Human Error? No, Bad Design Chapter 5 Human Error? No, Bad Design

Today’s Topics Humans, systems, and error More on data gathering Designing for communication and coordination

Part 1: Humans, Systems and Error Root cause analysis for error Why did Air Force stop at human error? Do not view the system being designed as one combining human and computer/machine actors (e.g. sociotechnical or cyber-human system) The five whys method Humans are good at creative and adaptive behavior. The more mechanistic the activity is required to be, … Deliberate Violations Occur in all types of contexts (e.g. driving, officework, security, …) Sometimes because official procedures are developed for legal or accountability reasons In many situations, work cannot get done without exceptions

Errors: Slips and Mistakes Slips: the goal is correct but the actions not done correctly Action based – inaccurate performance of planned action Memory lapse – forgetting to take a planned action Mistakes: the goal or plan is wrong Rule based – wrong plan due to poor planning (“wrong rule”) Knowledge based – wrong plan due to inaccurate/incomplete knowledge Memory lapse – forgetting at stage of goals, plans, or evaluation Examples from elevator Pressing button below correct one along with correct one Pressing the 1st floor button when going to the lobby

Slips Slips are more common for experts than novices Capture slips Result of more aspects of a task being handled unconsciously Capture slips When similar action sequences get confused Generally a more familiar or recent action sequence replaces another Example might be finishing a password with more common suffix More common for experts Description-similarity slips Performing an action with a similar item to the one intended Designers should reduce similar items and reduce number to choose from Mode errors Occurs when design has modes where the same user action results in different system actions Designers should try to avoid modes and make the mode visible when they are necessary Memory lapse slips Forgetting where one is in procedure (repeating or not performing steps) Often caused by interruptions Designers can show steps and indicate where one is or can create forcing functions enforcing prerequeisits

Mistakes Rule-based mistakes Knowledge-based mistakes Situation is misinterpreted so wrong rule is selected Correct rule is invoked, but it is faulty or did not account for current situation Correct rule is invoked, but the outcome is incorrectly evaluated Knowledge-based mistakes Occurs in situations requiring more general problem solving Can be due to a mistaken mental model Designers need to provide clear conceptual models Systems need to support users when they encounter previously unknown situations Cooperative problem solving system Memory-lapse mistakes Occur when user forgets previously gathered information about situation, or forgets the plan altogether

Social Issues and Error Social, time, and economic pressure lead to bad decisions Social pressure also can lead to errors not being reported Adding more responsible people can cause less attention to tasks Each may assume others are being more responsible Need to encourage reporting of errors to fix their causes Jidoka and Poka-Yoke NASA aviation safety reporting system

Detecting Errors is Non-Trivial Errors are generally preceded by a number of uncommon events that each by itself is not strongly indicative of error False alarm Response to correct action interpreted as failure May occur When there are conflicting indicators Due to different interpretations of plan/task Garden Path Incorrect action not discovered until later Losing context of original problem Machine can interpret action as correct for some alternative path Users can assume they know the process without machine Trivial breaches of understanding can become “fatal”

Designing for Error Warnings Computer systems can often add Checklists Need to be clear from one another and coordinated with one another Need to gain needed attention without being obnoxious/disabilitating Computer systems can often add Different visual cues and locations for different actions/features Sensibility checks on user actions Undo and multi-stage undo Checklists Common in aviation but not other industries Design of checklists is still a difficult task More generally, processes can focus on creating resilience The Swiss cheese model to minimize likelihood Notifications when features are lining up for error

Part 2: More on Data Gathering Interviews Questionnaires Observation Choosing and combining techniques

Four key issues Setting goals Relationship with participants Decide how to analyze data once collected Relationship with participants Clear and professional Informed consent when appropriate Triangulation Use more than one approach Pilot studies Small trial of main study

Data recording Notes, audio, video, photographs Notes plus photographs Audio plus photographs Video

Interviews Unstructured - are not directed by a script. Rich but not replicable. Structured - are tightly scripted, often like a questionnaire. Replicable but may lack richness. Semi-structured - guided by a script but interesting issues can be explored in more depth. Can provide a good balance between richness and replicability.

Interview questions Two types: Closed questions are easier to analyze ‘closed questions’ have a predetermined answer format, e.g., ‘yes’ or ‘no’ ‘open questions’ do not have a predetermined format Closed questions are easier to analyze Avoid: Long questions Compound sentences - split them into two Jargon and language that the interviewee may not understand Leading questions that make assumptions e.g., why do you like …? Unconscious biases e.g., gender stereotypes

Running the interview Introduction – introduce yourself, explain the goals of the interview, reassure about the ethical issues, ask to record, present any informed consent form. Warm-up – make first questions easy and non-threatening. Main body – present questions in a logical order A cool-off period – include a few easy questions to defuse tension at the end Closure – thank interviewee, signal the end, e.g, switch recorder off.

Enriching the interview process Props - devices for prompting interviewee, e.g., a prototype, scenario

Questionnaires Questions can be closed or open Closed questions are easier to analyze, and may be done by computer Can be administered to large populations Paper, email and the web used for dissemination Sampling can be a problem when the size of a population is unknown as is common online

Questionnaire design The impact of a question can be influenced by question order. Do you need different versions of the questionnaire for different populations? Provide clear instructions on how to complete the questionnaire. Strike a balance between using white space and keeping the questionnaire compact. Decide on whether phrases will all be positive, all negative or mixed.

Question and response format ‘Yes’ and ‘No’ checkboxes Checkboxes that offer many options Rating scales Likert scales semantic scales 3, 5, 7 or more points? Open-ended responses

Encouraging a good response Make sure purpose of study is clear Promise anonymity Ensure questionnaire is well designed Offer a short version for those who do not have time to complete a long questionnaire If mailed, include a stamped addressed envelope Follow-up with emails, phone calls, letters Provide an incentive 40% response rate is high, 20% is often acceptable

Advantages of online questionnaires Responses are usually received quickly No copying and postage costs Data can be collected in database for analysis Time required for data analysis is reduced Errors can be corrected easily

Problems with online questionnaires Sampling is problematic if population size is unknown Preventing individuals from responding more than once Individuals have also been known to change questions in email questionnaires

Observation Direct observation in the field Structuring frameworks Degree of participation (insider or outsider) Ethnography Direct observation in controlled environments Indirect observation: tracking users’ activities Diaries Interaction logging

Structuring frameworks to guide observation - The person. Who? - The place. Where? - The thing. What? The Goetz and LeCompte (1984) framework: - Who is present? - What is their role? - What is happening? - When does the activity occur? - Where is it happening? - Why is it happening? - How is the activity organized?

Ethnography (1) Ethnography is a philosophy with a set of techniques that include participant observation and interviews Debate about differences between participant observation and ethnography Ethnographers immerse themselves in the culture that they study A researcher’s degree of participation can vary along a scale from ‘outside’ to ‘inside’ Analyzing video and data logs can be time-consuming Collections of comments, incidents, and artifacts are made

Ethnography (2) Co-operation of people being observed is required Informants are useful Data analysis is continuous Interpretivist technique Questions get refined as understanding grows Reports usually contain examples

Other Observational Methods Direct observation in a controlled setting Think-aloud technique Indirect observation Diaries Interaction logs

Choosing and combining techniques Depends on The focus of the study The participants involved The nature of the technique The resources available

Data Gathering Summary Three main data gathering methods: interviews, questionnaires, observation Four key issues of data gathering: goals, triangulation, participant relationship, pilot Interviews may be structured, semi-structured or unstructured Questionnaires may be on paper, online or telephone Observation may be direct or indirect, in the field or in controlled setting Techniques can be combined depending on study focus, participants, nature of technique and available resources

Part 3: Designing for Conversation and Coordination Conversational mechanisms Coordination mechanisms Awareness mechanisms Examples of technologies designed to extend how people talk and socialise work together play and learn together

Conversational mechanisms Various mechanisms and ‘rules’ are followed when holding a conversation, e.g. mutual greetings A: Hi there B: Hi! C: Hi A: All right? C: Good, how’s it going? A: Fine, how are you? C: OK B: So-so. How’s life treating you?

Conversational rules Sacks et al. (1978) work on conversation analysis describe three basic rules: Rule 1: the current speaker chooses the next speaker by asking an opinion, question, or request Rule 2: another person decides to start speaking Rule 3: the current speaker continues talking

Conversational rules Turn-taking used to coordinate conversation A: Shall we meet at 8? B: Um, can we meet a bit later? B: Wow, look at him? A: Yes what a funny hairdo! Back channeling to signal to continue and following Uh-uh, umm, ahh

More conversational rules Farewell rituals Bye then, see you, yer bye, see you later…. Implicit and explicit cues e.g., looking at watch, fidgeting with coat and bags explicitly saying “Oh dear, must go, look at the time, I’m late…”

Breakdowns in conversation When someone says something that is misunderstood: Speaker will repeat with emphasis: A: “this one?” B: “no, I meant that one!” Also use tokens: Eh? Quoi? Huh? What?

What happens in technology-mediated conversations? Do same conversational rules apply? Are there more breakdowns? How do people repair them for: Phone? Email? IM? Texting?

Synchronous computer-mediated communication Conversations are supported in real-time through voice and/or typing Examples include video conferencing, VOIP, MUDs and chat Benefits include: Not having to physically face people may increase shy people’s confidence Allows people to keep abreast of the goings-on in an organization without having to move from their office Problems: Difficult to establish eye contact with images of others People can behave badly when behind the mask of an avatar

Asynchronous computer-mediated communication Communication takes place remotely at different times Email, newsgroups, texting Benefits include: Read any place any time Flexible as to how to deal with it Can make saying things easier Problems include: FLAMING!!! Message overload False expectations as to when people will reply

Coordination mechanisms When a group of people act or interact together they need to coordinate themselves e.g., playing football, navigating a ship They use: verbal and non-verbal communication schedules, rules, and conventions shared external representations

Verbal and non-verbal communication Talk is central Non-verbal also used to emphasize and as substitute e.g., nods, shakes, winks, glances, gestures and hand-raising Formal meetings explicit structures such as agendas, memos, and minutes are employed to coordinate the activity

Schedules, rules and conventions Schedules used to organize regular activities in large organizations Formal rules, like the writing of monthly reports enable organizations to maintain order and keep track Conventions, like keeping quiet in a library, are a form of courtesy to others

Shared external representations Common method used to coordinate collaborative activities, e.g., checklists, tables, to-do lists They can provide external information on: who is working on what when it is being worked on where it is being worked on when a piece of work is supposed to be finished whom it goes to next

Collaborative technologies to support coordination There are a variety of software tools designed to support scheduling, planning and coordinating e.g., group calendars, electronic schedulers, project management tools, and workflow tools Need to get balance between human and system control too much system control and the users will rebel too little control and the system breaks down

Awareness mechanisms Involves knowing who is around, what is happening, and who is talking with whom Peripheral awareness Keeping an eye on things happening in the periphery of vision Overhearing and overseeing - allows tracking of what others are doing without explicit cues

Lo tech awareness mechanism

Designing technologies to support awareness Provide awareness of others who are in different locations Early example was media spaces “extend the world of desks, chairs, walls and ceilings” (Harrison et al, 1997) Examples: Clearboard and Portholes

Portholes (Xerox PARC) Regularly updated digitized images of people in their offices appeared on everyone’s desktop machines throughout day and night Portholes, developed at Xerox PARC research labs (Dourish and Bly, 1992).The system presented regularly updated digitized video images of people in their offices from a number of different locations (in the US and UK). These were shown in a matrix display on people’s workstations. Clicking on one of the images had the effect of bringing up a dialog box, providing further information about that individual (e.g. name, phone number) together with a set of lightweight action buttons (e.g. email the person, listen to a pre-recorded audio snippet). The system provided changing images of people throughout the day and night in their offices, letting others see ‘at a glance’ whether they were in their offices, what they were working on and knowing who was around (see figure 10). Informal evaluation of the set-up suggested that having access to such information led to a shared sense of community.

Notification systems Users notify others as opposed to being constantly monitored Provide information about shared objects and progress of collaborative tasks examples: Tickertape, Babble

Elvin Elvin is a distributed awareness system that provides a range of client services (Segall and Arnold, 1997) It includes Tickertape, one of the first lightweight messaging systems

Conversation and Coordination Key Points Social mechanisms, like turn-taking, conventions, etc., enable us to collaborate and coordinate our activities Keeping aware of what others are doing and letting others know what you are doing are important aspects of collaborative working and socialising Many collaborative technologies systems have been built to support collaboration