Spousal Voting Behavior should be Analyzed

Great Debates
61

In case anyone thinks that protecting the vote of domestic violence victims is of any importance to conservatives, here’s a tidbit.

Many states have programs that assist the former victims of domestic violence in concealing their whereabouts from their abusers. These laws provide a mechanism for them to keep their personal info, including their address, from appearing in public records such as tax rolls and voter registration lists.

Of course, the people that are concern-trolling ( I’m not talking about anyone on this board, I’m talking about right wing public figures) about election integrity have seized on this and it’s one factor at the root of the “hundreds of voters registered with only PO Boxes as addresses, even though that’s illegal” - election fraud claim.

Even worse, Minnesota is trying to implement such a program, and it’s being stalled because some people seem to think it’s a cover for a voter fraud operation.

I’ll say it again…I’m pretty sure millions of women in the US vote the way their husband tells them to, whether from tradition, religion or just plain indifference. As a Democrat, I think it costs us votes, but I think women are gradually becoming more informed and that the practice will diminish with time. I also think most of those women would vote the way their husband wants them to whether or not he was watching.

But it’s not illegal or fraudulent unless you throw in bribery, blackmail or coercion. It’s not even particularly wrong.

If a woman is being threatened or coerced by her husband into doing something she doesn’t want to do, it’s abuse, it’s a crime and it should be addressed. But the problem here is not voter fraud.

You are looking to help the abused woman who is too scared or intimidated to tell her husband she wants to cast her ballot in private, or to engineer a situation where she can vote as she’s wishes, yet not too scared to go into a voting booth and defy his wishes, even though he’ll probably question her and she’ll have to lie, even though he may have convinced her that her vote isn’t private.

I’ve known many abused women, and that’s a pretty narrow sweet spot. It’s ridiculous to tailor legislation around it. If you want to combat this “spousal influence”, I’d start with public service advertising convincing woman to research issues themselves and emphasizing the fact that your vote is private and no one, not even your husband, can find out who you voted for - because I bet many abusive husbands gaslight their wives on this.

62

I asked before how you separate out all the non-spousal adults at an address from the spouses, but you didn’t answer.

Nor did you answer how to get a baseline for the in-person anonymous votes.

Determining who’s a spouse of whom is the critical step, before the discussion turns to hash values.

Hey, do we even know if mail-in ballots must specify an apartment inside of an address?

63

Really, any set of co-habitants is relevant, not just specifically legally married couples. Probably the larger concern would be if there’s any situation where you have like 30+ people who all have the same exact address.

I’d venture to guess that you wouldn’t gain any real additional data by adding in threesome or foursomes, and if you start including groups of 30+ co-habitants, that might allow someone to figure out the original address through logical deduction. I’d probably just crop it as exactly two during team 3’s step.

I’m not sure what you’re asking.

Any process that you can apply to mail-in, you can extend to walk-in. Maybe with the specifics of the setup in a particular state it might make sense to do something else, though, so I don’t really want to specify that there’s only one true way of doing it, just that there is at least one and probably many.

64

I wouldn’t say that “millions of women vote the way their husbands tell them to.” Thousands, perhaps, and the kind of men who are the kind of people Ann-Hedonia described probably wouldn’t let their wives vote anyway.

65

Men can be spouses, too. I don’t see any reason to believe that it’s purely or even majority single-direction male to female.

66

I’m still confused about the anonymity. Will the company see two votes from 123 Poplar Street and see two Republican votes? That would allow anyone to know who I and my wife both voted for.

Or if we split one D and one R, what is to stop a union or corporate hack from pulling me aside and telling me that they know I’m good, but I better get the wife in line? Or God forbid, 2 Ds and that is the wrong way for my conservative boss?

Am I missing something? Does the two step process only say “same household”?

67

You’re not missing anything. The end result of @Sage_Rat 's process is a database tying addresses to ballots. He does put some tiny speed bumps in the way of just having that information out in the open at the cost of maybe tripling the number of election workers.

All this to investigate a “problem” that he can’t point to a single known case of.

68

No. They would see that there were two majority Republican votes at 63178852257964311693741, which is just a randomly generated number that has no connection to and can’t be returned back to an address.

(For purposes of correcting for any urban vs. rural voter differences, the zip code might be chosen to be passed through.)

Yes.

If you had a person on team 1 with eidetic memory and a second one on team 2 with eidetic memory, and those two were friends, then my technique would have a problem. Given that that’s largely a TV fantasy, preventing any of the participants from using their phone, a camera, or taking notes should be sufficient to ensure that the process is fully anonymous. And I would expect that a security expert (e.g. Ron Rivets) could improve the system to remove that opening.

69

Growing up in the 80s, your average popular movie had guys pushing themselves on women. Your average rock star was sleeping with underaged groupies. Forty years later, people started to talk about how that was wrong and they started to reveal their stories.

Is there any rule that if you just give people forty years, they’ll figure things out? I would say no, because the movies of the 60s, 50s, 40s, etc. were also all creepy. For most of human history, men probably didn’t treat women the way that women wanted to be treated. The learning cycle is very far from being guaranteed to be forty years and that’s with something where you can see the guys taking to girls, in public, at bars and dances and being creepy in front of people. Here, we’re talking about private moments at home.

No one will tell you that Coca-Cola forced them to drink their drink. But if you did a double-blind taste test of all cola drinks in the world and ranked them, Coca-Cola would likely be somewhere in the middle. We would expect the most delicious to be the most common. Coca-Cola wins through advertising and by using contracts to create geographical monopolies. But, again, no one person will claim to you that he, himself, was factually manipulated and coerced into buying Coca-Cola.

This is the real world. Choosing to not look for proof is not the same thing as evidence that something isn’t real.

70

I actually agree with @Sage_Rat that technically this could be analyzed using a secure one-way hash.

And I think I agree that there is the possibly that mail-in ballots have a higher probability of “same voting” rate between spouses (since it is much more likely that the ballot is not “secret” between spouses).

It’s an interesting point, and one I haven’t really considered before.

I wonder if anybody has done an analysis comparing polarization in areas with primarily mail-in vs. in-person voting. I’m imagining that what you would see if this were really a significant issue is that results would be more skewed towards one position/party/candidate in mail-in regions - you would have fewer “close” elections, because potential split households would no longer be split.

To give a personal anecdote, I often give guidance to my spouse on how to vote on things like local school board or county ballot issues. She often doesn’t have time to research these and knows that our views align enough to generally vote the way I advise. If we voted mail-in, I imagine she would vote 100% in line with my recommendations on these ballots (possibly even having me double-check before sealing) but in-person it’s likely she might forget, etc and thus our ballots would not align.

An interesting point is that it seems likely (to me at least) that this bias would tend to hurt Democrats due to the general shape of spousal pressures in America. But the increased access to voting offered by mail-in ballots tends to help Democrats.

71

But thinking that something might be a significant problem isn’t the same thing as that thing actually being a significant problem. Sure, some spouses almost certainly would stand over their spouse’s shoulder to make sure they fill out a mail-in ballot correctly. But nothing prevents those same people from demanding that the spouse allow them to go into the voting booth or preventing the spouse from leaving the house to vote at all. And there are definitely cases where one spouse influences the other’s vote - but I’m not sure why there’s a problem if I vote a certain way to please or shut up someone I live with but apparently no problem if it’s to please or shut up someone I don’t live with. Or how that’s really all that different from my spouse influencing views that I believe are my own, like in your Coca-Cola example.

And that’s not even taking into account that a lot of that influence can’t really be identified - I sort of came back into contact with an ex many years later through Facebook. And one thing that almost immediately struck me is that if we had stayed together, either one of us would be very different than we turned out to be or one of us would have filed for divorce in 2016. And if one of us had turned out differently, it would have been impossible to know how much of that (if any) was due to the influence of the other.

72

Walk me through the steps. A secure one-way hash, when the inputs are public knowledge, will, at best, be a minor inconvenience to someone trying to de-anonymize ballots.

73

It seems to me you could hash addresses, then have a second stage match hashed pairs and give each pair a unique ID, then destroy the hashed values. All you would know is pairs of votes with no data to extract addresses.

Obviously there is the possible vulnerability of not properly destroying the hashes. You would have to have trusted actors at that step, but you already have that requirement at the “remove ballots from signed envelope” step in the current system.

But I could be very wrong here - it’s been awhile since I thought seriously about these technical issues. I would also agree that even having a temporary database of votes with hashed addresses is a risk that may not be worth taking to get this data.

74

The one way hash is used to obfuscate the address into something like B6FCA5206DB37AC0E5C4564828AA68A9516D4365 between team 1 and team 2. That value is then dropped (erased, destroyed, tossed, rendered unto void) in the final output and replaced with a random value (e.g. 15358842246753785312286229) that has no connection to the address. It is just a random number.

Again, there’s a theoretical issue that someone on team 1 would memorize one of these values, which is of interest, and a friend in team 2 would remember every single value he saw, how each of those people voted, and have seen every ballot. In that situation, the guy in team 1 could ask the guy on team 2 what the vote was for B6FCA5206DB37AC0E5C4564828AA68A9516D4365, and he would be able to answer. But that’s really not practical without taking notes on everything.

My expectation is that the people in the teams will have had their devices confiscated, will be too busy, and will be watched.

75

In this case, team 1 wouldn’t be able to pass things to team 2 until all the ballots were collected or someone would have to keep a list matching hashed addresses to UUIDs until all ballots were collected. Several someones actually.

76

And to what end?

Suppose you learn a lot of couples vote ballots in identical ways. What do you plan to do with that information?

Go out and interview couples to determine if they voted independently?

What if they didn’t? What if they sat together at a table and arrived at mutually agreed-upon votes they both liked? Lots of couples do this.

What if they argued over it, but in the end, one spouse decided the other made good points in their argument and capitulated? Lots of couples do this, too.

Who conducts these interviews? How much latitude do they have to accuse a couple of not voting “independently enough”?

Is this really a path for which you have any justification to pursue?

77

No. I wrote the specific methodology. Your assumption that there’s more to it is not correct.

Yes, if you add more then it’s less secure. Don’t add more.

78

If you don’t add more, it doesn’t work.

If one person from a household sends their ballot in early, and another waits until the last minute, team 2 will not be able to assign the same random number to this pair of ballots unless they wait to assign the number to the first ballot until the second comes in or someone keeps a list.

79

If you type “123 Poplar Street” into this site you will get B6FCA5206DB37AC0E5C4564828AA68A9516D4365. That will still be true in a year.

For the situation that we are dealing with, a more complicated system is needed than simply using sha-1 on the address directly. You would probably need to do something like creating a random salt and including that as a secondary value on the inner envelope, which would be invisibly used by team 2s software so they never actually see the address hash directly.

80

I know how hashing works. I’m a data scientist with a PhD in mathematics who has studied cryptography.

You don’t seem to understand the flaw I’m pointing out here.

Team 1 turns plaintext address 1 into hashed address 1. Then team 2 turns hashed address 1 into random number 1. Then they destroy any link between Hashed address 1 and random number 1.

Some time later team 1 receives another ballot from plaintext address 1, turns that into hashed address 1 and passes it to team 2. By what mechanism can team 2 turn this into random number 1?