Has polling demographics become politically correct?

I have been using a spreadsheet to track the popular vote Presidential race going back to the 2004 election. I have used a variety of weighing and demographics, but the number one demographic I have always relied on the most heavily is partisan breakdown. As in Republican, Democrat and Independents. The use of the partisan breakdown (among other things) along with a fairly traditional and basic manner of weighing the electorate has allowed me to come pretty damned close nearly every election. While the 2004, 2008, and 2012 elections were not necessarily surprises and didn't take much rocket science to get those close, not many people got the 2016 election popular vote as close as I did.

But lately I have found that many pollsters have been providing mainly demographics built almost entirely on your physical makeup. Gender. Age. Income. Race. Religion. There even tends to be an almost universal inclusion on what part of the country you are from, which seems bafflingly irrelevant. But there seems to be quite a few more pollsters who no longer appear to include any make up (or possibly don't even ask) about political affiliation. Even many of those who do, tend to micro-analyze down "leaners" which are then moved into the Republican or Democratic category, leaving an extremely low amount of independents to sample from which can lead to wildly different results. I saw one cross tab that appeared to show about a 50 point spread with Independents, at which point I was forced to throw it out since Independent make up about 30% of the electorate in every election. All of this shifting from a traditional polling demographic model to whatever we are seeing today seems puzzling at best and troublesome at worst. 

I read a couple of articles over the past couple of months suggesting that the use of political affiliation was a poor manner in which to weigh a poll and that it should be eventually just left out completely. There was little logic or explanation implied other than it appeared that new wave of people running your polling want something more tangible and relevant to how people view themselves today. In other words, what sex you are, how old you are, what race you are, what part of the country you live are not subjective to anything but a static answer. It is who you are, while what political affiliation you are is being considered less relevant because you decide it rather than being born into it. Or possibly many people just take more pride in other characteristics and therefore political affiliation is unimportant. 

I get that this is the year 2020 and everyone wants to be labeled and associated with some kind of group and I am pretty sure that the younger generation is not too much on political affiliation labels. They would rather be tied to their race, tied to a creed, or one of the 39 different genders that are currently out there. Being a Republican or Democrat is probably old school to them and yesterday's concept of what politics is made up from. But by deciding to totally ignore partisanship, at the expense that you get the proper amount of pansexual vs bisexuals vs agenders, that may well be missing the point. What a person wants to tell you about themselves isn't always what you "need" to know. Some things in life cannot just be what millennials want to make them to be.  

It's not to say that I have not found enough pollsters to create a 2020 spreadsheet. It's just not as robust as what I had seen in the past. I might have 60% of what I used to work with, and almost no daily tracking polls (which provide daily feedback). Moreover, even within those pollsters that provide what I would generally need, I find some of it cryptic, unusually reported, and mathematically hard to associate with the top line numbers. I find myself plugging here and there and hoping to see something that makes sense, occasionally giving up and tossing the poll out for lack of any semblance of statistical consistency. 

In 2016 I stated over and over that I did not believe the large state polling leads for Clinton or the multiple double digit lead polling for Clinton during the month of October. Obviously I was right. But right now I am having issues really "trusting" much of the polling at all, or at least to the degree that I feel comfortable making a hard call. That being said, I stand behind what I have for the most part. I still think Biden is up by around the figure I have. But I certainly wouldn't bet the mortgage on it.