He is saying that the y-value should stay constant and not go down linearly. But the y-value is this: y=c-x where x is linear from 0-100% and c is the percentage of votes Trump got on ballots that were mixed (meaning not straight Republican or straight democrat vote ballots). In order for the value of Y to stay flat "c" must increase at the same rate as "x" so they cancel each other out. But his assumption is that as a precinct becomes more Republican (moving left to right on his graph) mixed ballot voters should go for Trump in an increasing number (the same rate as the x-axis). But mixed ballot voters are independents, Democrats and Republicans who encompass every voter that did not vote straight ticket (Dem or Rep). Why should we assume those voters would increase their Trump votes from 0-100% as a district becomes more Republican? We should not. The percentage of Independents voting Trump should not increase linearly as a precinct gets more Republican - that doesn't make sense. In fact the graph shows that mixed ballot Trump voters go from ~0-25% as you move from left to right on the x-axis until you get to 20%. Then it rises to ~35% and essentially levels off there moving right. At this point the "c" becomes a constant while "x" increases giving a linearly sloping downward "Y" result.
Their assumption that the percentage of Trump Only voters should track with percentage of GOP straight ticket voters is fine. You do realize some GOP voted Trump only right?
I think I'm seeing Filetsmignon's argument. I'm not sure you can assume the remaining non-R population would vote for Trump at the same rate as the R population. If the non-R population voted for Trump at a consistent rate, which is a more logical assumption, the graph as constructed would always slope down as you move right along the x-axis.
You are saying that %TO is people who came in and voted for Trump only and they did not vote for anything else on the ballot - just a single vote for Trump and left everything else blank. That's not correct. What Shiva is saying is that y=(%TO)-x where %TO (to use your variable) is the percentage of voters who voted for Trump but ALSO voted for other candidates - which makes them a mix of Reps, Dems, Indy, Green, etc. That's an important distinction because that %TO variable represents Dems, Reps, Indy, etc. voters who voted for Trump. And I think we can likely agree that they are probably heavily Independents because they didn't vote straight party Rep or Dem - but we don't know the mix. Here is Shiva explaining what the two types of voters are - tell me if you get a different sense from his explanation and whether you still think %TO represents people who only voted for Trump and left the rest of the ballot blank or whether you think they are people who voted for Trump but also voted for other party candidates down ballot. https://youtu.be/Ztu5Y5obWPk?t=920
I understand the variables, but the logic is squishy. There are two populations we are considering, the straight R population and everyone else (the non-straight R population). As you move right along the x-axis, the non-straight R population gets smaller and smaller. Why should we assume that the non-straight R population would vote for Trump at an increasing rate the more republican a district? All I'm saying is that a better assumption is that the non-straight R population would vote for Trump at a consistent rate. If my assumption is correct, the graph has to slope down.
He is saying that the y-value should stay constant and not go down linearly. But the y-value is this: y=c-x where x is linear from 0-100% and c is the percentage of votes Trump got on ballots that were mixed (meaning not straight Republican or straight democrat vote ballots). In order for the value of Y to stay flat "c" must increase at the same rate as "x" so they cancel each other out. But his assumption is that as a precinct becomes more Republican (moving left to right on his graph) mixed ballot voters should go for Trump in an increasing number (the same rate as the x-axis). But mixed ballot voters are independents, Democrats and Republicans who encompass every voter that did not vote straight ticket (Dem or Rep). Why should we assume those voters would increase their Trump votes from 0-100% as a district becomes more Republican? We should not. The percentage of Independents voting Trump should not increase linearly as a precinct gets more Republican - that doesn't make sense. In fact the graph shows that mixed ballot Trump voters go from ~0-25% as you move from left to right on the x-axis until you get to 20%. Then it rises to ~35% and essentially levels off there moving right. At this point the "c" becomes a constant while "x" increases giving a linearly sloping downward "Y" result.
Their assumption that the percentage of Trump Only voters should track with percentage of GOP straight ticket voters is fine. You do realize some GOP voted Trump only right?
I think I'm seeing Filetsmignon's argument. I'm not sure you can assume the remaining non-R population would vote for Trump at the same rate as the R population. If the non-R population voted for Trump at a consistent rate, which is a more logical assumption, the graph as constructed would always slope down as you move right along the x-axis.
Not sure you understand the variables.
Y axis is %TO - %GOPstr8
X axis is %GOPstr8
%TO = Trump Only votes / all Only votes
%GOPstr8 = GOP straight ticket votes / all straight ticket votes
The denominators are different. It's more plausible TO and GOPstr8 votes would track than not.
You are saying that %TO is people who came in and voted for Trump only and they did not vote for anything else on the ballot - just a single vote for Trump and left everything else blank. That's not correct. What Shiva is saying is that y=(%TO)-x where %TO (to use your variable) is the percentage of voters who voted for Trump but ALSO voted for other candidates - which makes them a mix of Reps, Dems, Indy, Green, etc. That's an important distinction because that %TO variable represents Dems, Reps, Indy, etc. voters who voted for Trump. And I think we can likely agree that they are probably heavily Independents because they didn't vote straight party Rep or Dem - but we don't know the mix. Here is Shiva explaining what the two types of voters are - tell me if you get a different sense from his explanation and whether you still think %TO represents people who only voted for Trump and left the rest of the ballot blank or whether you think they are people who voted for Trump but also voted for other party candidates down ballot. https://youtu.be/Ztu5Y5obWPk?t=920
UPDATE: here is good explanation between the straight party vote and individual vote in Michigan: https://www.lansingstatejournal.com/story/news/politics/2020/10/24/straight-ticket-option-michigan-ballots-2020-november/3726907001/
I understand the variables, but the logic is squishy. There are two populations we are considering, the straight R population and everyone else (the non-straight R population). As you move right along the x-axis, the non-straight R population gets smaller and smaller. Why should we assume that the non-straight R population would vote for Trump at an increasing rate the more republican a district? All I'm saying is that a better assumption is that the non-straight R population would vote for Trump at a consistent rate. If my assumption is correct, the graph has to slope down.