Yes, this is how the Windows version works as well. The setting that is being used here is the “Use Fixed Row Height” toggle, at the very bottom of the menu. In that mode, the outliner acts a bit more like the corkboard in how it populates the cell with an excerpt of the synopsis (which can be of course adapted from the main text if it is left empty, which is kind of what we are seeing here). If the item has no title, then the synopsis/text occupies the entire cell to a maximum of four lines (three if you have a title, to remain fixed).
Which leads to the question, why do you have a formal title at all on this item, if the title is in the text as a styled heading? Wouldn’t your compile settings be leaving the Title off anyway, to allow the text to handle the heading?
To step back and clarify the design a bit here, while Scrivener’s core design is very much tilted toward item titles having the power to become headings in the output document, the purpose of this “empty=content one layer down” approach is so that you can, if you want, have the first line of your text be the heading, and then by leaving the synopsis and title for the item empty, let that text in effect become its title in searches, binder and bookmark listings, etc.
Having both is giving the software two competing working habits as data, so it isn’t odd I would say to see double the data. For that reason I don’t see this feature request itself having legs, as it would only serve to make less visually awkward an awkward data request, if that makes sense.
I use this capability myself now and then, but typically more at deeper sublevels. For example a tip box often starts with a box header, a table can start with a caption, and so it often makes sense to leave the title empty on such things, so that its representation in the outline is driven by the content. I wouldn’t want a heading on top of a tip box, figure or table anyway.
As far as I can see, this has no effect on the Outliner. Just on the output.
Right, otherwise changing compile Formats could radically alter the data you work with in the views, which would be kind of weird and force you into having to swap settings around between what you want to export and what you need to work comfortably—the opposite of Scrivener’s ethos.