclusion. The artifact (a bridge) was used as a technical arrangement to achieve a social/political order.

Question

What is value-sensitive design (VSD)?

Answer

An approach to design that systematically accounts for human values throughout the design process — not just usability or functionality. Developed by Batya Friedman and colleagues at the University of Washington. It asks: beyond whether the interface is useful and usable, is it consistent with the values of the people who use it and are affected by it? Example: designing a healthcare app that is easy to use but also respects patient privacy and autonomy.

Question

What are the three types of investigations in value-sensitive design?

Answer

(1) Conceptual investigations: thought experiments identifying stakeholders and how they're affected by the design. (2) Empirical investigations: user research exploring how people understand and prioritize values in actual use. (3) Technical investigations: system-level analysis asking whether the technology as designed is actually compatible with user values. All three are needed for a complete VSD analysis.

Question

What is the difference between direct and indirect stakeholders in value-sensitive design?

Answer

Direct stakeholders are people who actually interact with and use the system. Indirect stakeholders are people who never use the system but are nonetheless affected by it. Example: in a hospital patient records system, doctors and nurses are direct stakeholders; patients whose data is managed are indirect stakeholders. VSD argues designers must account for both groups.

Question

Why can values differ across cultures, and what challenge does this create for interface design?

Answer

Different cultures may hold fundamentally different values, and an interface designed for one culture's values may conflict with another's. Example: the EU's 'right to be forgotten' law values privacy and personal control over online information. Many US-based tech companies were not designed with this value in mind, requiring significant redesign. Meanwhile, some argue the law conflicts with free speech values. Global interfaces must navigate these conflicts.

Question

What are the three goals of HCI?

Answer

(1) Help a user do a task — the usability goal: make things easier and more efficient. (2) Understand how a user does a task — the research goal: generate knowledge about human behavior. (3) Change the way a user does a task — the values goal: use design to promote some value like safety, health, or equity, sometimes at the expense of usability. Example of goal 3: a car that beeps until you put on a seatbelt isn't usable by seatbelt-haters, but it advances a safety value.

Question

How can an interface create positive social change by design? Give an example.

Answer

Designers can intentionally build interfaces that nudge users toward socially beneficial behavior — without directly telling them to. Example: a weather crowdsourcing app that asks participants to go outside hourly to take readings incidentally gets people to exercise regularly. Pokemon GO got millions of people to walk more without ever framing itself as a fitness app. Facebook's Like-only button fosters positivity by making negativity structurally impossible.

Question

How can an interface create negative social outcomes unintentionally?

Answer

Innocent, logical design decisions can reinforce existing inequalities without any malicious intent. Example: the internet was built on top of existing infrastructure (phone lines → cable → fiber). Areas with more developed infrastructure — typically wealthier areas — got faster internet first. This widened the digital divide even though no one intended to disadvantage poorer communities. Good designers anticipate these second-order effects.

Question

What is 'designing for change' in HCI, and how does it differ from designing for usability?

Answer

Designing for usability makes a task easier. Designing for change uses an interface to modify user behavior or create social outcomes, sometimes at the direct expense of usability. Example: a car's seatbelt warning beep is annoying by design — that annoyance is the mechanism that drives behavior change. Facebook's lack of a dislike button limits social expression (a usability weakness) but reduces cyberbullying (a social value win).

Question

What is the bicycle example and what does it illustrate about unintended social change from technology?

Answer

The bicycle was invented as a transportation device, not as a social movement tool. But because it was affordable and targeted at individuals, it gave women the ability to travel independently without relying on men. This accidentally triggered a chain of social changes: new women's attire (dresses couldn't be worn on a bike), greater visible independence, and challenges to traditional gender roles. Susan B. Anthony called it the greatest contributor to women's emancipation. The lesson: technology can have profound unintended social consequences.

Question

What is the fluorescent lightbulb example, and what does it illustrate about political motivations in technology design?

Answer

In 1938, GE developed a highly energy-efficient fluorescent bulb. Power companies, fearing reduced electricity consumption, lobbied against it. After a legal battle involving the US government, the fluorescent bulbs ultimately sold were intentionally less efficient than they could have been — to protect power companies' profits. This illustrates how business and political interests can deliberately limit technological capability, resulting in an inferior product for consumers.

Question

What are the three models of the user in HCI, and what level of context does each represent?

Answer

(1) Processor model: narrowest — the user as a biological computer, defined by sensory, memory, and motor limits. (2) Predictor model: mid-level — the user actively predicts, plans, and interacts with a task through an interface. Gulfs of execution and evaluation live here. (3) Participant model: broadest — the user as a participant in the social, cultural, and political world. Distributed cognition, activity theory, value-sensitive design, and situated action all operate at this level.

Question

What is the gulf of execution?

Answer

The gap between what a user wants to do and the actions that are available in the interface — essentially, how hard it is to figure out how to do something. A small gulf of execution means the available actions clearly map to the user's intentions. A large gulf means the user struggles to translate their goal into the right steps. Good design minimizes the gulf of execution through clear affordances, labels, and logical action structures.

Question

What is the gulf of evaluation?

Answer

The gap between the system's output or state and how easily the user can interpret what happened — essentially, how hard it is to understand the result of your action. A small gulf of evaluation means the system gives clear, immediate, interpretable feedback. A large gulf means the user must guess whether their action worked. Good design minimizes the gulf of evaluation through clear feedback mechanisms.

Question

What does it mean for an interface to be 'invisible,' and why is this the goal?

Answer

An invisible interface is one where the tool disappears from the user's awareness — they focus entirely on the task, not on operating the interface. This is the ideal in HCI because the interface becomes a transparent conduit between the user and their goal. Example: an experienced driver doesn't think about the steering wheel; they think about where they're going. Good interfaces achieve this either immediately (no learning curve) or through a smooth learning curve.

Question

What is agile development, and why did it become viable as a software methodology?

Answer

Agile is a software development approach emphasizing rapid iteration, early delivery, and continuous improvement based on real user feedback. It became viable because: (1) development costs dropped dramatically, (2) software distribution became essentially free, (3) updates can be pushed instantly, and (4) usage data and user feedback can now be collected automatically at no cost. These changes made 'ship fast, fix fast' a better strategy than 'get it perfect before shipping.'

Question

Why is agile development exciting but also risky for HCI practitioners?

Answer

Exciting: it embraces rapid feedback cycles, gets designs in front of real users quickly, and allows continuous improvement — all values HCI already holds. Risky: it can bypass the careful, deliberate need-finding and prototyping stages that protect users from poor designs. In high-stakes domains (healthcare, aviation), shipping a flawed interface to learn from real failures is unacceptable. Agile works best where the cost of failure is low.

Question

What is live prototyping?

Answer

A prototyping approach where the 'prototype' is a fully functional, deployable interface — not a mockup or wireframe. This is possible because modern tools make building real interfaces nearly as fast as building paper prototypes. Example: tools like Optimizely allow drag-and-drop creation of real, live webpages. The line between prototype and product blurs, enabling faster feedback loops with real users in real conditions.

Question

What is AB testing in the context of HCI and agile development?

Answer

A method of comparing two interface versions (A = current, B = new variant) by rolling the new version out to a small subset of real users and measuring behavioral differences. If B performs better, it's rolled out to everyone; if worse, it's rolled back. No recruitment cost, no lab setting — feedback comes automatically from real usage. Example: testing whether showing a course as '420 hours total' vs. '10 hours per week' affects enrollment rates.

Question

According to Boehm and Turner, what conditions must be met for agile development to be appropriate?

Answer

(1) Low criticality: errors and poor usability won't lead to serious consequences (a bad game update is annoying; a bad surgical interface could kill someone). (2) Frequently changing requirements: the product's needs evolve quickly enough that long upfront design cycles would be obsolete by completion. (3) Small, adaptable teams: agile breaks down with large bureaucratic teams that require rigid structure.

Question

What is the parallel track method for combining HCI research with agile development?

Answer

A strategy where the HCI/design team works one sprint ahead of the development team. The design team spends their sprint doing need-finding, low-fidelity prototyping, and early evaluation. They hand the results to the development team, who implements in their sprint. This preserves user-centered design practices within the rapid cadence of agile development.

Question

What is participatory design?

Answer

A design approach in which actual users are included as active members of the design team — not merely as subjects of research or sources of feedback, but as genuine collaborators in making design decisions. This ensures the user perspective is continuously present throughout the process. The main risk: one user on the team might not represent the full diversity of the user population, so their view must not be over-weighted.

Question

What is action research?

Answer

A research methodology in which practitioners address a real, immediate problem by trying potential solutions and systematically reflecting on the results. Research and problem-solving happen simultaneously and iteratively. Importantly, action research is conducted by the actual users of the system, not by outside researchers. Example: a teacher tries a new instructional activity, observes student outcomes, reflects, and adjusts — generating both a solution and new knowledge.

Question

What is design-based research?

Answer

A methodology in which designers create interventions based on their current theoretical understanding of a problem, then use the success or failure of those interventions to refine their understanding. Unlike action research, it can be conducted by outside practitioners. Example: if you believe jaywalking at an intersection is caused by poor sign visibility, you install a more visible sign and measure whether jaywalking decreases — the result tests and refines your theory.

Question

What is the iterative design lifecycle in HCI?

Answer

The core methodology of HCI: (1) Need-finding — understand users and their context. (2) Brainstorm design alternatives — generate multiple possible solutions. (3) Prototype — build low-to-high fidelity representations of the best ideas. (4) Evaluate with real users — gather feedback. Then repeat, with each cycle increasing fidelity and understanding. The cycle never truly ends — even after launch, usage data feeds back into need-finding.

Question

How does Liu et al. define information visualization (InfoVis), and what is the central understudied dimension?

Answer

InfoVis is defined as 'the use of computer-supported, interactive, visual representations of abstract data to amplify cognition' (Card, Mackinlay & Shneiderman). The two well-studied dimensions are representation (how data is displayed) and interaction (how users manipulate those displays). The understudied third dimension is cognition — how thinking actually happens within the human-visualization system. Liu et al. argue distributed cognition is the right framework for studying this.

Question

What is Hutchins's critique of the phrase 'amplifying cognition'?

Answer

Hutchins argues that saying tools 'amplify' cognition implies we're making the same cognitive process stronger. In reality, tools transform the task into a completely different set of cognitive operations. Writing on paper doesn't make your memory stronger — it replaces remembering with writing, reading, and interpreting. The important distinction: the cognitive properties of a system of person + tool are fundamentally different from, not merely more powerful than, the properties of the person alone.

Question

What does it mean that cognition is an 'emergent property' of interaction, according to distributed cognition?

Answer

Emergence means the whole has properties that none of its parts have in isolation. A cognitive system — people + artifacts working together — can think, remember, and solve problems in ways that no single component can accomplish alone. Example: in an intelligence analysis scenario, a human analyst + a visualization software system + a notepad collectively generate insights that neither the human nor the software could produce independently.

Question

What are internal vs. external representations, and why does the distinction matter for InfoVis?

Answer

Internal representations are mental structures in the brain (mental models, schemas, memory traces). External representations are observable structures outside the brain (charts, notes, labels, graphs, physical markers). InfoVis is fundamentally about creating external representations that work in coordination with internal representations. Distributed cognition argues you can't understand InfoVis by studying only what happens in the brain — you must study how internal and external representations interact.

Question

What are problem isomorphs, and what do experiments with them reveal about distributed representations?

Answer

Problem isomorphs are different physical or visual representations of logically identical problems. Zhang and Norman studied the Tower of Hanoi in three versions: one where rules were physically enforced by the materials (external), one where some rules were external and some had to be remembered, and one where all rules had to be held in memory. Result: the more rules externalized in the representation, the faster and more accurately people solved the problem. Takeaway: the distribution of information across internal and external media directly shapes cognitive difficulty.

Question

What are pragmatic actions vs. epistemic actions, and why does the distinction matter for HCI?

Answer

Pragmatic actions directly advance a task toward completion (moving a chess piece to a better position). Epistemic actions modify the environment to make thinking easier, not to directly achieve the goal (rotating a Tetris piece extra times to mentally simulate its placement). Epistemic actions are a form of cognitive offloading — using the environment to reduce internal cognitive effort. HCI designers should ask: does our interface support the epistemic actions users naturally want to take?

Question

What is cognitive ethnography, and why does distributed cognition favor it as a research method?

Answer

Cognitive ethnography is in-depth, in-situ observational study of how people accomplish tasks using artifacts in real environments, focusing on how coordination between internal and external representations is achieved. DCog favors it because cognition is an emergent property of real interaction in context — stripping users out of their environment in a lab setting destroys the very phenomenon being studied. Cognitive ethnography can be complemented with controlled experiments when observations leave open questions.

Question

What is the reductionist approach in cognitive science, and what is DCog's objection to it?

Answer

The reductionist approach studies cognition by isolating and examining the individual human brain as if it were a self-contained information processor. It assumes that once we understand the individual, we can understand how cognition interacts with the environment. DCog's objection: this misses the point. Much of what we call 'intelligence' in everyday life is distributed across people and their environment. You cannot find the cognitive properties of a system by studying its components in isolation — you will only find properties that don't belong to the system.

Question

What is social visualization, and how does it connect to distributed cognition?

Answer

Social visualization extends InfoVis systems beyond a single user to include social groups — recognizing that data analysis is inherently social and contextual. Example: Many Eyes (IBM) was a platform where users could share and collaboratively discuss data visualizations. From a DCog perspective, this expands the cognitive system beyond one person + one computer to include other users, shared representations, and social coordination mechanisms. The whole group forms a richer cognitive system than any individual could.

Question

What is the 'reductionist' vs. 'environmental' perspective on cognition, and what does each imply for how we study HCI?

Answer

The reductionist perspective treats cognition as computation inside the individual brain; research focuses on abstract properties of isolated minds (e.g., 'working memory holds 7±2 items'). The environmental perspective (which DCog adopts) treats cognition as embodied, situated, and distributed — emerging from interaction between individuals and their social, cultural, and material environment. For HCI, this means studying users in context with their real tools, not in sterile lab conditions stripped of the artifacts they actually rely on.