I'm not going to present data - just a personal anecdote. If someone else has done a systematic analysis, it would be a much better answer.
Are standards rising?
Here is a brief history of Skeptics.SE.
In the early days of the site, we were still exploring what our internal standards for answers should be. For example, This discussion about references shows discussions from February-August 2011.
I see a moderately sharp line around August 2011 when the modern standards started gripping. If I had to mark the watershed, I would say this guideline post from July 30, 2011 which summarised the conclusion of several debates, and which I still try to follow (and I should probably read more regularly as a reminder!)
Questions and answers grandfathered in before that date tend to be lower quality. Questions and answers after that date (that survive the regulars beating them into shape) tend to be higher quality.
That, IMHO, was the major transformation period. It is nice to think that the standards have continued to rise, monotonically, since then, but I am not sure I perceive it.
What is being done? What should be done?
I don't think there have any systematic attempts to go through the pre-August 2011 answers, and subject them to modern scrutiny.
Every now and again, I will go on a systematic jag of delete old abandoned answers which still have the citation-needed banner. (While the guidelines suggest only a week is given before an answer might be deleted, generally, by the time I get within a year of the current date, working oldest to newest, I am ready to stop.)
Sometimes an old answer will come to my attention - someone has flagged it, or someone has posted a newer answer which brings the old answer in front of our eyes, and it might get judged by the modern rules. (This arguably introduces a bias - that only offensive old answers get deleted or marked as requiring a reference, whereas inoffensive answers get by without. This is an argument for systematic clean up, rather than ad hoc.)
There's also the option of locking "historic" questions. I don't have the figures, but I think that's only happened a handful of times.
My suggestion: Let's implement this first. Then, we can have a systematic campaign to clean up old posts!