With any story driven by a large data set, we know there will be limitations and concerns. So, we wanted to use this space to be as transparent as possible. 

The article entitled ‘Is Greek life at Duke as homogenous as you think?’ focuses just on organizations in the Interfraternity Council and Panhellenic Association. However, upcoming articles will look at the topic from the perspective of Multicultural Greek Council organizations, National Pan-Hellenic Council organizations and selective living groups. When we mention IFC, we also included Kappa Sigma. Even though it is not officially recognized by Duke, it still recruits new members, as evident by public social media posts and LinkedIn. The charter was in IFC five years ago and is recognized by the national organization currently.

In all, The Chronicle collected data on every member of the Class of 2018 listed in the “Freshman Picture Book” the University publishes every fall. That picture book contains every member of the senior class except for a handful of students who for unknown reasons were left out of it. A digitized version of the book is also posted on the Duke Library website for anyone to view, which is the case for every class dating back to 1982.

Some members of the Class of 2018 might have transferred from Duke since being in the Freshman Picture Book, but because that data is difficult to collect and for consistency, we included only those listed in the book. People might have also left Greek life since joining, which the data does not reflect. But, we chose to focus on those that were accepted through the rush process.

To find the tuitions and other information about private schools, we used education-review website Niche.com. If the school was not on Niche.com, we used information from the school’s website, news articles about the school and a number of other public sources. Still, there were some schools for which we could not find data for certain variables—this was true mostly for a few international schools, especially those in countries where English is not the primary language.

International schools also refer to themselves as public and private based on different criteria than schools in the United States. We identified schools the same way they are identified by their websites. Therefore, some public schools in our data set also have attached tuitions, and some private schools have no attached tuitions.

For schools that had at least a majority of boarding students, we included boarding costs along with tuitions. For schools in other countries, which sometimes post tuition just for one semester or term, we multiplied the cost of a term’s tuition by the number of terms in a year. We also converted all currencies to American dollars. 

To find the percentage of students eligible for free and reduced at U.S. public schools, we used the National Center for Education Statistics’ school look up tool. If a school wasn't in that database, we found the percentage on Niche.com, ProPublica's The Opportunity Gap pages, the school website or other public sources. There were a handful of public schools where this information could not be found.

All these figures are from the most recent years available.

To find out which students were in which IFC or Panhellenic organization, we used public social media posts, national and chapter fraternity and sorority websites, LinkedIn and a picture of all members in Delta Sigma Phi on the wall of the restaurant Devil’s Krafthouse. For a minority of students, neither they nor their Greek organization posted their Greek affiliations publicly. However, any of them whom we affiliated with a Greek organization were either tagged by name in a photo by another member of their Greek organization, or they were identified in a group photo of members of the Greek organization through a matching public image of them.

We also collected data on career and academic interests of each IFC and Panhellenic member. From publicly available data on LinkedIn and Duke-affiliated sites, each internship experience of an IFC or Panhellenic member was categorized into the subdivisions used by the Duke Career Center: engineering and technology (e.g. Microsoft, Google, etc.), government and politics (e.g. U.S. Congress, lobbying firms, etc.), health and life sciences (e.g. research, hospital work, etc.), media and sports (e.g. journalism, ESPN, etc.), business (e.g. finance, consulting, etc.) and nonprofit (e.g. charities). 

With this methodology, we actually found very little overlap. For the very few cases where there was overlap, we decided on a classification and we were consistent with those classifications throughout. When no information was available, we classified that situation as unknown for both work experiences and major. 

If a student worked in one field one summer (e.g. government and politics because they interned for a senator) and then another field the next summer (e.g. business because they interned for a consulting firm), both experiences were noted.

Lastly, we looked at the senior members of a number of clubs to test for an overrepresented Greek presence. But, we only included clubs for which we found that there was actually an overrepresented Greek presence in the article. To find the senior members of Dukes and Duchesses and Campus Enterprises, we just used their respective websites as they accept members by each class (i.e. Class of 2018, Class of 2019, etc.), which is clearly denoted on their websites. 

To find senior members of Investment Club and The Standard, we used web.archive.org to find Class of 2018 members on the clubs’ respective webpages since Oct. 2014. For Investment Club, we only looked at analysts and executive board members.

For FORM magazine, we looked at each issue published on issuu.com and a physical copy of the most recent issue for Class of 2018 members. For PWaves, a Chronicle member who is senior and attended the pre-orientation program identified the other participants.

Our data set can be viewed on GitHub, though the names of high schools and of students has been removed to protect their identity even though, again, the data is based on public information that any person could access.

Some other notes:

We could not collect data on race because it is self-identified. However, we recognize that race is a major factor in a discussion about Greek homogeneity, so we were able to get some sources to address that topic.

When looking at individual fraternities or sororities, one risk of our analysis is small sample sizes. This can be said about Delta Kappa Epsilon, Chi Psi and Alpha Delta Phi, which each had 10 members or less.

If you find any inaccuracies with our data, please contact Likhitha Butchireddygari at lb255@duke.edu or Jack Dolgin at jbd33@duke.edu.