I’ll address your concerns.
If exit polls tend to have a bias with too many younger voters in the sample, then Edison Research must have realized it after the first few primaries, instead of not noticing it for eighteen primaries. They have decades of experience with exit polls, so they would have known about the bias beforehand. Further, since they adjust for each demographic’s response rate in the exit polls, this would cancel out the younger voter bias in the exit polls, making them reliable.
As for the second point, the voting machines could have been hacked by election officials, remote hackers, or anyone with basic computer programming knowledge. Most of them can be hacked at any time before the primary voting night, without a trace, since their security systems are at least 10 years old. Many people handle the machines between their construction and voting time, and it would only take one of them to hack them, and they would only have to do so in the most populated counties of each primary.
I agree that this is a result of the VRA being gutted; however, it still does not discount that this is voter suppression targeted in the majority towards one candidate. We can’t know who are the perpetrators of it are until we fully investigate it, and the research and data confirm that we must do so.
On the caucuses — this would be more of an indictment of caucuses, which I don’t believe we should even have in the first place. A more sensible solution would be having paper ballots hand-counted in public view, or with one of each of the candidates’ representatives there, in every state. The section you referred to focuses more on voter suppression, not on whether the system is broken.
It is true that some of the voter suppression comes from Independents not being able to vote in closed primaries. As an example to the contrary the New York primary changed their laws: independent voters had to change their registration more than six months before the primary to vote. Six months before the New York primary, very few people outside of Vermont knew who Bernie Sanders was, and everyone knew Hillary Clinton was going to join the race. This would limit Independents (especially younger voters) from switching their registration so they could vote for a grassroots candidate, while the longer-established Democratic voters could still vote for Clinton. Also, many younger voters are still deciding which party they want to join. Bernie Sanders might have helped them realize that they largely share Democratic values, but in New York, they were denied the chance to realize this and vote for him.
Lastly, it is not voter fraud being discussed, but electoral fraud. Voter fraud is when someone votes twice, or uses another method to have his or her vote count more than once. You’re right in that this is extremely rare, because it is almost never committed. With electoral fraud, however, we have no idea how extensive it is or how many elections it has skewed, because our election results are unverifiable. Independent investigations and audits into every Democratic primary this year would clear up if electoral fraud did occur.
All in all, thank you for raising these concerns. To clarify, I didn’t just leave out the information you mentioned, but also left out a lot of research that indicates voter suppression targeted towards Bernie Sanders is even worse than I wrote, because it would have made the article far too lengthy. I hope I addressed all of your concerns adequately; if not, feel free to ask more questions.