I took a look through it and it rests on a couple of dubious assumptions:
-
Biden and Clinton can be assumed to have identical levels of support
-
There is no association between candidate preference and choosing to vote by mail or in person
If you applied the same logic, Trump's massive lead on election night would also appear "impossible"
I think the main issue with all of the analysis of NYT data is that the data is not official - it comes from Edison and is produced for the purpose of media reporting on election night. I don't know if there's full details anywhere on how the reported data is churned out, but it probably uses stuff like weighted races and exit polls to project vote totals, rather than just straight using results reported by county.
There's jumbled data in a lot of states from what I've seen - where the vote totals are out of sync with the timestamps. A lot of the irregularities disappear if you sort by vote total instead of timestamp.
https://rpubs.com/ElectProject/Early_Vote_2020G
Scroll down to PA, the data is for the primary, and the number of ballots requested (1,823,148) is exactly the number that Rudy quoted.
Yeah it's the NYT scrape, right? I think they're talking cumulatively over the various "spikes", but they don't mention their inclusion criteria for "spikes". Based on the other analyses of this data that I've seen, I'm guessing they had a one-way explanation of what constitutes a "spike", i.e. they counted stuff skewed toward Biden but excluded stuff skewed toward Trump.
Part of the difficulty with evaluating these soundbyte type results is recreating the results from the same data, and then evaluating if the methodology made sense. By the time you've done that, the idea is already loose.
I've also looked through this, the big anomalies are where the overall vote total decreases and goes out of sync with the time stamps. This could be a problem with the scraper, or something between NYT and Edison ( it's not official data)
Still, why would you want to look at the actual data behind a damning claim when you could just repeat it without scrutiny? That's why your comment is down in the cheap seats.
https://rpubs.com/ElectProject/Early_Vote_2020G
Scroll down to the data for the PA primaries. The number quoted by Rudy exactly matches the number of ballots sent out for the primary. There's no question that it's just the wrong number.
Giuliani is using the number of ballots requested in the primary election, not the general.
Source is for the primary election, not the general: https://data.pa.gov/Government-Efficiency-Citizen-Engagement/2020-Primary-Election-Mail-Ballot-Requests-Departm/853w-ecfz
This is why you shouldn't share stuff before verifying it, and you definitely shouldn't testify in a hearing without checking it
Hijacking this to say that this number is just a screw-up by whoever made the image. 1.82 million is the number of ballots for the primary election in June, not the general. Data seems to be sourced from here: https://data.pa.gov/Government-Efficiency-Citizen-Engagement/2020-Primary-Election-Mail-Ballot-Requests-Departm/853w-ecfz
From the primary election in June: https://data.pa.gov/Government-Efficiency-Citizen-Engagement/2020-Primary-Election-Mail-Ballot-Requests-Departm/853w-ecfz
I think this site just shows the present time lol