How, really, to determine if a trope or joke is still current? Can it be done?

Edit: There are some issues with Google's Ngram which anyone using it should be aware of. I think it can be very helpful, but others may not. I was going to add in some more details on the shortcomings of it, but I found some meta posts that I'll link to instead:

• Should we allow Google NGrams to be presented as statistical evidence without qualification? Should we define a set of standards for their usage?
• How accurate is Google Ngram as a language reference source?

Mostly piggy-backing off Robin's answer, using Google's Ngram database would be a great way to check the popularity of a word or phrase in print, with the exception that it currently looks like you can't search past 2008. Edit: Actually, it looks like there is a significant change in the data across the board for dates after 2000, especially around 2004, which is probably responsible for the dip at the end of the graph shown below.

To get accurate results, it is important to read the about page. Some helpful tools to use include:

• Wildcard searching
• Case-insensitive searching - useful in this case as "Freudian" is often capitalized, but "freudian slip" wouldn't show that without selecting "case-insensitive"
• Searching phrase variation, such as boughten (old usage - think Laura Ingals-Wilder books) vs store bought vs store-bought.

Notice the difference between Robin's screenshot and this one:

Regarding your example of "Freudian slip", according to Ngram, it looks like the phrase originated around 1925, and its usage peaked around 2002. It looks like it's moving out of usage, but may still be relatively popular, assuming its usage hasn't decreased more rapidly since 2008. Maybe not, see this link.

An Ngram:

This is what Ngram says

So, is there a way to determine the popularity of usage of [some idiom], from year A through to the present?

Nope, it can't be done.

There are a number of different ways to scan published materials, and I wouldn't be surprised if, 5-10 years from now, there comes some scheme for doing speech recognition on TV shows and the like and scanning that data base. But these schemes have two basic deficits:

1. They are always looking to the past. In most cases, to have a sufficiently large and diverse data base to scan, data must be effectively aggregated into periods of time of at least several months, and often several years. So asking if something is "current" is typically asking if it's been used in the past few years.
2. The data that is collected is always narrow in scope. Eg, Ngram used data mostly from "hard-cover" books, and a few magazines, and a lot of popular idioms will rarely if ever appear in such venues. Even if the idioms being researched manage to make it into print (a fairly high hurdle), they're generally going to be seen in your less erudite publications first, and only make it into Newsweek when the idiom is beginning to become stale.

The best you can do is use Ngram or one of the other tools for scanning published literature and live with the limitations -- the information will not be current (rarely "fresher" than about 2 years), it will be aggregated in periods of months or years, and the available data will be heavily tilted toward "establishment" cultural standards.