I would suggest that it depends on the case study and your audience. That said - this is a perfect scenario for an A/B test - you can test a couple of hundred random contacts - see which does better - then go with that for the rest of the blast.
I also check the email subject line and content with Litmus. It will tell you what will work in different browsers, browser versions, and mobile devices. We do a Litmus test before every blast. Learn more about Litmus at https://litmus.com/