It depends on which perspective you want to focus on (marketing, psychological, design, behavior, information management, usability, mobile news trends, etc.). You could also focus your research on "clickbait", "misleading", "online content", "false news", "fake news", etc.;  and find some leads as to what the current approach is when researching online news. I'd suggest you first decide on an approach (or find a new one) and take it from there.

I'll attach some references, hope some of them might be useful for your research.

Cheers!

Article Mixed-methods approach to measuring user experience in onlin...

Article Choosing and Reading Online News: How Available Choice Affec...

Article What Prompts Users to Click and Comment: A Longitudinal Stud...

Conference Paper Misleading Online Content: Recognizing Clickbait as “False News”

Article Mobile Online Journalism for smartphone. Analysis of multime...

Article SIX THINGS YOU DIDN’T KNOW ABOUT HEADLINE WRITING: SENSATION...

Article Stop Clickbait: Detecting and Preventing Clickbaits in Onlin...

Conference Paper Automatic Extraction of News Values from Headline Text

Dear Joseph,

Thanks for your prompt response, this will certainly help. I do have a research focus already (even though this project is certainly in the experimental phase), but in my field not a lot has been done on this topic. I also tried to extract relevant terms based on some machine learning tactics, but these results were not very promissing in predicting the popularity of the news item. But, as I'm not a a researcher within the communication sciences, I was hoping on a more in depth approach on this matter from this angle.

I'm sure these papers will already help a lot. Will look into these.

Thanks!

Kind regards,

Iris

I think that you could also look for guidance given to headline writers, both in actual journalistic media, and headlines for the web.  

One term I see often online today is "just."  Someone JUST did something.  This gives the story more urgency.

Michael's idea is a good starting point. If you're interested in working within the communication sciences, there are some theories you could work with, but since your research is an experimental one, I'd say it'd be interesting to combine more than one theory for different parts of the process, for example:

Use one theory (maybe some form of structuralism/semiotics) to identify which are the common elements or phrases used in headlines -though you'd have to limit your sample size to specific news sites. Then use another theory (could be functionalism/mass communication) to explore the effects these "triggers" have on citizens (though you could do this from conductivism or another psychology theory).

That was just an example (and those are pretty vanilla/classic theories, you might as well use more modern theories like the political economy of communication), but it has the potential to easily become a fascinating research.

It is fascinating indeed. I've got results as a side effect from another project, but these already show the importance of (non-)specific terms, based on an analysis of one newspaper (and around 45000 newsitems). These are not results I would have expected though. But my expectations relate to pure gut feeling here, hence the initial question.

I already wrote the algorithm for one source and will just run it over another one to verify if this results in a similar set. The initial one was rather related to a specific, though broad, topic, which can explain the strange effects. But even the difference between these sources might be interesting.

But as said, first need to catch up with some literature. As fascinating as it is, it is out of my field of expertise. So your help is much appreciated (even if you coin it as vanilla theories ;), it's a starting point).