CS 321 Exam 1 study guide

  1. Name the 5 interaction design principles.
    • affordance
    • consistency
    • constraints
    • feedback
    • visibility
  2. What is affordance?
    • Affordance is the idea that objects within an interface should have attributes that facilitate the user being able to perceive their purpose and functionality.
    • example from quiz: explain why the system arrow, the I-beam, and the crosshairs cursors are the appropriate mouse cursor to be activated in a particular location in terms of its affordance.
    • system cursor - arrow indicates pointing and selection (you point at what you wish to select) - also it is the default cursor and makes sense that it would be used when the others would not
    • I-beam - mimics appearance of the cursor in text and is in fact used to position the cursor to insert text.
    • crosshairs - connote positioning in 2-dimensions which leads one to believe it could be used to target non-textual items
  3. What is consistency?
    • An interface has consistency if the user uses similar elements and similar operations to perform similar tasks.
    • e.g., the iPhone interface for email appears similar to the iPhone's interface for voice mail. You perform the same operation to select an email to read as you do to select a voice mail to listen to.
  4. What is the principle of constraints?
    • It is a design principle that prevents users from making mistakes by restricting (or constrianing) the actions that they may perform. Unavailable options could be greyed out or simply not shown.
    • e.g., a system with poor constraints would display all options to all users, whether that user can access that functionality or not.
  5. What is feedback?
    • Feedback refers to sending the user information about what action has been done and what has been accomplished.
    • e.g., Nintendo Cat - this was meant as a tool to help users bet comfortable interacting with the Wii. The cat runs away from the cursor if it advances toward the cat too quickly. When "caught" the cat provides "helpful" tips for using the Wii. This is a novel way to provide tips about how to use the wii.
  6. What is visibility?
    • Visibliity refers to knowing the status the software and which choices are available to them (which priveleges have been granted).
    • e.g., in a car, the position of the controls is related to their function, so they are easy to find and use
    • * The more visible functions are, the more likely users will be able to know what to do next.
  7. What is a mental model?
    • A mental model is a collection of knowledge about how something works and a user's ideas about how something works based on how they interact with it.
    • e.g., when someone wants to make their home warm as quickly as possible, they may set the thermostat to the highest temperature thinking that their home will get warm more quickly that way.
    • That is, if they don't understand how a thermostat works.
  8. Name and describe three common metaphors (mental models) used in software design.
    • 1) servant (Clippy) - anthropomorphic (gives human qualities to an object,), 1st person, application-specific lingo (servant that helps with an application)
    • 2) desktop (Windows) - The computer screen is meant to evoke the idea of a desktop - things are "open" when you are working on them.
    • 3) tool metaphor - the interface provides the user with a set of tools to choose from to accomplish tasks (like making a drawing in Visio)
  9. Why are metaphoric interfaces used?
    Metaphors are supposed to help the user be intuitive about the interface.
  10. What are some problems with metaphoric interfaces?
    • 1) Computer software shouldn't be limited to emulating physical things. For example, a word processor could be written to emulate a typewriter, but that would seriously limit its usefulness.
    • Some of the best functionality provided by a word processor cannot be provided by a typewriter.
    • 2) Sometimes we don't want or need computers to emulate physical things - like the bell a typewriter sounds when you approach the end of a line.
    • 3) Sometimes the domains (computer interface and what it emulates) have common features that behave differently. For example, whitespace (tabs, line feeds, etc.) is handled differently in word processors and typewriters.
  11. What should a designer do when the user's mental model is very different from their own?
    • Speaking in terms of the general user, the designer should attempt to make the interface more closely match the user's mental model. The inner workings of the application don't need to be changed to match the user's mental model.
    • In other words, the designer develops a third model, called the represented model, that combines the simplicity of the user's mental model with the technical veracity of the developer's implementation model.
    • e.g. - Pizza Hut website. User - peruse, select, order, payment, credit card, delivery. Designer - menus, frames, icons, URLs, hyperlinks. Both: *security*
  12. Mental models are both __________ and ____________.
    • predictive and explanatory
    • *The user's mental model of an application enables the user to predict the consequences of interactions.
    • *The user's mental model of an application enables the user to explain the reason that events occur in the application.
  13. What is an implementation model?
    An implementation model can be represented as a flowchart. It is an accurate representation of the decision points and subtasks involved in a process.
  14. What are the design methods that support mental models? (Or how do we cater to the user's mental model)
    • 1) simplicity - include only needed functions - bury rarely used ones deeper to leave interface uncluttered.
    • - make frequently used functions easy to find
    • 2) familiarity - build on users' prior knowledge and experience
    • -allow users to start quickly and make immediate progress
    • -consistency in an interface reduces user's learning curve
    • 3) availability - provide the user with visual cues about what is available (people are better at recognition than recall)
    • 4) flexibility - a good interface will allow a user to choose from different interaction techniques (icons or pull-down menus, mouse or keyboard navigation, etc.)
    • 5) feedback - when the user has initiated an action the interface should let the user know that it received the command, and it should notify them when complete as well
    • -feedback that supports the user's mental model reinforces it
    • -feedback that doesn't support the user's mental model forces it to be changed
    • 6) safety - users should feel that if they try an action and it is not what they intended, they can undo it.
    • 7) perceived affordances - the application's interface gives users a sense of the system's affordances (if you see binoculars, you probably expect a search function)
  15. What do work models do?
    They provide concrete representations of the work of intended customers via diagrams.
  16. What are the 5 types of work models?
    • Flow (who?)
    • Artifact (what?)
    • Sequence (when?)
    • Physical (where?)
    • Cultural (how?)
  17. Describe the flow model.
    • It diagrams the relevant interactions between people.
    • Each person is a bubble - responsibilities and potential breakdowns are noted.
    • Objects important to the flow are indicated in boxes.
    • Indicate communication and interaction with arrows.
  18. Describe the cultural model.
    • It diagrams the policies, procedures, rules, laws, authorities, traditions, customs, practices, preferences, attitudes etc that influence the work.
    • Use bubbles to specify all entities (people, departments, businesses, etc) - annotate the relevant attitude(s) of each entity
    • Arrows indicate influences that entities have on each other
    • The amount that the bubbles overlap is proportional to the extent to which it influences the other entity.
  19. Describe the sequence model.
    • It diagrams the step-by-step structure of the work, including all intents, triggers, activities and strategies.
    • Specify the intent of each action.
    • Use arrows to demonstrate the order of steps, including branches and loops.
    • Specify the trigger of each sequence of actions.
  20. Describe the artifact model.
    • The artifact model reveals the structure and intent of the objects used in the work.
    • Illustrate the objects used in the work activity, emphasizing their structure and essential characteristics.
    • Use annotations to elaborate on the informal usage of each object (not readily apparent)
  21. Describe the physical model.
    it diagrams where the work is done, emphasizing how space is organized, where people are located, and how people and objects move (including data, input and output)
  22. What tool is used to consolidate the ideas and perceptions that the various design team members have accumulated?
    The affinity diagram
  23. How is an affinity diagram created?
    • Each team member uses her notes to re-record individual observations on post-it notes.
    • The team meets and places the post-its on a whiteboard.
    • Cluster the notes that appear to be related.
    • Rearrange the clusters into categories agreed upon by the team.
    • Label the categories.
    • Group each category hierarchically. (I don't think we talked about this much in class.)
  24. What is a persona?
    • It is a hypothetical archetype standing in for actual users to help drive decision making.
    • *not real people but represents real people
    • *not fictional - discovered by investigation (user interviews)
    • *realistic
    • *imaginary but rigorous and precise
    • *defined by needs and GOALS
  25. How do personas help designers?
    • *Help define measurable goals in relation to user's goals
    • *improve team dynamics by grounding interface design decisions
    • *focus the design team's communication
  26. List the information that is used to create a persona.
    • name
    • age
    • photo
    • personal information (family and home life)
    • work environment (tools used, working conditions)
    • technical proficiency and comfort level
    • pet peeves
    • attitudes
    • motivation for using technical product (end result - not just tasks)
    • information seeking habits and favorite resources
    • personal and professional goals
    • candid quotes
  27. How do you construct a persona from interview results?
    • 1) identify behavioral variables (use the interview results to identify relevant behavioral patterns)
    • 2) map interview subjects to behavioral variables (identify the degree to which a subject exhibited the variable)
    • 3) identify significant behavior patterns (identify sets of subjects who cluster in several variables.)
    • 4) synthesize characteristics and relevant goals
    • 5) check for completeness and redundancy
    • 6) expand description of attributes and behaviors
    • 7) designate persona types
    • ***I think really only the first 3 steps were focused on in class
  28. What are some problems with personas?
    • 1) designers develop personas who use product as intended, causing flaws to be missed
    • 2) designers focus on functionality rather than global look and feel
    • 3) culture lock - if the designer is very different from the persona's they may be unable to address the issues brought to light by that persona
  29. To understand user interfaces, the design team must understand the users' ______________.
    cognitive processes
  30. How do the users' cognitive processes impact design? Give an example.
    e.g. Nintendo DS - The dual screen interface was used by developers to show different things. A first person shooter game might have the 1st person perspective on one screen and the game domain map on the other.
  31. What is a mental model?
    The user's idea of how the application works.
  32. What is important about the user's mental model to the designer?
    The designer must take the user's mental model into account not in the design of the application's functionality, but in terms of the application's interface.
  33. The user's mental model of an application enables the user to ____________.
    • predict the consequences of interactions
    • e.g., browser back buttons
  34. The user's mental model of an application enables the user to ____________ .
    • explain the reason that events occur in the application
    • e.g. VRML navigation
  35. What is the developer's implementation model?
    It is a highly detailed description of all of the players and interactions in a process.
  36. What is an implementation model?
    A technically logical, truthful, and accurate representation of a process. It is not accessible (or usable) for users.
  37. What is the designer's represented model?
    A model that combines the simplicity of the user's mental model with the technical veracity of the developer's implementation model.
  38. List the 7 design methods that support mental models.
    • 1) simplicity - only needed functions
    • 2) familiarity - build on user's prior experience
    • 3) availability - provide visual cues of what is available rather than relying on user to find what is available
    • 4) flexibility - support alternate interaction techniques (keyboard and mouse)
    • 5) feedback - complete and continuous feedback should be given (hourglass, status bar) -
    • feedback can reinforce mental model or force it to be modified
    • 6) safety - users will try actions if they feel confident unintended actions can be undone
    • 7) perceived affordances - real-world representations of objects allow users to know at a glance what affordances are available (binoculars for search, magnifying glass, etc.)
  39. Name the 5 work models.
    • 1) flow (who)
    • 2) artifact (what)
    • 3) sequence (when)
    • 4) physical (where)
    • 5) cultural (how)
  40. What is shown in a flow model?
    The relevant interactions between people (bubbles), departments (bubbles), and things (boxes). Responsibilities and potential breakdowns should be shown. Communication and interaction is shown with arrows.
  41. What is shown in a cultural model?
    The culture (laws, traditions, customs, attitudes, preferences) that influence the work. Entities shown in bubbles, overlapping indicates the amount influence the top has on the bottom. Arrows indicate the influences the origin has on the destination.
  42. What is shown on the sequence model?
    The intent of each action, the steps required to take each action (arrows indicate sequence). The trigger (cause) of each action is shown.
  43. What is shown on the artifact model?
    The objects involved in the work and any functionality they provide that is not immediately obvious. Problems are indicated with red lightningbolts.
  44. What is shown on the physical model?
    Where the work is done - how the space is organized, where people are located, and how people and objects move.
  45. What is an interpretation session?
    A design team meeting held after a team member meets with a customer. New design ideas are recorded, work models are drawn.
  46. What are the 4 roles that should be filled during an interpretation session?
    • 1) interviewer
    • 2) moderator
    • 3) recorder
    • 4) rat-hole watcher - prevents team from getting sidetracked
  47. What tool is used to consolidate the ideas and perceptions that the various design team members have accumulated?
    Affinity diagrams. Team members record pertinent info on post its, gather it into related topics. Then groups are refined into final groups and titled.
  48. What is a persona?
    Personas are hypothetical archetypes standing in for actual users to help drive decision-making.
  49. What can be said about how real personas are?
    Personas are not real people but they are representative models of real people. They're not fictional; they are discovered through investigation. They are imaginary but defined rigorously and precisely. They are realistic, with names and personal details to flesh them out. They are defined by their NEEDS and GOALS.
  50. What is the purpose of personas?
    They provide a much needed focus to design projects - especially when used to supplement other user-centered design methods like usability testing.
  51. Name a few benefits of personas.
    • *help define measurable project goals in relation to users' goals
    • *improve team dynamics by grounding interface design decisions
    • *focus the design team's communication
  52. What details should be included in a persona?
    • *name
    • *age
    • *photo
    • *personal info (home and family life)
    • *work environment
    • *technical proficiency and comfort level
    • *pet peeves (esp. technical frustrations)
    • *attitudes
    • *motivation for using technical product
    • *information-seeking habits and favorite resources
    • *personal and professional goals
    • *candid quotes
  53. How can you construct personas from interview results?
    • 1) identify behavioral variables (freq. of photo taking, freq. of photo viewing, ...)
    • 2) map interview subjects to behavioral variables (user 1 freq. took photos but rarely viewed them, user 2 freq. took and viewed photos, ...)
    • 3) identify significant behavior patterns aka look for clusters (user 1 and 4 and users 2 and 6 are fairly similar in behavior variables)
    • 4) combine users
    • 5) check for completeness and redundancy
    • 6) flesh out details as needed
    • 7) designate persona types (heavy user, occasional and not technically proficient user, ...)
  54. What are a few problems with personas?
    • 1) focus - personas are often developed from the designer's view of the user - someone who will use the product as intended
    • 2) tunnel vision - look and feel are often ignored, instead concentrating on specific functional issues
    • 3) culture lock - personas may bring issues to light that the designers cannot relate to and therefore can't fully address
  55. Why should personas have goals?
    • The goals of the personas help to focus the design. Goals and tasks are not the same thing - tasks are done to accomplish goals.
    • *life goals may not help with the application you are designing but can help in some and help to flesh out the persona
    • *end goals are tangible outcomes that the design will produce - these can align with persona goals (i.e., accomplish tasks faster or with more accurate results)
    • *experience goals - how does the persona want to feel when they use the product? Secure? Not stupid?
  56. What are scenarios?
    Scenarios are types of situations designers expect personas to encounter when using their software. For example, a persona might take a trip to the Grand Canyon and take, organize and store lots of photos with their photo organization software.
  57. What are some benefits of using scenarios?
    • *assumption clarification - written scenarios help team members share and formalize assumptions so there are fewer surprises later
    • *full design exploration - scenarios provide a vehicle for exploring all of the important aspects of the interface.
    • *understanding the user - developing scenarios forces the team members to learn more about the product, the business, and the users
    • *providing context for reviews - features found in the interface are often clarified when using scenarios
  58. Why is prototyping used in design (as opposed to a textual description or diagrams of proposed interfaces)?
    • Prototyping is an event-driven approach which allows users to envision dynamics of the proposed system.
    • Text and diagrams are abstract and difficult to relate to the scenario the user will encounter when using the application.
  59. What are some advantages of prototyping?
    • *allows designers to get validation of usability of proposed system
    • *ensures a common understanding of what will and will not be accomplished with the proposed system
    • *provides an opportunity to clarify misconceptions between designers and users before significant investment in time and resources has been made
  60. What is the purpose of a low-fidelity prototype?
    It has some of the characteristics of the target product but is otherwise simple so that it can be produced quickly and broad concepts can be tested.
  61. What are some advantages of low-fidelity prototyping?
    • *fast - no coding
    • *quickly ID basic design flaws
    • *refine GUI pre-implementation
    • *allows multidisciplinary participation (b/c no coding required)
    • *encourages ideas from designers and users
  62. What are some disadvantages of low-fidelity prototyping?
    • *produces no code
    • *doesn't find all GUI problems
    • *incompatible with mere upgrades - when you already have a finished product it may not be worth the time and effort to low-fi prototype
    • *may seem unprofessional
    • *may be overwhelming for large systems - lots of screens
  63. What is wireframing?
    Wireframing is the modern-day approach to paper prototyping. Static versions of digital controls are laid out. Hyperlinks are used to connect the various versions of the interface's appearance.
  64. Why would HCI designers use wireframing rather than paper prototyping?
    It looks more professional and provides fewer distractions during presentation (from humans moving paper).
  65. Describe how a paper or low-fi prototyping session is conducted.
    • There are three participants:
    • *facilitator - leads the user through the walkthroughs. Answers questions, but doesn't actually help the user traverse the interface
    • *observer - takes notes and monitors video recording
    • *user - attempts to navigate interface
    • After the session:
    • 1) discuss where the user had problems
    • 2) indentify larger/deeper design issues and prioritize them
    • 3) discuss possible changes
  66. What is iterative refinement?
    • It is the strategy of refining the wireframe after each prototyping session and before the next one.
    • After early sessions make broad changes.
    • After later sessions make smaller changes.
  67. What is a high-fidelity prototype?
    It is quite close to the final product, with lots of detail and functionality.
  68. What are some advantages of high-fi prototyping?
    • *no delays for test users
    • *superior control over what the user can see
    • *resetting or undoing actions is possible and easy
    • *particularly useful for upgrading existing software
  69. What are some disadvantages of high-fi prototyping?
    • *curtails major changes
    • *not conducive to large design teams
    • *tends to make users feel that design is complete
    • *requires all designers to have technical backgrounds
  70. What can be said about prototypes as the design team iterates through prototypes?
    The structure and flow of the design are refined and the level of commitment to the design increases.
  71. Name three research methods HCI researchers employ to collect evidence.
    • 1) lab experiment - highly controlled but not native work env't
    • 2) field study - not highly controlled but native work env't
    • 3) survey - goog, generalizable results since survey allows greatest participation (quick to collect data)
  72. What are four factors you should seek to quantify when determining a system's usability?
    • 1) learnability (easy to learn)
    • 2) efficiency (fast)
    • 3) recoverability (errors few and recoverable)
    • 4) satisfaction (enjoyable to use)
  73. Name and describe 6 factors that affect the validity of usability tests/
    • 1) sample size (how many users are needed to ensure valid results)
    • 2) representativeness (how well does sample represent population)
    • 3) randomness (are non-participants fundamentally different than participants => invalid results?)
    • 4) data collection (how should the data be gathered)
    • 5) completion rate (how many participants successfully complete the assigned task during a usability test)
    • 6) task time (how long does a user spend on an activity during a usability test)
  74. What are the four major steps in a controlled experiment?
    • 1) start with a testable hypothesis (menu is faster to menu at bottom of screen than at top of screen)
    • 2) choose the independent variables to manipulate to test the hypothesis (e.g. y-position of screen)
    • 3) measure the dependent variables to test (time to access menu, error rate, user satisfaction)
    • 4) use statistical tests to accept or reject the hypothesis (analyze how changes in independent variables affected the dependent variables, and whether those effects were significant - indicating a definite cause and effect)
  75. What should you consider when designing an experiment?
    You should try to eliminate - or render harmless - the effect(s) of unknown and uncontrolled variables, in order to enable conclusions to be drawn regarding the effect of the independent variables on the dependent variables.
  76. What is a confidence interval?
    It is a range of values with a specific probability of containing the estimated value we seek.
  77. What are the three main factors that affect the confidence interval?
    • 1) confidence level - you choose this value - usually 90% or 95%
    • 2) variability - how much does the data fluctuate? - a data set with more variable data will have a wider confidence interval
    • 3) sample size - how much data do you have? - the size of the confidence interval has an inverse square root relationship with the sample size - to halve the confidence interval you must quadruple the sample size
  78. Given the same data, how would the 90% confidence interval compare to the 95% confidence interval?
    The 90% confidence interval will be significantly narrower than the 95% confidence interval. It does increase (double) the chance of making an error.
  79. What are the variables needed to compute the *completion rate* confidence interval?
    • *p-hat - the proportion of the sample that completed the task
    • *n - the sample size
    • *z (sub (1-(alpha/2))) - the critical value from the *normal distribution* for the confidence level
  80. Which distribution is used to calculate the confidence value when the data is discrete binary data?
    the normal distribution
  81. Which distribution is used to calculate the confidence value when the data is continuous?
    The t-distribution
  82. What is the difference between the t-statistic and the z-statistic?
    • Both of these statistics answer the question "What is the likelihood of having a sample mean given that the population mean is mu?"
    • The z-statistic is the number of standard error units that a sample's mean is from the population's mean - ASSUMING a NORMAL distribution
    • The t-statistic is the estimated number of standard error units that a sample's mean is from the population's mean and uses the sample's standard deviation
  83. What does the t-statistic allow you to do?
    It allows researchers to use sample data to test hypotheses about an *unknown population mean*.
  84. How should you choose between using the t-statistic and the z-statistic?
    • When your sample size is large (>= 30), you can assume that your sample is normally distributed and use the z-statistic.
    • When your sample size is small (< 30), you need to use the t-statistic that utilizes the actual sample size (minus 1, called degrees of freedom) in the calculation.
  85. What can be said about t-distribution curves and z-distribution curves?
    • z-distribution curves are classic bell shaped curves. They are used when working with large sample sizes of binary data.
    • t-distribution curves are "flattened out" bell curves. There middle of the curve is lower and the curve is flatter overall. This is because the the mean is composed of more complicated information (continuous data), so the distribution wont be as sharp as in a z-distribution.
  86. What must be taken into consideration when calculating confidence intervals for task times?
    While there is effectively a floor for task time (everything must take some time), there is no ceiling for task time. There are users who will take an exceptionally long time to complete a task. Therefore, the mean of a set of completion times will be skewed high.
  87. What statistical technique can you use to combat the positive skew of completion time data?
    You can use the geometric mean (rather than the arithmetic mean). To do this, calculate the average of the logarithmic values of a data set, then convert back to a base 10 number.
  88. Describe one- and two-tailed tests and when you would use each one.
    • A one-tailed test tests the probability of a value comparing one threshold (like completion time < 1 min). Therefore the entire area under the curve above the point at 1 min is the probability you are looking for.
    • The two-tailed test tests the probability of a value being between two values (like the number of calories for a meal replacement bar is between 180 and 200).
    • It is important to note that when looking for confidence of x% in a 1 tailed test, x% will be excluded on one side of the curve.
    • In a 2 tailed test, x/2 will be excluded on both sides of the curve.
  89. When you are testing to compare to a benchmark (like at least a particular satisfaction score was achieved), what type of test would you conduct and why?
    A one-tailed test. You are looking for a minimum threshold, not a range.
  90. What is the binomial distribution?
    It is the discrete probability distribution of the number of successes in a sequence of independent yes/no experiments.
  91. What test should be used when determining whether a benchmark is met with regard to completion rates when there is a small (<30) sample size?
    The binomial distribution should be used.
  92. What does the sum accomplish in the binomial distribution function?
    The binomial distribution function sums the probabilities of each completion rate percentage (e.g. if you want >= 80% completion rate it sums the probability of 80% completion rate + 81% + 82% ... + 100% completion rate probability)
  93. What is done to the binomial distribution when your sample is of a larger size?
    If your sample includes at least 15 successes and at least 15 failures, a normal approximation to the binomial distribution is used.
  94. Go through all statistical formulas and get a general idea of what the variables are used for.
  95. What statistical tool should you use when you want to compare an interface's satisfaction score (e.g., from a questionnaire) to a benchmark?
    Here we are dealing with a sample so we use a t-distribution. We don't know if our sample approximates a normal distribution or not.
  96. When dealing with benchmarked task times, what statistical tools/tests are used?
    • Because we are dealing with task times, we want to use geometric means (convert to log, average, convert back to base 10 - to normalize the data).
    • Use the t-test because we will use our sample data to approximate the population data.
Card Set
CS 321 Exam 1 study guide
exam 1 review