A few years ago, OrgTheory had a brilliant post tracking the trend of increasingly long article titles, and the growth in the use of colons and subtitles, in the journal Administrative Science Quarterly. The main finding was that average article title length had grown from 9 to 15 words from 1958 to 2008, and colon usage had gone from 20% to something like 90%.
A couple years before that, James Moody published a short piece in the American Sociologist, Trends in Sociology Titles. Moody’s findings were similar: mean words grew from 7 in 1935 to 12 in the early 2000s. Colon usage went from near 0 to 75% in the late 90s. Moody also rejects a suggestion proposed by Howard Becker a few years before that in the hilarious titled “Long-Term Changes in the Character of the Sociological Discipline: A Short Note on the Length of Titles of Articles Submitted to the “American Sociological Review” during the Year 2002.” Becker argued that the trend of longer titles might be an attempt to garner more citations, with the idea being that longer titles attracted more reader attention. Moody finds the relationship between article title length and citations to be negative (but not significant). Of course, Moody’s finding doesn’t mean that authors aren’t trying to garner attention with more specific wording, it just suggests that these attempts aren’t working. It may also be endogenous – more prestigious authors can afford to be vaguer and will still get cited more often. And so on.
Edit: A quick analysis of the last four issues of ASR shows the colon trend has not stopped. Of the 26 articles in the four issues, 17 have a colon, and another 4 have a question mark or exclamation point. So, a total of 21/26 have a subtitle of some sort, about 81%.