Dean

“That is essentially assuming the population grows at the same rate the set of people who could vote grows.”

No, I am essentially avoiding that assumption. And, at least for the period of we are looking at, the assumption is not correct; there was, for example, a change in 1971 with the passage of the 26th amendement when 18-2-year-olds were added to those eligible to vote, with no corresponding change in population. It seems that the use of reported voter turnout implicitly makes that assumption.

If the proportion of eligible voters grew at the same rate as the population, the two measures should be the same, except for a scaling factor. But they are not the same, with one going up and the other going down.

So the question is, to what degree does the fall in voter participation numbers represent a fall in actual voter interest in voting. Voter participation number can fall because fewer people are voting, or because the pool of eligible voters rises.

Take for instance the 19th amendment, approximately doubling the number of eligible voters. If women voter at the same rate as men, the voter participation rate would stay the same, although the proportion of the population who voted would approximately double. If no woman showed up to vote, the voter participation rate woud fall approximately in half, but the proportion of the population would stay the same.

It seems that finding the proportion of all people who voted is a better indication of democratic participation than just the proportion of those elegible to vote if the eligibility to vote is restricted. A highly restricted franchise seems to be less democratic than a broad franchise, even if a smaller fraction of those eligible to vote exercise that franchise.

“regressing the number of votes on the number eligible gives a negative slope but not a slope you gave.”

If would be interesting to know what number you did get for that, I could do that calculation to check my figures. But that’s not the calculation I did; my calculation was a regession for reported turnout over time. Perhaps I wasn’t clear. Certainly it is no surprise that that slope be negative.

“That regression gives a lousy fit”

Yeah, the data is noisy, and we cannot put too much reliance on the exact numbers. But regressive is is an easily available indication of whether voter participation is overall going up or down. In spite of the noise in the data, it does appear that the reported voter participation rate is going down, and this is one quick way of checking that appearance. Therenote = [[

Dean

“That is essentially assuming the population grows at the same rate the set of people who could vote grows.”

No, I am essentially avoiding that assumption. And, at least for the period of we are looking at, the assumption is not correct; there was, for example, a change in 1971 with the passage of the 26th amendement when 18-2-year-olds were added to those eligible to vote, with no corresponding change in population. It seems that the use of reported voter turnout implicitly makes that assumption.

If the proportion of eligible voters grew at the same rate as the population, the two measures should be the same, except for a scaling factor. But they are not the same, with one going up and the other going down.

So the question is, to what degree does the fall in voter participation numbers represent a fall in actual voter interest iin voting. Voter participation number can fall because fewer people are voting, or because the pool of eligible voters rises.

Take for instance the 19th amendment, approximately doubling the number of eligible voters. If women voter at the same rate as men, the voter participation rate would stay the same, although the proportion of the population who voted would approximately double. If no woman showed up to vote, the voter participation rate woud fall approxiamtely in half, but the proportion of the population would stay

It seems that finding the proportion of all people who voted is a better indication of democratic participation than just the proportion of those elegible to vote if the eligibility to vote is restricted.

“regressing the number of votes on the number eligible gives a negative slope but not a slope you gave.”

If would be interesting to know what number you did get for that, I could do that calculation to check my figures. But that’s not the calculation I did; my calculation was a regession for reported turnout over time. Perhaps I wasn’t clear. Certainly it is no the that slope be negative.

“That regression gives a lousy fit”

Yeah, the data is noisy, and we cannot put too much reliance on the exact numbers. But is is an easily available indication of whether voter participation is going up of down. In spite of the noise in the data, it does appear that the reported voter participation rate is going down, and linear regression is one quick way of checking that appearance. There may be other ways, and they may be better, but for a quick, back-of-the-envelope type calculation it seems not unreasonable.

“Using total population as the base is an attempt to measure democratic participation in a way that elegibility to vote can, and in some cases does, hide. ”

That is essentially assuming the population grows at the same rate the set of people who could vote grows.

But more directly — did you get the population data from your link above or from somewhere else?

Just playing with the data from your previous link

> regressing the number of votes on the number eligible gives a negative slope but not a slope you gave (and it isn’t a percent since none of the data are percentages)
> That regression gives a lousy fit — lots of curvature in the residuals (due to the lack of linearity seen in other plots) AND the votes in 1982 and 1994 were identified as having high amounts of influence on the fit. As bad a measure a p-value is, removing those two votes from the data set dropped the p-value by a factor of 4. Those issues tell me that looking at this with anything related to regression is
worthless

I’m still not sure where the general population values came from

For the voters/population figure, I took the total voters form the same Secretary of state’s website, and the historical population figures from the large table at the bottom of this page:
https://en.m.wikipedia.org/wiki/List_of_U.S._states_by_historical_population
Since that data only goes to 2015, I googled “Minnesota population 2016” and used the census bureau’s figure.

But I think this is just a sideline to the overall problem which is that political parties don’t represent the interests of their voters, but rather a subset of those voters that have influence. When it comes to serving the best interests of the country as a whole that’s out the window entirely!
I believe this is actually the biggest driver behind dropping voter turnout – the increasing understanding that the people we elect will not represent our interests. This also contributes to the tribalistic voting that dominates current politics – the reason to support a party doesn’t depend on it’s platform if no platform actually matches your own interest.
This is a much more difficult problem to solve, but the only one worth solving anyway. Even if more people vote, what does it matter if the elected representatives don’t actually represent them anyway?

The first is simply practical – it’s available. I can find the yearly estimates of total population for the states, and I don’t right off know of a source that gives me, say, the population of each state over 18, for example.

But the more important reason flows out of democratic theory. Voter eligibility can be, and has been historically, limited and changed. And limiting elegibility to vote can be an effective anti-democratic tactic. Until 1920 women could not vote in most states, when the 19th amendment became ratified, those eligible to vote suddenly doubled. Literacy requirements limited elegibility to vote. Race has been used to limit elegibility. Age has been, at one time 21, not 18.

In general, expansion of the franchise is seen as a pro-democratic good, and looking only at voter eligibility can make a very un-democratic system look more democratic that it is. Limiting the franchise to adult, property owning white men? – it’s been done, and is not very democratic, even if it produces a high turnout rate.

Using total population as the base is an attempt to measure democratic participation in a way that elegibility to vote can, and in some cases does, hide. Is it perfect? No, but it is, from a democratic participation standpoint, probably better than such measures as percent of register voters.

Running a linear regression line though the general election turnout data reported by the MN secretary of state, the slope is indeed negative, going down by about a 0.32% a decade. But the regression line through the proportion of voters per total population has a positive slope, going up by about 1.4% per decade.

What data are you getting these runs from? Not from the final link you posted?

“the totsl population was 5,519,952,”

Are all of those people eligible to vote?

There is more to be concerned about — the person currently pegged to head the 2020 Census is on record as being in favor of racial gerrymandering — he’s the ass-clown responsible for North Carolina’s redistricting attempt to marginalize and suppress black voters. That, coupled with the DOJ’s attempts at tampering with questions on the census and the lack of funding for it, shouldn’t make anyone with an understanding of the Census’ purpose and/or correct statistical methodology very nervous.

It seems that what we are really interested in is the percentage of people who actually vote. So, instead of reported voter turnput data, another measure we can use is the percentage of the population that actually vote, which is quite different than reported turnout.

For example, in the presidential election of 2016, the reported turnout in MN was 74.72%; there were 2,968,281 voters, and the totsl population was 5,519,952, or about 54% of the total population voted. In contrast, in the 1952 presidential election, the reported turnout was 77.22%, and there were 652,825 voters andthe population of MN was about 3,030,000, so only about 22% of the total poulation voted. So despite the fact the reported turnout fell slightly from 77% to 75% the proportion of the population that voted rose from 22% to 54%.

Running a linear regression line though the general election turnout data reported by the MN secretary of state, the slope is indeed negative, going down by about a 0.32% a decade. But the regression line through the proportion of voters per total population has a positive slope, going up by about 1.4% per decade.

So, from the viewpoint of actual voter participation, things are much better now than in the 50’s and 60’s, and has been moving in the right direction, despite reported turnout data seeming to show the opposite.