Undergraduate Research and Creative Collaborations

Faculty Name: Chandler Rosenberger 

Faculty Email:  crosen@brandeis.edu 

Faculty Position Title: Associate Professor 

Faculty Department or Program: International and Global Studies/Sociology 

Hourly commitment (range in hours/week): 3-5; 4-6; 5-7 

Project Title: “Reading the Air”: How Machine Learning to Can Help Us Understand the Emotions Behind Public Opinion in China 

This research assistant position will support Professor Rosenberger. 

Position description 
I am looking for an undergraduate research assistant with strong computer programming skills; the ideal candidate would also have an interest in using those skills to study public opinion as reflected in social media.   

My current project is an extension of work I began last year with the help of a 2022-23 Hearst Fellow, with whom I studied online movie reviews.  Commenting on controversial films from the past, we discovered, has often given ordinary Chinese the chance to express opinions indirectly and thus without censorship.  We reviewed hundreds of these film reviews and quickly identified patterns in the language contributors were using. These patterns revealed deeper patterns in the thoughts of contemporary Chinese about the future of their country. 

The next step for such research is to find new, unanticipated patterns: not just the things we can imagine but do not know, but instead the things we don’t know and that we don’t even know we don’t know.  The goal is to train a computer to find patterns in the use of terms that researchers have identified, but also to search for and find other patterns that we have not seen.  We could look, for example, at terms that show up consistently near the terms we have already chosen to study.  This could show us patterns we did not even know to look for. 

Skills the student will learn 

A computer programmer on a Hearst Fellowship could help this project by: 

1. Write programs to “scrape” web sites and organize data in searchable databases
2. Write machine learning algorithms that can be trained to search for both known and unknown patterns in these databases.  We will be especially interested in terms that appear frequently in the collected posts but not in other parts of the internet: these terms may be coded language referring to something else
3. Run analysis on the patterns to identify which are statistically significant.

The project’s focus will be social media sites with posts about the writer Lu Xun (see the description of the research project). A Hearst Fellow learning to apply computing skills to the study of social media would quickly become one of the most sought-after collaborators in arguably the hottest field of social science research.  Such a Fellow would already have highly-valued programming skills. Under the tutelage of a sociologist, however, the fellow would learn social science skills as well, such as:  

1. how to frame social science questions properly; 
2. how to identify socially significant language, such as metaphors with social import (e.g. “the tide of history”);
3. how to identify plausible patterns of meaning behind terms that appear frequently together.  We have noticed, for example, that language comparing human events to natural events (floods, earthquakes, etc) is strongly tied to a fatalistic mindset;
4. in combination with the computer skills, how find these patterns, and unanticipated ones, in large databases of social media posts.

An undergraduate with a combination of these computing and social science skills would be reviewed as a top candidate at the best social science graduate programs worldwide.  They would also be top candidates for jobs in politics, marketing, and consulting.   

Faculty Name: Jytte Klausen 

Faculty Email:  klausen@brandeis.edu 

Faculty Position Title: Professor 

Faculty Department or Program: Politics 

Hourly commitment (range in hours/week): 5-7 

Project Title: Why Do They Do It? A Study of Domestic Violent Extremists and Their Radicalization Trajectories 

This research assistant position will support Professor Klausen. 

Position Description 

The student will be trained in the use of a detailed codebook to extract relevant information from public sources--news media, documents from trials, and statements by convicted extremists--used in the research to track how people who commit criminal acts on behalf of political figures and ideologies radicalize.  

Skills the student will learn 

The student will learn how to use an online portal and spreadsheets to archive and analyze the data. We are at this point still coding data but as I progress on the manuscript, the student will be taught how to make manuscript-ready tables and graphs.  

The student will join a small team of students who have worked with me for the past two years. Teamwork is essential to the type of research I do. The successful student will progress from data collection to analysis, and is expected to learn how to use an EXCEL spreadsheet. The students may be asked to take a short course provided by the library to improve their data analysis skills. Group meetings are an important part of the learning process 

Faculty Name: Jingyi Huang 

Faculty Email:  jyhuang@brandeis.edu 

Faculty Position Title: Assistant Professor 

Faculty Department or Program: Economics 

Hourly commitment (range in hours/week): 5-7 

Project Title: Text as Data: Use Historical Records to Analyze Cartel Collusion 

This research assistant position will support Professor Huang. 

Position Description (2 available)

The research assistant will play a central role in constructing the dataset. Tasks include: 

1. Write Python scripts to extract information from unstructured texts 
2. Clean and standardize text and convert it to computable data
3. Create descriptive statistics and network figures to present the results

Skills the student will learn 

1. Coding: Developing Python code to extract board member names from raw texts with varying layout structures
2. Data management: Cleaning and harmonizing the output to construct a dataset across multiple years
3. Data visualization: Creating figures to present the the findings on inter-firm network
4. Communication and project management: Coordinating with others on the team to optimize the code and build a consistent data structure

Faculty Name: Jill Greenlee 

Faculty Email:  greenlee@brandeis.edu 

Faculty Position Title: Associate Professor 

Faculty Department or Program: Politics (and WGS) 

Hourly commitment (range in hours/week): 3-5 

Project Title: Gender, Family, and Running for the White House 

This research assistant position will support Professor Greenlee in a research project and participate in her Gender Lab, which is dedicated to enriching the research experience of undergraduate research assistants. 

Position description (2 available)

The undergraduate RA will continue building out a dataset of articles from The New York Times to identify how and when candidates' families are mentioned in connection to their presidential campaign. The student will also help think through how to organize these data, help to identify patterns in data collected by another research assistant, and aid in identifying if any archives have materials that are of interest to me for this project. 

Skills the student will learn 

The student working on this project will gain skills in systematic data collection, critical thinking to identify patterns and themes qualitatively, and writing skills (when writing up summaries). The student will also identify and explore electronic archives and the sites of presidential archives. Learning about how the written record can be used as data is another skill the student will gain. 

Faculty Name: Lucy Goodhart 

Faculty Email:  lgoodhar@brandeis.edu 

Faculty Position Title: Lecturer 

Faculty Department or Program: IGS and POL 

Hourly commitment (range in hours/week): 3-5;4-6;5-7 

Project Title: How do Moderate Candidates Message Voters in Primary Elections for the US House? 

This research assistant position will support Professor Goodhart. 

Position Description 

Revised Project Description Lucy Goodhart 

The undergraduate RA would contribute to the analysis  of television advertisements in primary elections for the US House, comparing the messages of more centrist and more ideological candidates.  The undergraduate RA would be responsible for the following: 

• Maintaining and updating the data set of video files by US House primary candidate with original data from the Wesleyan Media Project (WMP). 
• Running ads through Zoom in order to obtain a preliminary transcript of each ad and revising those transcripts for greater accuracy. 
• Using simple, descriptive statistics to compare ad features across candidate categories using existing coding categories from the WMP. 
• Watching a sub-sample of ads, as will the PI, Prof Goodhart, to jointly generate a coding scheme of ad types based on features the candidate wishes to message.   
• If interested, the RA would also research the best, AI-based, contemporary software for textual analysis to be applied using the coding framework we develop. 

Given the nature of the project, applicants will ideally have the following skills: some background in American politics; descriptive statistics at the level of POL 52A, including hypothesis tests and their interpretation, and the ability to log your work and write do files, plus facility in either Stata or R. 

Skills the student will learn 

The research will reinforce proficiency in simple, descriptive statistics (using Stata or R), will develop skills in hand coding of qualitative, textual data and will increase knowledge of machine-based coding tools. 

Faculty Name: Charles  Golden 

Faculty Email:  cgolden@brandeis.edu 

Faculty Position Title: Professor 

Faculty Department or Program: Anthropology 

Hourly commitment (range in hours/week): 4-6; 

Project Title: GIS Analysis of Ancient Maya Landscapes in Guatemala and Mexico 

This research assistant position will support Professor Golden.

 Position description 

Student researchers will work with faculty at Brandeis, Brown, and the University of Florida on this project to identify ancient settlement, with the possibility for participating in professional publication and presentation opportunities. Research assistants should have a foundational knowledge of GIS and particularly ArcGIS Pro or QGIS software.  

Skills the student will learn 

Research assistants will need a foundational knowledge of GIS software - using ArcGIS Pro or QGIS software - to make maps. Building on this foundation, the project will engage researchers in higher level analysis and data manipulation, particularly the study and transformation of raster data and studies of spatial clustering that provide advanced GIS skills applicable in many professional trajectories. 

Faculty Name: Zachary Albert 

Faculty Email:  zalbert@brandeis.edu 

Faculty Position Title: Assistant Professor 

Faculty Department or Program: Politics 

Hourly commitment (range in hours/week): 3-5;4-6 

Project Title: Sore Losers and Negative Rhetoric in Political Campaigns 

This research assistant position will support Professor Albert. 

Position description 

Democracy requires peaceful transitions of power from election losers to winners, but increasing partisan polarization and technological changes raise the possibility that “sore losers” have incentives to go negative, target their political opponents, and even challenge the results of the election itself. This project will address the question: Are losing candidates more likely to engage in negative rhetoric in post-election communications with their supporters and, if so, what are the effects? The project will analyze candidate email rhetoric and candidate fundraising hauls before and after the 2020 election, exploring how an election loss affects rhetoric and how this rhetoric affects popular support for candidates.  

The undergraduate research assistant will aid in data collection and basic data analysis. First, the RA will collect candidate-level data  –  such as election results and campaign finance records – from public databases, enabling the analysis of candidate rhetoric in emails based on their post-election status and the effect of sore losers' rhetoric on future fundraising. This process will involve the RA navigating online data sources and entering information into spreadsheets. Second, the RA will participate in textual analysis of candidate emails, including training machine learning models to assess negativity and identify instances where candidates mention their political opponents. All models will be trained and applied in the R programming language; part of the RA's responsibilities will include learning and practicing intermediate R skills independently and with the PI's guidance. Lastly, the RA will use these R skills to produce basic data visualizations to convey descriptive findings. 

Skills the student will learn 

1) Ability to navigate and collect data from online databases (Ballotpedia, OpenSecrets, Federal Election Commission) 

2) R programming language (intermediate level), including the ability to read and combine datasets, manipulate variables, perform basic descriptive analyses, and produce data visualizations 

3) Methods of textual analysis (including sentiment analysis and keyword extraction) and machine learning (including automated content analysis) in R 

4) General research design and research practice skills (e.g. structuring data collection processes; operationalizing variables; pairing methods, data, and questions)