Skip to main content
University of Washington UW W Logo Human Centered Design & Engineering Wordmark Logo

Research

Kate Starbird's Research Group Archive

The following research group descriptions are archived because they are no longer offered, the faculty member is on sabbatical, or the group is taking a break. Please contact the faculty member or an advisor to learn more about these groups.

Spring 2024

Dehumanizing and Problematic Imagery about Latin American* Migration in the 2024 Presidential Election

Run by: Nina Lutz, PhD Student, HCDE
Supervisor: Professor Kate Starbird

This DRG aims to develop methods, research questions, literature reviews, and tools to analyze rumors around harmful rhetoric at the US-Mexico land border. In particular, this DRG will focus on visual media (memes, short and long-form videos, photos, infographics, etc) that aims to dehumanize and spread rumors about and within intersectionally marginalized migration populations, particularly Latin American migrants and refugees at the US-Mexico land border. 

Students will first work together to create a literature review of related work. Then they will divide into two groups to design and embark upon either a primarily qualitative or primarily quantitative research study. Students will have an opportunity to continue this research in Fall of 2024 to complete the research study and to be on subsequent publications from this research.

Capacity: 8 students 

Student Requirements:

  • Strong interest in visual media
  • Strong interest in mis/disinformation
  • Strong interest in social media 
  • Basic understanding or willingness to learn about the American Presidential Election and American Legislative Government Structure
  • Power users of social media are encouraged to apply (TikTok, Instagram, X, etc)
  • Spanish language proficiency is a huge plus but not a requirement

The Quantitative Team will meet for 1 hour a week. The Qualitative Team will have a 2 hour weekly meeting. These meetings will be scheduled based on the students admitted. 

* We will for this DRG focus on migration and refugee populations, particularly from Latin America (including Haiti and the Caribbean Islands) but open to other migrations based on real-world events (ie, individuals from the Middle East). 

Winter 2024

Transgender Science Communication DRG

Organizers: Andrew Beers, albeers@uw.edu, Izzi Grasso, grassoi@uw.edu 
Faculty Sponsors: Dr. Kate Starbird, HCDE, Dr. Emma Spiro, iSchool

The last five years have seen an escalating number of legal challenges towards transgender’s people, and particularly transgender children’s, right to access healthcare. Many of these challenges have been couched in scientific terms, suggesting in misleading fashion that gender-affirming care is neither safe nor effective. In this Directed Research Group (DRG), we conduct an extended case study of one of the first and most severe proposed laws to restrict access to children’s healthcare in the United States, and subsequent legal challenges to this law. Particularly, we focus on expert testimonies submitted by both the law’s defenders and its detractors, which contain extensive reviews of the supposed evidence for or against providing gender-affirming care for children. Our goal is, through an analysis of the citations offered in these expert testimonies, to understand how disparate scientific, journalistic, and activist information sources are collected and mobilized to define the legal terms of transgender’s people’s access to healthcare. More broadly, we seek to understand the long-term collaborative work that goes into producing evidence disinformation campaigns, and how that work is mobilized into the legal sphere.

The specific work of this DRG is three-fold. The bulk of our energy will be spent reading and systematically annotating citations of the expert testimonies submitted in this case, and continuing this process into the past to create a genealogy of information sources regarding transgender healthcare in the last decade. The second part of this work will be, over time, to create qualitative memos recording our developing insights as we annotate this literature. The third part will be to conduct a weekly journal club where we read and discuss prior published work relating to transphobic and scientific disinformation.

We’re interested in students with all sorts of backgrounds and experience levels, who are passionate about this topic and interested in studying issues of disinformation from a research-driven perspective.

Winter 2024

Dehumanizing and Problematic Imagery about Latin American* Migration in the 2024 Presidential Election

Run by: Nina Lutz, PhD Student, HCDE
Supervisor: Professor Kate Starbird

This DRG aims to develop methods, research questions, literature reviews, and tools to analyze rumors in real-time during the 2024 U.S. Election Cycle. In particular, this DRG will focus on visual media (memes, short and long-form videos, photos, infographics, etc) during the 2024 Primary Elections. To limit our problem space, we will center questions on visual media that aims to dehumanize and spread rumors about and within intersectionally marginalized migration populations, particularly Latin American migrants and refugees. 

Students will first work together to create a literature review of related work. Then they will divide into two groups to design and embark upon either a primarily qualitative or primarily quantitative research study. Students will have an opportunity to continue this research in Spring and Fall of 2024 to complete the research study. Our goal is that each research study will be published at an academic conference as a paper or poster. 

Capacity: 8 students 

Student Requirements:

  • Strong interest in visual media
  • Strong interest in mis/disinformation
  • Strong interest in social media 
  • Basic understanding or willingness to learn about the American Presidential Election and American Legislative Government Structure
  • Power users of social media are encouraged to apply (TikTok, Instagram, X, etc)
  • Spanish, Hebrew, or Arabic language proficiency is a huge plus but not a requirement

Questions? Email ninalutz@uw.edu 

* We will for this DRG focus on migration and refugee populations, particularly from Latin America (including Haiti and the Caribbean Islands) but open to other migrations based on real-world events (ie, individuals from the Middle East). 

Autumn 2023

Analyzing TikTok user behavior changes

This DRG will be run by HCDE PhD Student Joey Schafer with guidance from Professor Kate Starbird.

In this DRG, we will be analyzing social media trace data  from TikTok users, to understand how their use of the platform has changed.

We are looking for 2-4 new students who have experience with social media (especially TikTok), qualitative coding, and/or visual data analysis for this 2-credit DRG. Students from all academic levels are invited to apply and participate in this project. DRG students will meet for approximately 1 hour weekly and are expected to contribute an additional approximately 5 hours per week. The beginning of the DRG will focus on qualitative coding of previously-collected data, and further analysis as well as other components of the research process, such as interviewing, cleaning transcripts, analyzing transcripts, and writing up research findings will occur in the later portion of the DRG.

If you have any questions about this project please contact schaferj@uw.edu or kstarbi@uw.edu.  

Autumn 2023

Negative Affect Research Group

Led by HCDE PhD Candidate Andrew Beers and advised by Associate Professor Kate Starbird.

Social media metrics privilege “positive” affect. Famously, Facebook, Twitter, TikTok, and other platforms do not have “dislike” buttons, and consequently many social media datasets available to researchers are concerned with positivity: likes, shares, follows, endorsements, etc. And yet public debate often centers around social media’s potential fornegative affect. Many are concerned about the extent and severity of targeted social media harassment, which terrorizes individuals and can mute the online expression of entire communities. Group expressions of disgust, contempt, or ridicule seem to typify everyday interaction on social media platforms, causing a moral panic around “cancel culture” for some and a discourse around the beneficial effects of shaming and consequences for others. Some platform users seem to intentionally cultivate negative affect, transforming their controversial statements into social capital via the influencer economy. Even the utopian dream, rarely reached, of the internet as a forum for productive debate presupposes that disagreement would be a common feature of its everyday usage.
 
This research group seeks to understand how negative affect is expressed on popular social media platforms, what the consequences of those expressions are, and popular debates on the prospect for designing for (or against) negative affect. Half of our time will be committed to reading and discussing published research which investigates negative affect on social media from a variety of methods and perspectives. The other half will be committed to annotating a large dataset of quote tweet interactions on Twitter between popular United States political accounts, including politicians, journalists, activists, influencers, and media outlets. By annotating quote tweet interactions, frequently noted for their negative affect, we aim to both A) develop a repeatable codebook for and a deeper understanding of negativity online and B) generate a seed dataset which can be used to automatically classify future data and better understand the qualities of negativity at scale.

Spring 2023

Cross-platform influencers research group

We are looking to recruit 3-4 students to participate in a Directed Research Group led by Center for an Informed Public PhD students Kristen Engel and Morgan Wack, and advised by Professors Kate Starbird (HCDE) and Emma Spiro (iSchool). 
 
This project will examine how misinformation about elections spreads across social media platforms. Specifically, we will focus on the accounts of individual users that spread large quantities of false and misleading information during the 2020 and 2022 elections to study how their posts differ depending on the social media platform they are using. Students will work with CIP researchers to manually identify and categorize social media posts and comments via qualitative coding methods. 
 
Students would enroll in HCDE 496 (for undergraduates) and HCDE 596 (for graduate students) for 2 credits during the spring quarter and be expected to spend 6 hours per week on the project. Interested students should complete the application form below. Selected students will be given an add code for the course. You can find a description of the study aims and desired student qualifications below. 
 
Project description (2 credits/6 hours per week) 
Despite growing user-bases and influence in American politics, limited data currently exists regarding the spread of misinformation outside of mainstream platforms. This project will enable look to fill this gap by collecting and comparing the actions of “repeat spreaders” of misinformation across four text-based social media platforms: Truth Social, Parler, Gettr, and Twitter. To generate novel insights on both differences in the actions of these repeat spreaders and their interaction with distinct platform users, this project will leverage new data from repeat spreaders of misinformation on Twitter linked to the 2020 and 2022 U.S. Elections. Using this data, we will collectively be able to answer several interesting questions, including:

  • How does the content and subject matter of prominent spreaders of misinformation differ by platform?
  • How does engagement with misinformation-linked posts differ based on the platform policies and user bases?
  • Do different strategies implemented by prominent spreaders predict the virality of misinformation across platforms?
  • How do users on non-mainstream platforms engage with misinformation compared to mainstream social media? 

Qualifications: Students with prior qualitative coding experience, experience developing qualitative coding schemas, and a general understanding of online misinformation, U.S. politics, or non-mainstream social media platforms will be strong candidates for this project.

Spring 2023

Negative Affect Research Group

Led by HCDE PhD Candidate Andrew Beers and Associate Professor Kate Starbird

Social media metrics privilege “positive” affect. Famously, Facebook, Twitter, TikTok, and other platforms do not have “dislike” buttons, and consequently many social media datasets available to researchers are concerned with positivity: likes, shares, follows, endorsements, etc. And yet public debate often centers around social media’s potential for negative affect. Many are concerned about the extent and severity of targeted social media harassment, which terrorizes individuals and can mute the online expression of entire communities. Group expressions of disgust, contempt, or ridicule seem to typify everyday interaction on social media platforms, causing a moral panic around “cancel culture” for some and a discourse around the beneficial effects of shaming and consequences for others. Some platform users seem to intentionally cultivate negative affect, transforming their controversial statements into social capital via the influencer economy. Even the utopian dream, rarely reached, of the internet as a forum for productive debate presupposes that disagreement would be a common feature of its everyday usage.

This research group seeks to understand how negative affect is expressed on popular social media platforms, what the consequences of those expressions are, and popular debates on the prospect for designing for (or against) negative affect. Half of our time will be committed to reading and discussing published research which investigates negative affect on social media from a variety of methods and perspectives. The other half will be committed to annotating a large dataset of quote tweet interactions on Twitter between popular United States political accounts, including politicians, journalists, activists, influencers, and media outlets. By annotating quote tweet interactions, frequently noted for their negative affect, we aim to both A) develop a repeatable codebook for and a deeper understanding of negativity online and B) generate a seed dataset which can be used to automatically classify future data and better understand the qualities of negativity at scale.

Spring 2023

Understanding TikTok User Behavioral Changes After Sudden Bursts of Increased Attention

Led by HCDE PhD Student Joey Schafer with guidance from Professor Kate Starbird

Current social media platforms like TikTok facilitate sudden convergence on particular users or videos, giving them a much larger audience than they were previously accustomed to. This sudden increase in attention can be quite disorienting for users, who are suddenly thrust into a much more visible, public online space. In this DRG, we will be interviewing U.S. TikTok users who have experienced what they self-identified as viral events, in order to understand what this experience was like and the impacts that this had on their use of social media platforms.

We are looking for 2-4 students who have experience with social media (especially TikTok), interviewing, and visual data analysis for this 2-credit DRG. Students from all academic levels are invited to apply and participate in this project. DRG students will meet for approximately 1 hour weekly and are expected to contribute an additional approximately 5 hours per week, such as through background readings, participant scheduling and interviews, transcribing interview recordings, and memoing on themes found in interviews.

Winter  - Spring 2023

Examining the Spread of Election Rumors Online

Led by Sukrit Venkatagiri (Postdoctoral Scholar), Emma Spiro (iSchool), and Kate Starbird (HCDE)

The 2022 midterm elections were the focus of a wide range of rumors and conspiracy theories. In Autumn, in a first stage of research, our team at the Center for an Informed Public identified hundreds of different claims about the election that were false, misleading, and/or unsubstantiated. Now, in a second stage, our team aims to classify these different claims, identify social media posts from a variety of platforms related to these claims, and analyze social media content around these claims to answer a variety of research questions about how rumors spread online during the 2022 election period.

We are looking for students with a range of different skills. First and foremost, all students must have familiarity with social media, an interest in the processes and procedures around elections, and a willingness to engage in qualitative analysis of social media posts. We are also looking for students who, in addition to those interests/skills, have experience with data science (writing code to analyze data), network science, visualization, and statistical/machine learning. We encourage students with journalism and political science backgrounds to apply as well.

Spring 2022

Investigating Content Integrity and Disinformation Risks Across Wikipedia Language Editions

Facilitated by HCDE PhD student Zarine Kharazian, with guidance from faculty advisors Kate Starbird and Benjamin Mako Hill 

This directed research group will conduct a qualitative interview study to better understand “content integrity” risks across Wikipedia language editions, particularly non-English Wikipedia projects. Specifically, we are interested in understanding whether some Wikipedia language editions are more vulnerable to disinformation campaigns and ideologically-motivated editing than others, and why. 

A small team of students will conduct semi-structured interviews with Wikipedia stakeholders and community members, including editors of specific language editions and contributors involved with various cross-wiki monitoring activities. Students will also transcribe and qualitatively code the interviews. While interviews will be conducted in English, students with foreign language skills are strongly encouraged to apply, as there may be opportunities to supplement the interview data with analyses of digital trace data from various Wikipedia language projects.

Students will conduct 1-2 interviews a week with participants over Zoom (scheduling of interviews TBD). Most interviews will last about one hour. Additionally, the DRG will meet in person once a week for 1-2 hours. Outside of meeting times and interviews, the expected time commitment per week is approximately 3 hours — for a total of 6 hours per week. Students should register for 2 credits of HCDE 496/596.

We are looking for students with a range of skills. This DRG would be a great fit for those who have one or more of the following: 

  • Experience with qualitative interviewing
  • Foreign language skills (specifically strong reading ability in a language other than English)
  • Interest in or experience with Wikipedia or other peer production platforms
  • Background in political science, communication, or information studies

Spring 2022

Evaluating Disaster Adaptation through Coding Short Form Internet Videos

This research group is studying social media data from the 2021 Texas Freeze Power Crisis to determine how people adapt to natural disaster events. This DRG is open to both undergraduate and graduate students. Students will qualitatively code short form Internet videos, such as TikToks and Instagram stories, shared in tweets that were posted during the 2021 Texas Power Crisis. We are looking for students who have had some experience with qualitative coding text or other data (experience coding videos is not required). We will meet a total of 1-2 hours per week. Outside of meeting times, the expected time investment per week is approximately 4 hours. This DRG will be 2 credits.

Please note that the videos students will be coding may potentially include distressing content of people experiencing a crisis event. Please keep this in mind if you decide to apply for this DRG. 

Winter 2022

Evaluating Disaster Adaptation through Coding Tiktok Videos

This research group will analyze social media data from the 2021 "Texas Freeze" Power Crisis to determine how people adapt to disaster events. This DRG is open to both undergraduate and graduate students. Students will develop a coding scheme and qualitatively code TikTok videos (shared in tweets) from the 2021 Texas Power Crisis. Some students may also assist with automating detection of TikTok videos from tweets. We are looking for students who have had some experience with qualitative coding text or other data (experience coding videos is not required). Some experience with object recognition within videos would be a plus, but is definitely not required. Outside of meeting times (~2 hours per week), the expected time investment per week is approximately 4 hours. The course will be 2 credits.

Please note that the videos students will be coding may potentially include distressing content (e.g. of people experiencing a crisis event). Please keep this in mind if you decide to apply for this DRG. 

This DRG is closed for autumn and no longer accepting applications. If you have questions, please email Shengzhi Wang at shengzw@uw.edu or Alexa Schlein at alexa412@uw.edu.

Winter 2022

Designing for “Rapid Response” to Electoral Misinformation

This directed research group will design innovative methods and systems to detect and categorize online claims about election integrity. This work is situated within a larger project that aims to rapidly detect, collect, process, and analyze public data from Twitter and other online sources — to uncover misinformation and influence operations related to elections. A primary goal of this DRG is to design processes and tools to support this work going forward. We are looking for students with a range of skills, but most important is a willingness to engage closely with the data (i.e. read and categorize a large number of social media posts), and a comfortability with working with developing technology. Experience with qualitative coding and human-centered design methods are a plus, but not required. Power users of social media are welcome. Students with an interest in journalism, political science, policy, discourse analysis, online activism, and the design of social media platforms are all encouraged to apply. This DRG is at capacity and no longer accepting applications.

Autumn 2021

Misinformation and the 2020 U.S. Election

This directed research group will study how people mobilized around misinformation during the 2020 U.S. election cycle. Students will work in a small team to qualitatively code and quantitatively analyze public data collected from Twitter. We are looking for students with a range of skills, but most important is a willingness to engage closely with the data (i.e. read and categorize a large number of social media posts). Experience with qualitative coding is a plus, but not required. Students with an interest in journalism, political science, policy, online activism, and the design of social media platforms are all encouraged to apply. The DRG will take a hybrid approach to in person and online work and learning. We will meet in person at the beginning of the quarter and will have the option to switch to remote work as the quarter progresses.  

Winter 2021

Tracking The Rise and Fall of Mask-Related COVID-19 Theories

Led by

  • Andrew Beers, 2nd-year PhD student in HCDE
  • Sarah Nguyen, 1st-year PhD student in the iSchool
  • With guidance from faculty advisers Kate Starbird, HCDE, and Emma Spiro, iSchool.

What

The Center for an Informed Public at UW has been archiving posts from Twitter users engaging in arguments about whether or not to wear a mask. We’ve spent this last year qualitatively analyzing a subset of ~5,000 posts from these arguments, to understand the many different theories that Twitter users are employing to argue for or against public mask mandates. We found that users have a variety of theories about masks — about their ability to block virus particles, their potential harm to users, when they should be used, and more — and defend these theories aggressively using the language of science and a wide variety of external links, images, and videos.

This Winter quarter, we would like to expand this project to our full dataset, which currently numbers in the tens of millions of posts and is growing every day. Specifically, we’re aiming to use automated techniques to classify this larger tweet dataset of arguments into the theories we identified in the first stage of this project. We want to see how the popularity of certain theories about masks changed over time, and whether they responded to external events, such as the publication of a new scientific paper, a change in the severity of the pandemic, or a comment by a politician.

Who

We’re looking for up to three students with an interest in public health communication, misinformation, and/or natural language processing. We are interested in students familiar with Python and with the Twitter platform, but we strongly encourage students without prior experience in either of these areas to also apply.

What You’ll Be Working On

We’re looking for a relatively small team, and we can see several tasks for this project over the course of the quarter:

  • Qualitative coding of tweets according to a pre-existing protocol
  • Case-study analysis of events causing changes in the prevalence of different theories
  • Data analysis and visualization of data produced during the course of this project

Expectations/Commitment

  • Attend either one 2-hour or two 1-hour meetings each week, time TBD upon registrants schedules.
  • Work 6 hours outside of the class meeting.
  • Register for 3 credits of HCDE 496/596.

 

Autumn 2020

Exposing trails of misinformation: How did this content reach me?

We are looking for a few students to join our Fall 2020 quarter Directed Research Group—an extension of the Summer 2020 DRG—to investigate solutions for improving the quality of online engagements. In particular, we will explore whether providing contextual details about the content you see (e.g. information about the other accounts that have shared it) affects your decision to re-post information. Our study looks specifically at information sharing on Twitter.  We will ask the following questions:

  • What informational cues about a piece of online account are important to a user’s decision to share content?
  • How can we design new contextual cues about other accounts to better inform a user about who those accounts are and what they are sharing?
  • How do these new contextual cues impact the perceived trustworthiness of the information and the other accounts that have shared it?
  • Can providing these cues for online accounts that previously shared a piece of information alter users’ decisions about whether and how to reshare it?

As a part of the research process, we will be designing/refining interventions with informational cues that we find are important to information sharing on Twitter (mostly done in first DRG on this topic). We will then study the effectiveness of these interventions — using both qualitative and quantitative methods — and refine the interventions accordingly. The findings will inform the design of social media platforms that make it easy for users to discern which information is worth sharing, and which is not.

What makes you a suitable candidate? 

  • Experience with user research and facilitating user studies
  • Experience with analyzing data
  • Users of Twitter platform (highly preferred!)
  • Looking to register for 2-3 credits (i.e. 4-6 hours of weekly work)

The group will be facilitated by PhD candidate Himanshu Zade with guidance from Prof. Kate Starbird and Prof. Gary Hsieh. Meeting time is TBD.

Autumn 2020

Research through Collaboration: Understanding How Journalists Analyze and Report on Online Misinformation

Do you have strong data analysis and coding skills that you want to use for good? Are you passionate and knowledgeable about combating misinformation and disinformation? Are you a social media maven? Is documentation your strong suit? This DRG might be for you!

Journalists are increasingly investigating and reporting on problematic online content such as misinformation, disinformation and conspiracy theories, leading to the creation of a new misinformation beat. In the first phase of this research, we found that one way journalists overcome some of these challenges is by seeking help from experts. 

By partnering with journalists this fall, we will gather valuable information on how we, as researchers, can support them with tools, collaboration frameworks, and data support. Students in this DRG will become part of a response team that is able to quickly triage and answer data requests from journalists who are investigating misinformation, disinformation, or conspiracy theories online.

This DRG will be a three hour course, suggesting nine hours of commitment per week. This DRG is led by Melinda McClure Haughey with guidance by Professor Kate Starbird.

Winter 2020

Tools for Exploratory Analysis of Social Media Communities

This DRG will develop and apply techniques in the exploratory analysis of online communities found in large social media datasets. Such techniques could include quantitative metrics, data visualizations, or other methods that produce knowledge in pursuit of a research question. We will develop public-facing tools to implement these techniques across multiple contexts.

We will develop our tools in the context of specific research questions, to ground our tools in the practical needs of research. Participants in this DRG should either already be exploring a research question in a specific social media platform(s), or be prepared to adopt a question at the start of the DRG. More junior students can work with DRG organizers to determine a question and/or dataset. This group will be facilitated by HCDE PhD student Andrew Beers.

Goals

This DRG has several goals. We want its members to push and learn from each other in the development of these tools. To that end, we are looking for participants that are comfortable with at least one programming language, or otherwise have prior experience with the analysis of social media datasets.

We want to develop tools that stay relevant to the context of our research questions, while still being portable for use by future researchers. We will aim to make tools that can be used by audiences wider than ourselves, and to conduct our work in a reproducible and accessible fashion. One of the end-products of this DRG will be to share open-source code in formats that can be easily used by our fellow researchers.

We finally want to understand what knowledge our tools can and more specifically cannot make available to us. We will discuss and reflect on the limitations of these techniques throughout the quarter. We will also consider ethical issues in social media analysis, such as privacy and anonymization techniques.

Course Details

This will be a two-credit course, with 2 hours of meeting time per week. Meeting times will be chosen by a survey of DRG participants in the weeks before the quarter starts. The course will progress from a brief literature review to collaborative design and coding sessions.

Autumn 2019

HCI, Journalism & Misinformation: Analyzing Findings and Designing Solutions for the Future

In today’s online environment, journalists deal with an evolving landscape of information available through new and social medias, requiring them to process, analyze and visualize large amounts of data. Journalists also need to deal with the presence of media manipulation, the artificial amplification of certain issues in order to get journalists to cover them. In order to help combat these issues, HCI can help study journalists’ work and provide tools and techniques to better enable them to work in this type of environment.

Building on the progress of the Spring 2019 quarter, we will start by qualitatively analyzing findings from interviews with journalists and ideate on collaborations or tools that could be most impactful for those trying to understand online activity. We will follow a human-centered design process to ensure that our solutions meet the core needs of our users. We will spend the second half of the quarter designing and planning a second phase of research to enhance our findings.

We are looking for motivated students who are interested in journalism, and addressing the problems of misinformation and disinformation online. 

This will be a 2-credit hour course for HCDE 496 / 596. The group will likely meet on Tuesday afternoons. The group will be facilitated by PhD student Melinda McClure Haughey. 

To reply or to quote?: Conversational frames on Twitter

Summer 2019

How we interact with simple design features on online platforms can lead to larger consequences. For example, downstream threads that emerge from replied tweets vs quoted tweets can support different kinds of conversational frames. For investigating this further, we are organizing a Directed Research Group (DRG) 'To reply or to quote?: Conversational frames on Twitter' this summer 2019. There is also a possibility to get involved in a research publication at the end of the DRG.

We are looking for students who:

Are curious about human conversations on social media
Have some experience or want to learn qualitative research methods
Are keen to learn about experimental designs
Want to register for 3-4 summer credits (i.e. 9-12 hours of weekly work)

The group will be facilitated by PhD student Himanshu Zade and advised by Prof. Kate Starbird and Prof. Gary Hsieh. 

HCI, Journalism & Misinformation: Understanding how journalists process, analyze and make sense of data

Spring 2019

In today’s online environment, journalists deal with an evolving landscape of information available through new and social medias, requiring them to process, analyze and visualize large amounts of data. Journalists also need to deal with the presence of media manipulation, the artificial amplification of certain issues in order to get journalists to cover them. In order to help combat these issues, HCI can help study journalists’ work and provide tools and techniques to better enable them to work in this type of environment.

The goals of this DRG are to better understanding what challenges journalists face when using data, what tools have been developed to combat those challenges, and how those tools can be improved upon. We will be examining current practices, conducting interviews, and analyzing the data to formulate the needs of the journalists and designs that could help reach their goals

We are looking for motivated students who are interested in journalism, and addressing the problems of misinformation and disinformation online.  

This will be a 2-credit hour course for HCDE 496 / 596. The group will meet on Tuesday afternoons. The group will be facilitated by PhD students Melinda McClure Haughey and Spencer Williams with guidance from Dr. Kate Starbird. 


Rumor permutations: studying tweets surrounding the Hawaii missile false alarm

Autumn 2018

HCDE Asst. Professor Kate Starbird is seeking a small number of students to join an autumn quarter Directed Research Group (DRG — research for credit) looking at rumor permutations — how rumors permute, branch, and otherwise evolve over the course of their lifetime.

Our case study for this DRG are tweets surrounding the Hawaii missile false alarm: At 8:08 am on January 13, 2018, people in the state of Hawaii received a phone alert warning them of an incoming missile. Although (fortunately) this later proved to be a false alarm, for 37 long minutes residents, visitors to Hawaii, and their families across the country (and worldwide) were trying to make sense of the situation (seeking answers to questions such as: Is there a missile? Several missiles? Is/are it/they nuclear? Where did it come from?), and also sending emotional messages in what it was assumed could be their final ‘words’ on Twitter. Initial data exploration has yielded a rumor pertaining to the relationships between the missile alert and North Korea, and this will be the focus of the research.

We are looking for students with a range of different skill sets and abilities, including data science, and qualitative content analysis. Students with programming experience in Python and MySQL are encouraged to apply. Applicants should note that the research will involve in-depth analysis of tweets that were sent during the 37 minutes when people thought there was an incoming missile, and that as a consequence the content may be emotionally-charged.

This research group will be run by HCDE PhD student Tom Wilson and and supervised by HCDE Professor Kate Starbird. If you are interested, please send an email to tomwi@uw.edu that includes:

  • Current CV;
  • A brief description of why you interested in the group;
  • Any relevant experience;
  • What you feel you can bring to the group;
  • The number of credits you are seeking.

If you have particular research questions you would be interesting in pursuing you can also note them in your email.

Applications will be reviewed on September 19, 2018, and selected students notified by September 21. Meetings will be held weekly on Monday, Tuesday, or Wednesday and are mandatory for all registered students. The meeting time will be agreed with the selected students. Accepted students should register for 2-3 credits of HCDE 496 / 596 and are expected to conduct 3 hours per week of work outside the classroom for each registered credit.


Designing for reflection on social media activity to support resilience to misinformation

Autumn 2018

This directed research group seeks to understand how we can support student reflection on social media trace data related to news and politics. Reflection can be viewed as a kind of thinking that involves stepping outside of a personal situation to acquire deeper understanding and prepare for future action. Our research begins with the premise that supporting reflection on social media data could promote resilience in today’s information environment. In this DRG, we will explore this possibility through designing, participating in, and evaluating activities that allow us to reflect on our own use of social media. 

Our approach will be based on a combination of the principles of “participatory design” and "research through design". Researchers in this DRG will be designing tools (e.g. digital systems, visual provocations, classroom activities) to help students reflect on their social media usage and how that might intersect with phenomena like online misinformation and disinformation. To inform our brainstorming and design efforts, researchers will be reflecting on their own social media data and maintaining a record of their experiences as we ideate and try out new things. Creativity, critical thinking, an interest in cultivating self-awareness and exploring the nature of information we find problematic will be key components of our work in this DRG. 

This will be a 2-credit hour course for HCDE 496 / 596. The group will meet on Wednesdays from 3:30 to 5:30 p.m. The group will be facilitated by PhD candidate Ahmer Arif and Dr. Kate Starbird. If you are interested in registering, please contact Ahmer at ahmer@uw.edu.


Understanding the human impact of hurricanes through social media 

Autumn 2018

HCDE Asst. Professor Kate Starbird is seeking students to join an autumn quarter Directed Research Group (research for credit) looking at the human impact of hurricanes through social media. The goal of this research is to use social media trace data to understand how people adapt to the impacts of hurricanes. This is a mixed methods project — using qualitative and quantitative methods to examine social media posts shared in the aftermath of hurricanes in 2017. We hope to develop new “coding schemes” that identify the different kinds of impacts and adaptations that are visible within social media data as well as new methods for exploring social media data to generate those coding schemes.

We are looking for up to three students to do qualitative analysis and up to three students with some data science experience (Python, R, MySQL or Postgres, etc.) to help with analysis. Students from all departments and all levels are welcome to apply. 

This DRG would be a great fit for students who:

  1. have some interest in analysis of large scale social media data sets and/or want to do research to help people in crisis 
  2. can commit 6-9 hours of work per week (2-3 credit hours) for throughout the quarter

We are particularly interested in (but not limited to) applicants who are fully bilingual and/or culturally fluent in contemporary Puerto Rican culture.

To apply, please send your current resume and a one-page statement about who you are and what you hope to contribute to the project to himanz@uw.edu and ddailey@uw.edu. 

Understanding conversation strategies on social media: How can we redesign discussion forums?

Winter 2018

We are looking for students to join our winter 2017 quarter Directed Research Group (research for credit) investigating day-to-day conversations on social media platforms like Twitter, Facebook etc. We wish to investigate the following in the context of social media conversations about scientific and political topics:

  • How does individual bias impact these conversations? 
  • What strategies do people use to preserve their self-esteem in such a conversation? 
  • When do people learn from these conversations? How can we acknowledge such informal learning?

In particular, we are interested in identifying opportunities for redesigning interfaces to facilitate online conversations such that people can overcome their preconceived beliefs. We will conduct a literature review (read and summarize papers) about conversations on social media, create physical chat rooms for mediated discussions, and create low fidelity (paper) prototypes to gather some preliminary user data.

What makes you a suitable candidate? 

  • Curiosity about human conversations on social media 
  • Some experience with user research and facilitating user studies
  • Prior knowledge (or some understanding) about experimental design
  • Passion for designing new user interfaces 
  • Looking to register for 2-3 credits (i.e. 6-9 hours of weekly work)

The group will be facilitated by Prof. Kate Starbird and PhD student Himanshu Zade. This will be a great opportunity to learn about research through design and user studies. Meeting time is TBD.

Doubt, Disgust & Disinformation: Designing to Support Emotional Reasoning for Networked Societies

Winter 2018

In this directed research group, we will think critically about 'fake news' and the human instinct to focus more on protecting one's world-view and less on discerning truth from falsehood. We will learn about how this instinct is being exploited in increasingly sophisticated ways by propagators of online misinformation to promote their ideas and commitments. 

We will also explore how cultivating self-awareness around our emotions can help us guide our thinking and behavior around some of these challenges. To do this, we will try out different activities to learn about the challenges we become subject to when we read material we don't agree with. And we will try out activities that might help us step outside of those challenges with an eye toward designing future interventions.

This work will help us understand what we can do in educational and online spaces to address the alarming deterioration of trust and emotional manipulation that is occurring in our public spheres. On a personal level, this work might help you learn about things like: 1) Flavors of online misinformation; and 2) What you can do differently when you engage with information to protect yourself as a citizen.

This will be a 2-credit hour course for HCDE 496 / 596. The group will be facilitated by Ph.D. candidate Ahmer Arif and Dr. Kate Starbird. If you are interested in registering, please contact Ahmer at ahmer@uw.edu.

Understanding the Impact of Hurricane Maria on Puerto Ricans through Social Media (En Español)

Winter 2018

HCDE Asst. Professor Kate Starbird is seeking a few students to join a Winter Quarter Directed Research Group (research for credit) looking at the impact of Hurricane Maria on Puerto Ricans. Particularly, we are looking for students who can help with qualitative analysis of Spanish language social media and/or have contextual knowledge of Puerto Rico. Students from all departments and all levels are welcome to apply. 

If you are ...

fully bilingual and/or culturally fluent in contemporary Puerto Rican culture
have some interest in qualitative analysis of large scale social media data sets and/or want to do research to help people in crisis 
can commit 6 hours of work per week (3 credit hours) for Winter Quarter

please drop a line to ddailey@uw.edu expressing your interest and relevant background. 

2017–2018 DRG on Online Rumoring, Misinformation and Disinformation during Crisis Events

Autumn 2017-Spring 2018

We are exploring several research questions about how rumors, misinformation, and disinformation spread online during crisis events (e.g. natural disasters like Hurricane Harvey and terrorist attacks). In related projects, we are looking at 1) how online rumors change over time; 2) the role that journalists and other “professional” media play in spreading and correcting rumors; 3) how intentional disinformation is spread during crisis events.

We need students with various skill sets, including:

motivated and dedicated qualitative researchers with extensive experience studying and/or using social media (for example: Twitter, Reddit, Snapchat)
students with programmatic data science skills (e.g. Python, MySQL, R, Tableau, Gephi)
students with web scraping skills (e.g. using APIs, creating a web scraper for specific URLs)
students with some experience applying machine learning techniques to big data

Meetings times are TBD. This DRG can be taken for 2-3 credits.

If you are interested in applying to participate in this group, please send an email to kstarbi@uw.edu that includes a resume or CV and a written statement or cover letter describing why you would like to be in the group, what you can offer the group, and what you hope to learn by being a part of the group. If you have specific research questions (ours or your own) that you hope to explore within our project, please describe those.

Sketching a Field Guide to Contemporary Crisis Information Work

Winter 2017

Social media and other information and communication technologies are transforming how people communicate in times of crisis. In this directed research group we will be drawing on recent Human-Computer Interaction literature to make brief sketches (in words and otherwise) of contemporary crisis information workers, what they do, and how they do it. In this way, we will make contemporary scholarship more accessible for practitioners. 

Activities will include doing a close reading of selected academic research (about one reading per week). We will then synthesize what we learn in words and images aimed at practitioners.  

We are looking for students with a range of different skill sets and interests, including crisis informatics; information design; visual communication; and/or bridging the research-practitioner divide. 
Students may use this DRG as an opportunity to learn about contemporary communication in crises and/or develop writing or visual portfolio pieces. 

This research group will be led by PhD Candidate Dharma Dailey, with guidance from Assistant Professor Kate Starbird.

Interested students should send Dharma Dailey a resume/CV along with a statement about why you want to be involved and what you can offer to the team. Please mention both relevant past experience and experience you would like to gain. (ddailey@uw.edu) 

Tracking Information Flows across Social Media during Disaster Events

Autumn 2016

This research project examines how social media is used during disaster events such as earthquakes, hurricanes, floods, and acts of terrorism. Expanding upon previous work tracking online rumors, this quarter we will be investigating how information moves through and across social media platforms during disaster events. For example, we may track specific stories or pieces of information as they spread on Twitter, Reddit, Facebook and NextDoor.

Activities this quarter will include: literature review, defining research questions, data collection, and exploratory analysis of social media data. We will collate and summarize existing research in the crisis informatics field—especially studies that look at cross-platform use. From there, we will identify open and important research questions for our group to answer. At the same time, we will work to identify and collect data on emerging events—using existing infrastructure where we have it and building new infrastructure (collection software) where we need it. Once we have data and preliminary research questions, we will conduct exploratory analysis to begin to develop methods of answering our research questions.

We are looking for students with a range of different skill sets and abilities, including software development, data science, and qualitative content analysis. Students who are expert users of Twitter, Reddit and/or other relevant social media sites are encouraged to apply, as are students with programming experience in Python and MySQL.

Interested students should send Professor Kate Starbird a resume/CV along with a statement about why you want to be involved and what you can offer to the team. (kstarbi@uw.edu)

Increasing Motivation and Participation of Digital Volunteers/Crisis Mappers
 
During crisis events, thousands and sometimes millions of social media posts are shared by those in the affected area and observers around the world. These data could be very valuable to affected people and emergency responders, but to make them truly useful, we need better mechanisms of filtering, categorizing, and mapping individuals posts. One promising method for processing these data involves crowdsourcing - using volunteers in the crowd to help classify the data.
 
In this project, we will be experimenting with an existing system for categorizing and mapping social media posts that has been deployed during several crisis events. This project is a collaboration with the creators of that system. Using experimental methods, we will be examining the effects of different kinds of feedback on the motivations of volunteer crowd workers. 
 
We are looking for a small number of students to help with A) system design (for each of several experimental conditions); B) software development; C) data analysis and writing a final research paper. Students can take this DRG for 2-4 credits.
 

We will be building on top of the open source microtasking platform, PyBossa. Students should have web development experience and be comfortable using HTML 5; Javascript; JQuery; and Python or Java for the back-end. 

Tracking the Online Spread of Misinformation after Disaster Events

We are examining social media data to better understand how rumors spread after disaster events. Students can see descriptions of this work in past DRG announcements regarding research on the Boston Marathon Bombings. We have created (and continue to develop) an infrastructure to identify emerging crisis events, collect data in real time on those events, find rumors within that data, then "code" and analyze that data. In this large, collaborative project, we are pursuing a diverse set of research questions—e.g. looking at "correction" behaviors, tracking permutations, understanding "sensemaking" behaviors, exploring better models and metaphors to describe rumoring dynamics, etc. Student researchers can contribute by manually coding Twitter data, analyzing that data (qualitatively, quantitatively, and visually), helping to build automatic (machine learning) classifiers to detect misinformation, and writing up reports and papers from those analysis. 

We are looking for enthusiastic and dedicated student-researchers with a range of different skill sets and abilities who are willing to commit at least 9 hours (3 credits) per week to the project. 

Specific skills for new students in Autumn 2015: we are specifically looking for a small number of students with programming/development experience to help develop our social media data collection framework. For this aspect of the project, ideal students would be interested in user experience, social media, and web applications and would have experience with at least 3 of the following: Django, Django Rest Framework, Angular.js, Node.js, D3.js, PubSub patterns in a web context, and familiarity with Rest APIs.

Social Media Use & the 2014 Oso Landslide

Spring quarter 2015, the emCOMP lab will be analyzing social media concerning the the 2014 Oso Landslide, a recent mass fatality event in our region. Students will have the opportunity to work individually on analytical deliverables that may lead to publications in the context of an active collaborative research project. We are looking for one or two students to do a qualitative analysis of a key local information resource that made ample use of social media. We are also looking for one or two students with a bit of programming experience who would like to do some quantitive analysis of social media data. Ideal students will have an interest in this event, qualitative analysis of social media, and be self-motivated. This research group will be co-facilitated by two of Kate Starbird's PhD students user researcher Dharma Dailey and software designer John Robinson. If you are interested email us (ddailey@uw.edu and soco@uw.edu) with a few lines about why your interested and any related experience you have. Previous experience is not required. One weekly meeting and three hours of work per credit hour is required. Minimum 2 credit hours. 

Tracking the Online Spread of Misinformation after Disaster Events

We will be examining social media data to better understand how rumors spread after disaster events. Students can see descriptions of this work in past DRG announcements regarding research on the Boston Marathon Bombings. This Autumn Quarter, we will begin to focus on more recent events, including the bombing of MH17, hurricanes in Hawaii, and violence in Iraq.

We will be doing manual coding of Twitter data, big data analysis (qualitative, quantitative, and visual), and building automatic (machine learning) classifiers to detect misinformation. We're looking for students with a range of different skill sets and abilities, including qualitative coders, data scientists, and programmers with some experience using machine learning algorithms. Students who sign up as qualitative coders will have the opportunity to learn data science techniques (e.g. MongoDB, R, Tableau).

Interested students should send Professor Kate Starbird a resume/CV along with a statement about why you want to be involved and what you can offer to the team. (kstarbi@uw.edu)

Understanding Public Information Needs in Crisis Contexts

When a disaster strikes how will you get the information that you need?
How do emergency professionals reach the public?  
What are the prevailing information needs in a crisis? Are they all met the same way?
There are many challenges to getting information out to the public during a crisis. Crises are inherently unpredictable, often interrupting predetermined strategies for getting information to the public. People are also ever more diverse in terms of the communication tools and platforms that they turn to for information. Given these challenges, the Public Information Needs in Crisis Directed Research Group examines the current strategies employed to get information out during crises from two perspectives: official response operations and observed recent information sharing behaviors among the public.  
In Spring 2014, we'll use document analysis and informal interviews with domain experts to create a basic model of how the work of informing the public during crises is taking place for one locale in the United States. With guidance and support students will be responsible for conducting and documenting a portion of the analysis.
This research group will be run by HCDE PhD student Dharma Dailey and iSchool PhD candidate Beth Patin, and supervised by HCDE Professor Kate Starbird. If you are interested, please send an email to both Dharma (ddailey@uw.edu) and Beth (bethp@uw.edu) with a brief description of why you interested in the group, any relevant experience, and the number of credits you are seeking. Meetings are mandatory for all registered students. Students should register for 2-3 credits and are expected to conduct 3 hours per week of work outside the classroom for each registered credit.
Designing/Developing a Content Curation Web App for Disaster Volunteers (2014)
 
Disaster events have long been catalysts for pro-social behavior, and spontaneous volunteers are known to “converge” onto the scene to help, often by improvising solutions to unexpected conditions and gaps in response efforts. In recent years, people have begun to turn to social media after disaster events for a variety of reasons: locals use social media to seek and share information, emergency responders use these tools to communicate with their constituents, and remote individuals come together on these platforms to offer help of various kinds. Members of this latter group are sometimes referred to as digital volunteers. One of their primary activities involves curating—i.e. filtering, classifying, and synthesizing—the massive amounts of information available on social media during disasters, helping to make this information usable to emergency responders and affected people.
 
The primary focus of this directed research group is to understand the behavior of these digital volunteers and to develop tools that support their activities and goals. In preliminary design work, we have generated design ideas focused around a collaboration curation web application where multiple volunteers can filter and categorize social media posts (tweets, photos).
This is an ongoing project with a small, dedicated group of student-researchers. We have developed preliminary designs based on contextual interviews with several digital volunteers. We are now moving towards developing a web application based on our initial designs. We plan to continue with an iterative design-develop-test cycle where we complete user testing on the application at various stages.
The group has multiple, related goals:
  • To develop and deploy a web application to support the collaborative, connected work of digital volunteers
  • To better understand the phenomenon of digital volunteerism
  • To better understand how groups collaborate and coordinate online 
  • To publish research on digital volunteerism and online collaboration 
We are currently looking for a small number (2-3) of additional team members, primarily to help with the development of functional prototypes from our preliminary designs and to participate in user testing of these prototypes. Students will be able to sign up for 2-3 credits each, depending upon the amount of time they will be able to devote to the project.
Desired Skills:
  • web programming experience (e.g. Javascript, PHP, Ruby on Rails, other)
  • database experience (e.g. MySQL, MongoDB, other NoSQL database)
  • comfortable working as part of a team
If you are interested in applying to join our team, please send an email to Kate Starbird (kstarbi@uw.edu) that includes a statement for why you are interested in working on this project as well as a current CV/resumé. Please highlight your relevant skill sets.
Researching at the Intersection of Social Media and Disasters: Tracing the Spread of Misinformation after the 2013 Boston Marathon Bombing (2014)
 
In recent years, disaster events have become catalysts for massive convergence online. After major events like the 2011 Japan Earthquake/Tsunami and Hurricane Sandy, hundreds of thousands—and in some cases millions—of users turn to social media platforms to share information and to collectively make sense of the event. This activity leaves a significant digital record that can be used to study and better understand mass convergence behavior during disasters as well as the role of these new technologies in our lives.
 
This research group will examine Twitter data collected after the 2013 Boston Marathon Bombing. We collected over 10 million tweets during a seven day window beginning a few hours after the explosions, and we have gone back and captured all of the links in these tweets as well as the content of the linked-to webpages. As part of an ongoing research effort, we plan to examine this “big data” with a goal of better understanding the spread of misinformation during disasters. Using a combination of qualitative and quantitative methods, we will investigate the spread of rumors, the crowd work to identify suspects (that went awry), and several conspiracy theories that propagated through Twitter as well as the surrounding information space of the wider Internet.
 
We are looking for a small group of students with complementary skill sets to continue this research. We are looking for researchers with qualitative skills who will:
  • help develop coding schemes for our analysis
  • read and classify 1000s of tweets and web pages
  • write papers describing our methods, analysis, and findings
We are looking for researchers with quantitative skills who will:
  • make network graphs and other visualizations
  • implement machine learning algorithms to automatically classify tweets and links
  • complete statistical analyses on these data
For this latter group, the following programming skills are recommended:
  • Programming in a scripting language (Python, Ruby or another) 
  • Database programming (MySQL or MongoDB) 
Team members may contribute to either the qualitative or the quantitative/computational side of the project (or both!). Other important skills include: comfort working with a team and great communication skills (writing/presenting).
 
If you are interested in applying to join our team, please send an email to Kate Starbird (kstarbi@uw.edu) that includes a statement for why you are interested in working on this project as well as a current CV/resumé. Please highlight your relevant skill sets.
 
 
Collaborative Crisis Curation: Designing a Social Media Processing Tool for Emergency Managers
 
In this second quarter of a multi-quarter project, we will be continuing our work to design a collaborative platform for emergency managers to help them filter and organize the information streaming in from social media during crisis events. Emergency managers and disaster responders are increasingly turning to social media as a potential information source during disaster events. As they do this, they are faced with new challenges related to processing this flood of information, including finding strategies and tools to deal with the huge volume, noise, lost context, misinformation, etc.
In the Fall Quarter, we identified a research opportunity within this space and developed a research plan for designing a platform to help a virtual group of emergency managers track global events using social media. This quarter, we will continue the human-centered design process, moving from user studies to prototyping of our new tool. Our end-goals for the quarter are to produce a high-fidelity prototype of this tool and to write a paper describing our research and design.
 
Though this research is ongoing, there is some opportunity for a small number of students to begin in the Winter Quarter. We are particularly interesting in students with experience in the human-centered design process and with advanced prototyping skills (web and mobile), and/or students with development skills that can help with the software design and implementation of our tool (web scraping, web development, mobile development, databases, machine learning, etc).
 
If you would like to apply to join the existing group (for 2–3 credit hours of credit/no credit grade of HCDE 496/596), please send an email to Kate Starbird (kstarbi@uw.edu) with a few paragraphs describing why you are interested in the project, your relevant skills, and the number of credits you are seeking. Meetings are mandatory for all registered students.
 
Investigating ICT use during Mass Convergence Events
 
This quarter, my directed research group will focus broadly on the use of ICT, including social media and mobile technology, during mass convergence events. Impactful events in the physical world are now triggering digital convergence in the online sphere. We will look at large-scale interaction and collaboration—including behavior that takes place completely online as well as online-offline coordination activity—during events such as natural and man-made disasters, entertainment and sports events, and political protests.
 
Possible events for analysis include the 2010 Deepwater Horizon Oil Spill, the 2011 Joplin Tornado, the first week of the Occupy Wall Street protest, the 2012 Olympics, or future events.
The goal of this group is to bring together students who can approach this domain from different perspectives, pooling a variety of skills (including qualitative, quantitative and computational). You will be encouraged to bring or find your own research questions within this space.
 
Three (broad) areas for research are:
  • Address specific questions about ICT-enabled human behavior during mass participation events, e.g.
  • Design tools for research and/or real-time analysis of social media interaction during these events, e.g.
  • Design tools to support the activities of those participating in these events (on-the-ground participants/fans/etc., emergency responders, organizers, digital volunteers, etc.)
Initially, the group will seek to understand the domain, sharing and discussing readings from research in the areas of crisis informatics, social media dynamics, crowdsourcing, etc. As the quarter progresses, we will work to identify and coalesce around specific research questions. End products may vary, depending upon the trajectory we take—publishing empirical research, designing and developing tools, etc. Long-term goals are to publish resulting work at appropriate conferences.