https://www.pewresearch.org/short-re...lling-in-2024/
General article about some of it, obviously impossible to know how accurate the polls will be this cycle and past performance is not always an indication of future performance although obviously pollsters (the non partisan ones anyway) want accurate results or their product is worthless. Partisan polls for either party are obviously serving a different purpose (to push whatever narrative the party wants) so should be discounted
2022 polling was pretty good, previous few cycles were a miss, polling had Romney closer than he was, Clinton and Biden both up more than they ended up being in national popular vote but again pollsters adjust to try and get their samples to better match the electorate as a whole over time if they realise they're over or under sampling certain likely voters etc in hindsight so who knows if they'll be accurate or off this cycle and if they're off, by how much as there's a huge difference between a 1% polling error and a 4-5% polling error
Hypothetically if we go into election day with Harris up 3-4% or whatever, she's definitely the favourite, but it could be anything from a landslide (Harris wins national vote by 7% and sweeps the swing states or similar) or if there's a 3% gap in the other direction and Harris wins the popular vote by 1% and the electoral college is close or Trump edges it
If we go into election day with polls tied nationally, Trump's probably a fav due to electoral college math and the last two cycles polling errors but again past performance does not guarantee future performance in that the last time there was a black candidate for President in 2012, polling actually under-sampled Obama's support and over-sampled Romney's.
A lot will depend on turnout too. If turnout is low, conservatives tend to do well, if turnout is high, progressives tend to do well. The right needs to try and disillusion the left over Palestine etc and try and paint Harris as 'both parties are the same/bad' to try and get people that will never vote for a Republican to stay home. If youth turnout (or single female or black turnout) is super high, Republicans are ****ed. If they're super low, Democrats are ****ed.
I think it's going to come down to turnout again. Trump got 74m to Bidens 81m votes last time. USA has about 6m more citizens than last time, let's say 4m are eligible voters and 3m of them actually vote - i'd expect Trump gets about 74-75m votes again (i'm expecting him to be slightly less popular as a non incumbent and not running against a super old dude). If Harris matches Biden's number turnout wise, Trump won't win. Now turnout could be higher or lower with Trump out of office but i'd say if Trump gets to 80 million votes he's a favourite, by 70 million he loses in a massive landslide and loses all of the swing states. In 2016 it was 66m for Clinton 63m for Trump approximately, but Trump ran well expectation with where the votes landed even with Republicans EC advantage. If he loses the popular vote by 3 million or so he's probably a tiny dog, if he loses the popular vote by closer to the 7 million he lost to Biden by he isn't going to win but it could be anything from a 1-3 state victory to Harris sweeps the swing states depending on the distribution. By 8-10 million votes i'd expect Harris covers the -99.5 electoral votes line and wins 7/7 swing states. If Trump wins the popular vote outright I don't see a scenario where Harris wins the electoral college she would have to run way, way above expectation in the exact 270 electoral votes she needs (probably the midwest plus Nevada or ME district route) and massively underperform in the sun belt or similar
Anyway, none of this is particularly helpful but there are a ton of similar articles to the one I linked at the start of the post that discuss how pollsters are trying to change their methodology to get accurate samples. Polling was quite good in 2022 overall in that people expected the polls to underestimate Republican support and form a red wave and it never eventuated.
It's more likely that polling under-samples Trump voters than over-samples them but it's not impossible that the reverse is true or that the polls are accurate within a percentage point both of those are very possible as well.
If black or female turnout vs white male turnout is significantly higher or lower than expected or if young people show up in force to vote or don't show up at all it'll obviously have a huge impact