Justin Ebert Dot Com
The official website of the illustrious Justin Ebert

Headlines at Duke February 24, 2018

This weekend I’m visiting friends at Duke University in Durham, North Carolina. Duke is home to an independent student newspaper, The Chronicle. As practice for another piece I’m working on I’ve been running an analysis of their headlines.

To collect the 918 headlines I analyzed I used scrapy, a web crawling framework written in python. In oversimplified terms, the program works by finding links on a starting webpage (in this case dukechronicle.com) and following them. Then the program downloads the headline, date, and author of the post it finds.

Once I got the data, I ran some quick formulas in excel to get to know the headlines. The Chronicle’s favorite words (besides common ones like “to,” “and,” etc.) in headlines are:

1 DUKE 226 1.41% 24.07%
2 STUDENT + STUDENTS 103 0.64% 10.97%
3 NEW 85 0.53% 9.05%
4 CAMPUS 57 0.35% 6.07%
5 DSG 44 0.27% 4.69%
6 CENTER 37 0.23% 3.94%
7 SEXUAL 32 0.20% 3.41%
8 COUNCIL 30 0.19% 3.19%
9 FACULTY 27 0.17% 2.88%
10 DURHAM 28 0.17% 2.98%

Based on these, the perfect Chronicle headline would be something along the lines of “DSG to Duke Students: Faculty, Student Council Unveil New Durham Campus Sexual Center,” which would also be within standard deviation of the average length of headlines in the past year, 10.3 words.

In fact the average word count for Chronicle headlines has been rising since they began focusing on their digital platform. Of the headlines in my data set about 70 (I know, not a great sample considering the thousands of headlines they published) were from before 2010, with an average word count of 5.7 per headline. Fast forward to 2017 and the average Chronicle headline has nearly doubled to 10.5 words.

Character count rose at a similar rate, and the average length of words remained constant. One possible that because The Chronicle has more space for headlines online, they don’t feel the need to limit themselves the way they did in print, and have been writing more descriptive headlines. The longest headline in my dataset by characters: “City Councilwoman Jillian Johnson, Board of Education member Sendolo Diaminah speak to Allen Building sit-in students Sunday morning,” published in 2016. This headline is so long, if you wanted to tweet it you’d only have 8 characters leftover for the link.

No Comments on Headlines at Duke