What is the generally accepted difference between registered voters poll sampling and likely voters? Can it be adjusted accurately to mean anything useful?
Qpac say in their methodology paragraph that they will change to Likely Voters vs Registered later in the year.
Likely Voters are identified several ways and not all polling is the say. They may ask how likely are you to vote, they may ask in how many recent elections did you vote, and they may actually look up the records for if the respondent voted.
I do think the advantage in the past the GOP had with likely vs registered likely has faded because voting is easier now. VBM etc. This is why the GOP must embrace VBM. There are lazy voters on both sides of the fence and to just award that category to the Democrats makes no sense.
Maybe even more important is that the GOP and Democrats are fringe parties now. Independents are the overwhelming bulk. So targeting weighting by party likely is inaccurate.