I think any study or analysis of the way feeds are refreshed on websites and applications would need be to done in the context of the type of user accessing the information and also the type of content being published.
The reason why it wouldn't make sense to look at only the type of user accessing the information or just the type of content being published is because these two combined provide the best indication of the context of usage:
Type of user accessing the information
- Emphasis on the 'freshness' of the content
- Emphasis on the 'completeness' of the content
Type of content being published
I'll provide and example of how different design strategies are applied in the case of news broadcasts. In a typical news broadcast, the presenter and the viewer is focused on recent information. Generally when the news presenter is speaking, we are provided with what would be 'relatively' new information. In addition, there are often news flashes being displayed across the bottom of the screen for news information that are probably 'fresher' than what the presenter is providing. On top of this, the presenter will interrupt what they are presenting so they can cross over to 'breaking news'.
So here we can see that the news bulletin is a 'complete' content that has a stable shelf life, compared to the 'fresh' content that also has a stable shelf life. The breaking news is 'fresh' content that has a relatively short shelf life.