I am not a statistician. I gathered the following data for the state of Michigan:
1)
Party Shift from 2016 to 2020 Election by Michigan county.
2)
Determined which counties used Dominion Voting Machines.
I created two different populations from this data set:
The first Population was Counties that used Dominion voting machines accompanied with their Party Shift Percent.
The second Population was Counties that used voting machines other than Dominion accompanied with their Party Shift Percent.
Shift to the Democrat Party was recorded as positive, whereas shift to the Republican Party was recorded as negative.
Next: I used the F-Test to determine if these two populations have the same standard deviation / variance. The result with a high level of certainty P.001 was that they did have the same std dev / variance.
That allowed me to utilize the two-tailed T-Test for these 2 populations with the same variance. The P result of this T-Test was .105 Which is greater than my chosen Probability threshold of .001 used to determine statistical significance.
Research Hypothesis: The mean of Pop 1 Party Shift is different from the mean of Pop 2 Party Shift.
Null Hypothesis: The mean of Pop 1 Party Shift equals the mean of Pop 2 Party Shift.
Results: Inconclusive (Which is to say the means of Party Shift for Pop 1 & Pop 2 appear to be from the same population. This lends to the belief that the mean of Political Shift votes tallied with Dominion Voting Machines is comparable to the mean of Political Shift votes tallied with voting machines other than Dominion. The lack of difference in party shift is not a reliable indicator of which voting machines were used).