Week 1 notes done!

Author: AustinTSchaffer

OMSCS/Courses/HCI/3.1 - Introduction to Methods.md

@@ -9,7 +9,97 @@ tags:
 
 > Novelty must have a purpose.
 
-- User-Centered Design
-	- design that considers the needs of the user throughout the entire design process
-	- you are not your user
-- 
+## User-Centered Design
+- design that considers the needs of the user throughout the entire design process
+- you are not your user
+- Principles
+	- based on an understanding of users, tasks, and environments
+	- users are involved throughout design and development
+	- the design is driven and refined by user-centered experiments
+	- the process is iterative
+	- the design addresses the whole user experience
+	- the design team includes multidisciplinary skills and expertise
+
+## Stakeholders
+- the user themselves is the primary stakeholder
+- secondary users are impacted by the output of the interface/tool
+- tertiary users are impacted by the secondary users
+
+![[Pasted image 20250515203406.png]]
+
+## The Design Lifecycle
+
+![[Pasted image 20250515205332.png]]
+
+We're going to discuss a 4-phase life cycle
+
+- need-finding / needfinding
+- design alternatives
+- prototyping
+- evaluation
+
+![[Pasted image 20250515205438.png]]
+
+![[Pasted image 20250515205522.png]]
+
+> Remember that this is a survey course. We're not going to become experts in any part of this feedback cycle.
+
+## Qualitative vs Quantitative Data
+
+- descriptions (qual)
+- measurements (quant)
+- preferences (qual/quant)
+- performance (quant)
+- quantitative = data that's summarized numerically (can be generated from qualitative data, narrow scope, (what))
+- qualitative = observations described or summarized non-numerically (broader, prone to biases, (how, why))
+- in HCI, we want to focus on both qualitative data and quantitative data
+
+### Types of Quantitative Data
+
+![[Pasted image 20250515205957.png]]
+
+- nominal
+	- categorical
+	- subcategories
+		- single nominal (radio buttons)
+		- multi-nominal (check boxes, can select multiple options)
+		- nonbinary
+		- binary
+- ordinal
+	- categorical with explicit ordering
+	- "on a scale of 1 to 5", where the numbers are labeled
+	- subcategories
+		- binary
+		- nonbinary
+- interval
+	- interval between different options are the same (e.g. times, temperatures)
+	- measurements that have "no zero point"
+	- subcategories
+		- discrete
+		- continuous
+- ratio
+	- like interval data, but has a zero point
+	- subcategories
+		- discrete
+		- continuous
+
+![[Pasted image 20250515210345.png]]
+
+### Types of Qualitative Data
+
+![[Pasted image 20250515210416.png]]
+
+> Determined by how it's gathered.
+
+- No upper bound on the space of possible qualitative data formats
+- Coding is the process of converting qualitative data to quantitative data
+
+![[Pasted image 20250515210532.png]]
+
+
+
+
+
+
+
+

OMSCS/Courses/HCI/3.2 - Ethics and Human Research.md

@@ -5,3 +5,53 @@ tags:
 ---
 # 3.2 - Ethics and Human Research
 
+- origin of research ethics
+- basic ethical considerations
+- Institutional Review Board (IRB) origins
+
+## Origin of IRBs
+
+- Milgram experiment
+- Tuskegee syphilis experiment
+- Stanford prison experiment
+- National Research Act
+	- Benefits to society must outweigh risks to participants
+	- participants must be selected fairly / ethically
+	- participants should be able to back out
+	- perception of coercion
+
+## IRB Basics #Research
+
+- A **protocol** is a description of a particular research project, waiting to be or already approved by the IRB.
+- **Human subject interaction**
+- we need scientific justification for the number of subjects
+- inclusion and exclusion criteria
+	- "There are no exclusion criteria beyond the inclusion criteria."
+- recruitment procedures
+- is compensation provided?
+	- real research probably involved compensation
+	- graduate research usually doesn't include compensation
+- direct interaction? we need consent procedures
+	- written consent?
+	- waiver of consent?
+		- does the observation affect the people from whom the data is generated?
+		- is the data anonymized?
+	- waiver of documentation of consent?
+- is deception or concealment proposed?
+- data management
+	- how do we keep the data safe?
+	- how do we safeguard the data?
+- is there DoD involvement?
+- conflicts of interest?
+- primary investigator?
+
+## Research Ethics and Industry
+- companies will do research on their users
+- infamous "emotional contagion" paper by Facebook
+
+
+
+
+
+
+

OMSCS/Courses/HCI/3.3 - Needfinding and Requirements Gathering.md

@@ -5,3 +5,181 @@ tags:
 ---
 # 3.3 - Needfinding and Requirements Gathering
 
+> Try to find out what the user really needs.
+
+- defining questions about the user
+- generating answers about the user
+- formalizing models of the user
+- data inventory
+	- who are the users?
+	- where are the users?
+	- what's the context of the task?
+	- what are their goals?
+	- what do they need?
+	- what are their tasks?
+	- what are their subtasks?
+- the problem space
+- user types
+	- adults
+	- professionals
+	- hobbyists
+	- gym rats
+
+## Avoiding Bias in Needfinding
+
+- confirmation bias
+	- "We see what we want to see."
+	- looking for signs that you're wrong
+	- testing beliefs empirically
+	- involve multiple people
+- observer bias
+	- we might accidentally phrase questions in a way that elicits the responses we're hoping for
+	- separate experimenters with motives from the participants
+	- heavily script interactions with the users
+	- have someone else review interview scripts and surveys for leading questions
+- social desirability bias
+	- people tend to be nice, people want to help
+	- hide what the socially desirable response is
+- voluntary response bias
+	- people with strongest opinions are often the loudest responders
+	- we risk oversampling extreme views
+	- limit the amount of content available before they view the survey
+	- confirm conclusions with other methods
+- recall bias
+	- people aren't good at recalling what they did/felt
+	- study tasks in context
+	- conduct interviews during task itself
+
+## Observation
+### Naturalistic Observation
+> Observing people in their natural environment.
+
+quick tips
+- take notes
+- start specific, then abstract
+	- individual actions first
+	- generalize later
+- spread out your sessions
+	- shorter sessions are better
+- find a partner
+- look for questions
+
+### Participant Observation
+> Try putting yourself in the shoes of the user.
+
+- you are not your user
+- don't overrepresent your own opinions and experiences
+
+### Hacks and Workarounds
+> look at the hacks that users presently employ
+
+### Errors
+> We can use errors to learn about a users mental model.
+
+Looking for mistakes made by the user can be useful.
+
+### Apprenticeship Approach
+> We might need to become experts ourselves in order to build prototypes.
+
+- integrate yourself into the domain
+- train under your users
+
+### Interviews and Focus Groups
+> Asking your users for feedback directly.
+
+- Focus on the 6 W's (who, what, when, where, why, how)
+- Be aware of bias. Don't predispose the user to say what you want them to say.
+- listen. Participant should do vast majority of the talking
+- organize the interview
+	- introduction
+	- ...
+	- summary
+- practice
+	- rehearse your questions
+
+### Think-Aloud
+- give interviewees voice recorders
+- this is significant observation, may affect user behavior
+
+### Surveys
+- larger, quicker responses
+- low cost
+- quick setup and turnaround
+- not as thorough as in-person interviews
+- tips
+	- less is more (minimum number of questions. questions should be easy to respond to)
+	- be aware of bias
+	- tie questions to the data inventory
+	- test out the survey
+	- iterate
+- good questions
+	- be clear
+		- make sure user understands what you're asking
+		- avoid overlapping ranges
+		- provide extra detail when in doubt
+		- timebox frequency-based questions
+		- at odds with "be concise"
+	- be concise
+		- ask questions in plain language
+		- at odds with "be clear"
+	- be specific
+		- avoid questions on super big ideas
+		- avoid "double-barrel" questions (questions that can be rephrased as 2 questions)
+		- avoid questions that allow internal conflict (questions that have nuanced answers should be broken up into smaller, more specific questions)
+	- be expressive
+		- emphasize the user's opinions
+		- use ranges instead of yes/no questions
+		- give levels of frequency or agreement
+		- allow multiple selections when possible
+		- allow users to add nominal categories
+	- be unbiased
+		- allow users to add nominal categories
+		- see above for more tips on reducing bias
+		- allow for open-ended responses
+		- avoid leading questions
+		- avoid loaded questions
+	- be usable
+		- provide a progress bar
+		- don't make it too long
+		- make your page lengths consistent
+		- order questions logically
+		- alert users about unanswered questions
+		- don't force users to answer all questions 
+		- preview the survey yourself
+
+### Other Data Gathering Methods
+- critiquing existing interfaces
+- go look at product reviews
+- get some logs of user interaction that have already been aggregated
+
+## Iterative Needfinding
+![[Pasted image 20250515234341.png]]
+
+- Draw double-ended arrows between all pairs of nodes
+
+## Revisiting the Inventory
+- pay attention to places where the data seems to conflict
+- revisit these questions
+	- who are the users?
+	- where are the users?
+	- what's the context of the task?
+	- what are their goals?
+	- what do they need?
+	- what are their tasks?
+	- what are their subtasks?
+
+## Represent the need
+
+![[Pasted image 20250515234612.png]]
+
+![[Pasted image 20250515234627.png]]
+
+## Define our Requirements
+- functionality
+- usability
+- learnability
+- accessibility
+- compatibility
+- compliance
+- cost
+- ...