Post Voting Discussion

[livius drusus]

The Post Voting System (http://www.freethought-forum.com/forum/showthread.php?t=50) is the key to what we hope to accomplish here. We didn't want a forum where free speech was an excuse for abuse, just as we didn't want our desire to protect people from abuse converted into stringent rules and hardcore moderation.

Anyway, it's all there in the thread above. Bring on the ruthless criticism. :)

[Adam]

I'll save the ruthless criticism until I've seen it in action but, juding solely from what you've written describing it, it's a beautiful thing. I'm glad that you decided to go with such a system, and I'd like to thank galiel (if he's around) for selling me on the merits of such a system.

[viscousmemories]

I'll save the ruthless criticism until I've seen it in action but, juding solely from what you've written describing it, it's a beautiful thing.
We really hope so.

I'm glad that you decided to go with such a system, and I'd like to thank galiel (if he's around) for selling me on the merits of such a system.
He hasn't come by yet, but I hope he does. I'm almost positive we would not have gone the route we did with the site development without his push in the direction of self and community governance instead of yet another authoritarian structure. :yup:

[lisarea]

The only real problem I can see with that sort of thing is that you'd need either a large or very selective group for it to work.

They do this sort of thing, basically, on /. and kuro5hin, but it works because there are enough people there to balance the wingnuts. Or to establish a status quo for wingnuttery at least.

Plus, once, on kuro5hin, this guy was saying some really dumbassed shit, and I kicked his ass, so he and his friend went through my entire post history and modded every single post I'd made there down. Which didn't affect my reputation much, but if it had been a smaller board, it could have.

On a smaller board, it runs the risk of looking like a popularity contest, which will cause me to complain bitterly, as I am wont to do.

So I guess it depends. Do all votes count the same? Is the moderation transparent? I am all kinds of tempted right now to go screw around with it a bunch to see how it works, but I won't.

Really, though, I think it's a good idea. I'm just advocating for my Dark Lord Satan.

[livius drusus]

I'll save the ruthless criticism until I've seen it in action but, juding solely from what you've written describing it, it's a beautiful thing. I'm glad that you decided to go with such a system, and I'd like to thank galiel (if he's around) for selling me on the merits of such a system.

I hope he'll be around too. It's not quite what he had in mind since it is, after all, just vB and nowhere near as flexible as the slashcodey kinda thing he had in mind, but we do have options and it sure is purtier than Slashdot. :tongueout:

[livius drusus]

The only real problem I can see with that sort of thing is that you'd need either a large or very selective group for it to work.

Yup. It's designed for large numbers of people and multiple levels of voting points. We've got it down to its barest bones right now because the group is so small, but it really is meant for a growing community.

They do this sort of thing, basically, on /. and kuro5hin, but it works because there are enough people there to balance the wingnuts. Or to establish a status quo for wingnuttery at least.

We checked out kuro5hin during the research phase. A status quo for wingnuttery is a great description of the aims of the system, except we have a wee stick in the form of a Bad Reputation group which doesn't get to play with the coolest toys.

Plus, once, on kuro5hin, this guy was saying some really dumbassed shit, and I kicked his ass, so he and his friend went through my entire post history and modded every single post I'd made there down. Which didn't affect my reputation much, but if it had been a smaller board, it could have.

That is our single greatest concern at this point. Our original idea was to have mods and admins with votes worth 5 points so we could step in and quickly counter any attempts to twist the system to nefarious ends. vm and I still have one privilege: we're the only ones who can vote on a post more than once, so in case of dire emergency, we could stem an asshole tide.

We do have the capability of making any number of groups with varying vote power. For example, we could set up an automatic promotion system where anyone who reaches x rep points gets y voting points. We could also make this time-limited, so the electors would revolve a la Slashdot mods. For now, though, we thought it best to start off simple and see where it takes us.

On a smaller board, it runs the risk of looking like a popularity contest, which will cause me to complain bitterly, as I am wont to do.

What if we make sure you're one of the popular ones? Will you still complain bitterly then?

So I guess it depends. Do all votes count the same? Is the moderation transparent? I am all kinds of tempted right now to go screw around with it a bunch to see how it works, but I won't.

Okay. Yes, right now, every vote is worth a single point. The voting is anonymous; neither users nor admins have any idea who voted what. Everyone starts out with 10 rep points. It would take 100 down votes to put you in the bad rep category, so since you would need no up votes, 100 posts and one enemy, or no up votes, 50 posts and two enemies, you're unlikely to get screwed by a wingnut even now when we're still small.

Really, though, I think it's a good idea. I'm just advocating for my Dark Lord Satan.

Speaking of, when is Maturin registering, anyway? Let him advocate for his damn self.

[Scotty]

I was thinking about the nature of the post voting system being anonymous, and I wonder if it wouldn't be a good idea to keep it anonymous to everybody but the admins.
This is why: You could run a report daily that correlates a user to post voting. Thus, you wouldn't see what post(s) they voted on, but you could tell if they were voting on every single post (which would seem odd), or you could correlate it from one user to another.
So, say the report (for Scotty) would look like:

User User +votes -votes
Scotty viscousmemories 0 5
Scotty liv baby 5 0


Or some such thing, sorted by how many votes they have cast. This would let you see if somebody is just cascading through an individual and giving them all positive or negative posts, and you could just capture the top 5 or 10 individuals who have done the most votes (or the top number of votes, because that would tell you the user doing the most voting).

It would take an SQL query to get that data, and I bet it is all in the database, but it might be worth having if you are getting issues with people screwing around with the voting system (good or bad votes).

So, if somebody knows the internals of the database structure, the table names and what data is contained in it, they could write a query (I could probably write the query if I knew the data structure, but somebody else, I am sure, knows more then I do).

Just a thought.

-Scott

[livius drusus]

That is a great idea, Scotty baby. I'll let Tom speak to the technical details (including the fixed font issue), but I think it's just what the doctor ordered. :yes:

[viscousmemories]

I think your idea is excellent, Scotty, and I'll put it on the list of possible enhancements to the voting system. One question: Despite the fact that there may be inherent benefits to an anonymous voting system, your idea doesn't seem to necessitate anonymous votes. Or does it?

I'm also looking into the fixedsys problem. Have you used that on other boards? And what exactly is the problem?

[Scotty]

I think your idea is excellent, Scotty, and I'll put it on the list of possible enhancements to the voting system. One question: Despite the fact that there may be inherent benefits to an anonymous voting system, your idea doesn't seem to necessitate anonymous votes. Or does it?

I'm also looking into the fixedsys problem. Have you used that on other boards? And what exactly is the problem?

It requires that you keep track of who voted and how many times associated with (or against) an individual. So, in terms of being anonymous in what posts you voted for or against, this is true. But, it does say who you voted against because you would look at top votes by user to user (I don't know if the system keeps track of the post owner and who voted against/for the post, it might not keep that information).
But, a simplistic way of doing it would be seeing how many times a person voted in a day. Again, the database would have to keep track of that information. If it just updates a table of the post number and a vote, but not who voted or who owns the post, then this is pretty much a wash.
I hope that all makes sense.

As for the formatting, I was wrapping the little table of information in a fixed font (courier) and the fixedsys font (whatever that is) and it never changed.

-Scott

[viscousmemories]

It requires that you keep track of who voted and how many times associated with (or against) an individual. So, in terms of being anonymous in what posts you voted for or against, this is true. But, it does say who you voted against because you would look at top votes by user to user (I don't know if the system keeps track of the post owner and who voted against/for the post, it might not keep that information).
But, a simplistic way of doing it would be seeing how many times a person voted in a day. Again, the database would have to keep track of that information. If it just updates a table of the post number and a vote, but not who voted or who owns the post, then this is pretty much a wash.
I hope that all makes sense.
It does. I'm pretty sure that data is all in the db, but I'll have to verify.

As for the formatting, I was wrapping the little table of information in a fixed font (courier) and the fixedsys font (whatever that is) and it never changed.
Hmm. Odd. Lemme see:

This is courier font
This is fixedsys font
This is verdana font

I'll post again to say what I see there.

Edited to add: I see three different fonts there. Do you?

[Scotty]

Yes, it looks all good on yours, I must have just messed up a tag or something, let me try it again.

This is a test
test fixedsys test

I changed it from "Courier New" which is in the pull-down menu, to just courier and it worked, I assumed Courier New was a fixed width font, and assumed that with the fixedsys, which doesn't seem to be the case.

-Scott

[viscousmemories]

Yes, it looks all good on yours, I must have just messed up a tag or something, let me try it again.

This is a test

-Scott
How strange. The e-mail notification I got shows the words all spaced far apart, as does this quote box, but in the post it looks like a normal sentence.

[viscousmemories]

Oops. I think we crossposted while you were changing fonts, 'cause when I quoted you the tag said Courier New. So when you use courier it retains the spacing between the words? I've always wondered what fixed width fonts were all about. :?

[Scotty]

I might have edited it right underneath you. :)

-Scott

[Scotty]

Oh, and fixed width just means it uses the same spacing for all characters (spaces included) so that you can format stuff a tad easier. It is easier to read in some circumstances too.

-Scott

[viscousmemories]

Cool, thanks. :)

[Godot]

Dammit Scotty! I was going to suggest keeping track of votes. Bastard, you beat me to it.

I would think it go without saying, but if such a list were generated, it should be visible only to the admins and not the rank and file. Anonymity has many advantages, after all.

[squian]

Through the Post Voting system, everyone has a voice here, even those who choose not to speak out loud.

Does everyone deserve a voice here?

[viscousmemories]

This is why: You could run a report daily that correlates a user to post voting. Thus, you wouldn't see what post(s) they voted on, but you could tell if they were voting on every single post (which would seem odd), or you could correlate it from one user to another.
Okay. I have confirmed that sufficient post voting information is stored in the database to generate a list of who has voted, whose posts he/she has voted on, and whether the vote was positive or negative. Since we are still technically beta-testing the voting system I think it would be useful to publish the current vote tallies so we can all examine some patterns and evaluate the effectiveness of the system with some raw data.

I'll discuss this with livius later tonight or tomorrow, but I wanted to let you all know right away that this data is accessible and that I have compiled it, once again in the interest of full disclosure.

Do any of you object to this data being posted?

[viscousmemories]

Does everyone deserve a voice here?
That's a good question, Squian. Our thought was that every registered member in good standing deserves a voice here. Is there a reason you think any such member should not have a voice here?

[Godot]

Do any of you object to this data being posted? I cannot begin to tell you how bad an idea it is for this information to be shared. It is the most despicably unethical course of action you could possibly choose to take.

The data collected was collected under the pretenses of anonymity. I can see where the admins might have a need to analyse the data that is generated, but under the current setup it is still going to be full of bias. To decide to share the data with the general public after it has been collected is deplorable. If such an action were taken in a research situation, the researcher responsible would be drawn and quartered for the amusement of the Ethics Committee that approved the research in the first place. It. Just. Isn't. Done.

If my complaint is ignored and you choose to share it with everyone anyways, I insist that all personally identifiable information is removed prior to disclosure. I would still maintain that the action is despicable and unethical, but at least you'd be working on damage control.

[squian]

Our thought was that every registered member in good standing deserves a voice here. Is there a reason you think any such member should not have a voice here?

I think my original question oversimplified the issue. Perhaps the real question is, "Does every member deserve the same voice?"

For me, the notion of "one person, one vote" is a gross oversimplification of what it means to be democratic. The right to vote requires an appropriate level of maturity and responsibility. I do not feel that simple registration is sufficient to indicate these. I would rather that the right to fully contribute to the community be earned after demonstrating maturity and responsibility in posting.

Don't get me wrong, I am not opposed to lurking or lurkers. I find myself more of a lurker than a contributor. But lurkers always have the right to be heard, they just choose not to be. I would rather that people who have the passion and the time to post drive the community than see it secretly manipulated by people who add nothing.

[Godot]

I have another concern about the starting conditions of this system. As it stands, full access to the site is granted to you once you register, correct? And if you lose rep points, your access to certain areas will be removed, correct?

It has been amply demonstrated in the literature that positive reinforcement is more effective in altering behaviour than negative reinforcement. With that in mind, what rewards are there for having a higher rep if you already have the forum unlocked from Day 1? Why bother with what ends up being little more than a popularity contest?

[viscousmemories]

I cannot begin to tell you how bad an idea it is for this information to be shared.
Well as I said I haven't made any decisions in this matter and won't until I talk to livius and collect more feedback from others, but I appreciate your input.

[squian]

The data collected was collected under the pretenses of anonymity.

I could not find any explicit mention of voting anonymity. What provides the "pretenses of anonymity"?

I can see where the admins might have a need to analyse the data that is generated, but under the current setup it is still going to be full of bias.

You might have different biases in mind. However, I can see that while we are "in beta" we should be careful how we interpret the data. For example, if I am testing, I might try to see if the voting system really works by voting myself down multiple times. If the system fails to catch this exception, I should not bear the social stigma of a "bad poster".

vm and liv, are there provisions to "roll-back" from "user error"? Would you be open to fixing my rep if I damaged it while proving a bug exists? As a "beta" are we not testing the social constraints here as well as the technical ones? As such, what if I damage my reputation while making a point about the social system here?

I would propose that data created during beta testing should not be carried forward into the live, production system. At the go-live date, we should all be given a clean slate. Otherwise, we are going to treat the system with kid gloves, not the rigor necessary for good system testing.

Anyway, that's my take from a software engineering standpoint.

[Godot]

Well as I said I haven't made any decisions in this matter and won't until I talk to livius and collect more feedback from others, but I appreciate your input.
I understand that we should wait for the input of others, but that isn't going to stop me from hammering in the final nail of this here coffin before I stop for dinner.

I direct your attention to The Rules: (http://www.freethought-forum.com/forum/showpost.php?p=133&postcount=1) 3. You may not post any messages that violate the Freethought Forum Privacy Policy.

* For example:
o Member's undisclosed personal information [Emphasis mine.]
Q.E.D.

[viscousmemories]

I would rather that the right to fully contribute to the community be earned after demonstrating maturity and responsibility in posting.
How would that maturity and responsibility be quantified?

I would rather that people who have the passion and the time to post drive the community than see it secretly manipulated by people who add nothing.
Are you suggesting that only people who post should have the right to vote, or only people with a certain reputation level, or...?

[Adam]

I direct your attention to The Rules: (http://www.freethought-forum.com/forum/showpost.php?p=133&postcount=1)

3. You may not post any messages that violate the Freethought Forum Privacy Policy.

* For example:
o Member's undisclosed personal information [Emphasis mine.]
Q.E.D.

I'm not arguing with your basic premise regarding the privacy of voting records, but I don't understand how one's vote on a post constitutes undisclosed personal information. I reviewed the Privacy Policy (http://www.freethought-forum.com/forum/showthread.php?t=49), and I don't see any mention in there of one's voting record being confidential. The matter of voting confidentiality has been discussed in this thread, but it's not part of any "official" proclamation that I can find.

[Godot]

I could not find any explicit mention of voting anonymity. What provides the "pretenses of anonymity"? I saw no explicit mention of it anywhere. Yet it is implicit. The very nature of voting in Western culture carries the implication of anonymity. If you do not state otherwise up front, that is the default point of view that many of us will have. Have a look at some of the comments in this (http://www.freethought-forum.com/forum/showthread.php?goto=newpost&t=89) thread if you disagree. It more than makes my point for me.
Second, the very nature of how voting occurs in this system implies anonymity since the tallies cannot be seen right away.
Third, that it was not mentioned up front that the data would be shared is a serious brach of protocol. You don't tell people to go and play around with something and then tell them that their preconceptions would be incorrect. It is incumbent on the admins to make any such change abundantly clear before things begin. To do so during is inethical.
You might have different biases in mind.I most certainly do. I think you and I may not be on the same page with this one. I am referring to bias in the analysis of the data. Feel free to analyse it in any manner you wish to test the vigour of the software, but do not include any personally identifiable information with it. The only people that have any right to see which way I voted on anything are the admins, and quite frankly, I have a problem with them having access to the raw data as well. My concerns of bias stem from the human element involved: if you see how a particular person voted on a particular post, you have no idea why they voted that way. You only know how it was they voted. But that won't stop those with access to the raw data from inferring invalid trends or ascribing motive based on their own previous interactions with any one poster. Good intentions are one thing, but bias of this sort is plainly inevitable.
However, I can see that while we are "in beta" we should be careful how we interpret the data.Agreed. For example, if I am testing, I might try to see if the voting system really works by voting myself down multiple times. If the system fails to catch this exception, I should not bear the social stigma of a "bad poster". Of course, you can't vote for yourself, nor can you can't vote more than once on a given post, which also includes changing your vote from a (+) to a (-) (I've already tried all of these). The only way you can be voted down is by others.

As a "beta" are we not testing the social constraints here as well as the technical ones? As such, what if I damage my reputation while making a point about the social system here? That is the inherent problem with the system as it exists. Your point may be the best researched and most well thought out post ever crafted, but if it offends a number of people (or even if they just plain don't like you it will get voted down. I see no way around this at all.
I would propose that data created during beta testing should not be carried forward into the live, production system. At the go-live date, we should all be given a clean slate. Otherwise, we are going to treat the system with kid gloves, not the rigor necessary for good system testing. I would think the need to go to restore a universal tabula rasa is the most fair outcome, but I have to disagree with doing so on principle.

[Godot]

Are you suggesting that only people who post should have the right to vote, or only people with a certain reputation level, or...?
Not that I'm presuming to speak on behalf of squian in the least, but I think that my post here (http://www.freethought-forum.com/forum/showpost.php?p=439&postcount=24) might provide some food for thought on this matter (i.e. provide the opportunity to vote as an incentive of participation. Set a minimum post number/rep nunmber to unlock areas of the forum rather than remove them when your rep gets too low.).

[squian]

How would that maturity and responsibility be quantified? Are you suggesting that only people who post should have the right to vote, or only people with a certain reputation level, or...?

In conjunction with Godot's suggestion that people should work uphill to achieve a positive reputation, voting could be a feature that is granted to those with a good reputation.

Of course, that leaves the question of how much posting and reputation should be necessary. In other words, how high is the barrier to entry? Any number we pick is going to be completely arbitrary at this point. Just thinking about my own ability to contribute, I doubt I can post more than 1 good post a week. If I could do that consistently for 4 weeks, I would hope that I could earn the right to vote.

Alternatively, maybe it's better to think about what % of the community should have the right to vote. 50%? 80%? 20%?

I have no idea how hard it is to get reputation so I'm just trying to throw out some ways to constrain the problem.

[Godot]

I'm not arguing with your basic premise regarding the privacy of voting records, but I don't understand how one's vote on a post constitutes undisclosed personal information. I reviewed the Privacy Policy (http://www.freethought-forum.com/forum/showthread.php?t=49), and I don't see any mention in there of one's voting record being confidential. The matter of voting confidentiality has been discussed in this thread, but it's not part of any "official" proclamation that I can find.
Adam, I believe that I addressed this concern of yours in the subsequent post (the one in which I responded to squian, which should be two posts above this one if I get in fast enough). Just let me know if I need to clear any of that up for you.

[squian]

Of course, you can't vote for yourself, nor can you can't vote more than once on a given post, which also includes changing your vote from a (+) to a (-) (I've already tried all of these). The only way you can be voted down is by others.

This is a very important point. In fact, after I wrote my post I felt compelled to try it out as well. I daresay it also illustrates by point about the responsibility of voting. People who have not gone through this "testing" may not understand all of the nuances to voting. The Post Voting System (http://www.freethought-forum.com/forum/showthread.php?t=50) gives some indications about the indended use and says that you can only vote one time. Still, I completely missed the fact that voting is irreversible -- somehow it was just not obvious to me from the notion of one vote one time (I'm obviously dense).

Perhaps we need to rethink the voting system based on the natural inclinations. Do people have a tendancy to want to use voting as a way to agree/disagree with the post (I do)? Do people feel compelled to vote on every post? Are the thresholds appropriate for users to be able to separate the wheat from the chaff?

[Adam]

I have another concern about the starting conditions of this system. As it stands, full access to the site is granted to you once you register, correct? And if you lose rep points, your access to certain areas will be removed, correct?

It has been amply demonstrated in the literature that positive reinforcement is more effective in altering behaviour than negative reinforcement. With that in mind, what rewards are there for having a higher rep if you already have the forum unlocked from Day 1? Why bother with what ends up being little more than a popularity contest?

One idea that was mentioned earlier in the thread that I sort of liked was that achieving X reputation points would make one's votes count for more.

I was also thinking...is there any way to limit the impact that any given individual can have on any other given individual's reputation? In other words, can we fix it so that, no matter how much ass I might think Godot kicks, there's no way I can personally cause his reputation to skyrocket? Perhaps limit the number of times a given user can vote for another given user's posts? Or, better yet, allow us to vote for any number of User X's posts, but only factor a certain number of those votes in when determining reputation?

Yeah, that's sort of complicated...but I have faith in the power of code hacks...

[Godot]

Adam, if you could write an algorithm to handle that kind of differential loading, I guarantee you'd be a very, very, very, very, very, very rich man. :yup:

[Godot]

Still, I completely missed the fact that voting is irreversible -- somehow it was just not obvious to me from the notion of one vote one time (I'm obviously dense). Dense-shmense. I think the idea of being able to go back and alter your vote is a lovely one, but I don't think the software to handle it has even been written.

Perhaps we need to rethink the voting system based on the natural inclinations. Do people have a tendancy to want to use voting as a way to agree/disagree with the post (I do)?Sometimes I vote according to quality of post, and sometimes I vote according to clarity of thinking. Sometimes I vote because I feel compelled by the argument, and sometimes I vote because the post made me laugh. I'm sure that others vote based on how they feel the forum should act; as in a fluff derail in a serious thread will get some (-) votes to attempt to discourage such behaviour. Do people feel compelled to vote on every post?God, I hope not. Are the thresholds appropriate for users to be able to separate the wheat from the chaff? I don't know what the default values for rep are in this setup (vm, liv?) But I do recall that you can set your own threshold level in your profile so you don't have to bother with what you consider to be the chaff.

Another idea comes from a forum that liv recently linked me to. There, they had a forum for all the fluffy sillyness you could require and none of the posts therein counted towards your post tally. If the voting system could be worked into the equation in such a place, then we'd be able to create one level of filter to encourage participation in the higher areas of the board.

[viscousmemories]

Still, I completely missed the fact that voting is irreversible -- somehow it was just not obvious to me from the notion of one vote one time (I'm obviously dense).
To be honest I didn't know that either. It's a variable we didn't consider. Of course that's the main reason y'all are here (most by personal invitation, some by happening by): We wanted to test the mechanics and efficacy of the site in general, and particularly the voting system. We are trying to break new ground here. There is no other online discussion forum (that we've found) that has implemented an identical system.

Perhaps we need to rethink the voting system based on the natural inclinations. Do people have a tendancy to want to use voting as a way to agree/disagree with the post (I do)? Do people feel compelled to vote on every post? Are the thresholds appropriate for users to be able to separate the wheat from the chaff?
Those are some good questions. Our assumption was actually that voting would be used infrequently, reserved only for those posts that were particularly good or bad, and never used to indicate an opinion of the poster. That was our intention, anyway. Of course we considered that there would be some who would abuse the system and use it to give a thumbs-up to their friends and a thumbs-down to their enemies, but we hoped such people would be in the minority. What we did not consider is the possibility that some (many? most?) people wouldn't understand that intention and would intuitively surmise that they were expected to vote yay or nay on every post.

I'd like to discuss the data I've compiled in general terms at least, to help make my point, but again I'll refrain until livius and I have a chance to meet and we discuss this further here. For the record I am definitely leaning toward agreement with Godot that there was an implicit assurance of anonymity with the voting. But I wanted to announce the fact that I have access to the data as early as possible.

Whether we administrators should access that data at all is another good question. People who are not technically savvy may not realize exactly what data we as administrators have access to and will access. In fact I will post an official clarification of that. For example, we have access to the members' private messages and passwords. But while I agree that it would be unethical to access such personal information, I disagree that the same ethical consideration applies to post votes. Simply because we did specify that the system was new and untested, and that our intention - before we begin announcing our site to the general public - was to use this "beta group" of members to help us experiment and perfect the voting system.

[viscousmemories]

I want to reply to every point y'all have made, but first I want to give you some data about our system as it is configured. So here's those stats:

Reputation and post voting: The default reputation a member starts with is a variable we control. It is currently set at 25.
The reputation levels are a variable we can control. We currently have configured the following three levels:
Less than 0 points - Bad reputation
0 points - Could go either way
Greater than 0 points - A member in good standing
The percentage of a reputation point each post vote counts as is a variable we control. As configured, 1 post vote = 1/100 of a reputation point. Hence, it would currently take 2500 negative votes, and 0 positive votes, for a member to get a bad reputation. (I hadn't really done the math on this one until just now. The idea was that we want to make it hard to get a bad reputation, but methinks it's flawed. With this configuration it's damn near impossible.)

Other points to consider:

First, the software is flexible. This voting system was not built in to vBulletin nor was it an existing hack. We paid someone to write it for us. So if we find it needs modification, that avenue is available.

Second, the reputation system is secondary to our intention with the post voting system. Our idea was that we want the community to decide collaboratively what is and isn't "quality" content, and label it accordingly. Thusly anyone who generally agrees with the community opinion can set a filter to allow only that content which bears the community stamp, so to speak, and anyone who prefers free and unencumbered speech could view everything available.

The reason we had the software designed so that post votes impact member reputation is because we thought that if the community overall disapproved of the majority of the content from a particular member, we would discourage the member from hanging out here by making it a less desirable place to be. To that end we decided to make it so that users with a bad reputation would lose access to the chat, arcade, and journals, and we considered revoking other privileges like custom user title, avatar, signature, voting rights, etc.

[Adam]

Adam, I believe that I addressed this concern of yours in the subsequent post (the one in which I responded to squian, which should be two posts above this one if I get in fast enough). Just let me know if I need to clear any of that up for you.

Yeah, I think I see where you're comign from now...even if you did use my "cold, dead fingers" post against me. Bastard. :)

[Adam]

Adam, if you could write an algorithm to handle that kind of differential loading, I guarantee you'd be a very, very, very, very, very, very rich man. :yup:

It shouldn't be that hard with the right database. Unfortunately for my prospects for wealth, however, I have no idea what the database looks like.

[livius drusus]

I apologize for my delay in posting to this thread; I should have known better than to attempt to have a life 2 days after opening. I will try to address some of the major issues I see here, but will have to do it in parts, I'm afraid, as my bedtime soon approaches. I'll do my utmost to reply promptly.

Regarding the matter of voting data access and publication, now that we know that the admins do indeed have access to who has voted what post up or down, I agree with Godot that we should respect the tradition of anonymous voting and not publish detailed information. vm and I have had a quick discussion about it (no time for more; so much to write), and we are in agreement that any voting data we collect from the database should be discussed, if at all, in terms of broad pattern analysis. I do believe some analysis would be handy at this juncture in order to address some of the other issues squian, Godot and Adam have raised about the way voting works in practice.

I disagree, however, that the Rule 3 - which prohibits violations of the current privacy policy - would prohibit disclosures of voting data. As Adam pointed out, the privacy policy has nothing at all to say about voting data, mainly because we had no idea if it even existed until Scotty's post on the subject. If anything, the paragraphs on aggregate data or IP information are more apt as they refer to information collected by the software based on usage as opposed to identifying information about the person behind the handle (which is what the example Godot quoted from Rule 3 refers to).

Still, even those paragraphs don't apply, because voting data is quite different from, say, browser information. This lacuna will be rectified shortly. Thank you for pointing it out.

As to whether the admins should have access to the raw data, I'll have to ask you to clarify your objection, Godot. In your first post (http://www.freethought-forum.com/forum/showthread.php?p=275&postcount=18) on this thread you stated:

Dammit Scotty! I was going to suggest keeping track of votes. Bastard, you beat me to it.

I would think it go without saying, but if such a list were generated, it should be visible only to the admins and not the rank and file. Anonymity has many advantages, after all.

Later (http://www.freethought-forum.com/forum/showthread.php?p=436&postcount=22) you seemed to suggest that the proposed data collection and analysis is an intrinsic violation of voter anonymity. If I misread you, I'm sure you'll correctly me promptly. It's rather late and this is a very information-rich thread.

Anyway, when I wrote the privacy policy I tried to cover every base I could think of to make clear that nothing is entirely private (hence the paragraphs about PMs on the forum and in the Chatroom), because the database in its entirety is available to the administrators. Since we didn't know the data existed until this discussion, we shall now endeavor to make it clear that voting is not, in fact, anonymous to the administrators, although it is and will remain anonymous to everyone else.

But that won't stop those with access to the raw data from inferring invalid trends or ascribing motive based on their own previous interactions with any one poster. Good intentions are one thing, but bias of this sort is plainly inevitable.

Bias is human of course, but as long as no biased conclusions are posted and no administrative action is taken as a result of it, I'm not sure I see what problem our bias could cause. If your sole concern regarding administrator access to the raw data is that we will see maliciousness where there is none or see conspiracies to play the system where there is only a sincere desire to be an active participant in the community, then again, unless we talk shit about people with third parties, or seek to use administrative privileges against the members our biases delude us into believing are out to dick with us, the problem seems to me to be one of character, and not a forum issue at all.

So, I suppose the next question is what if any data we should publish and to what ends. Who voted will not be posted. This is a firm position. How about which posts were voted on? Should the anonymity of a vote extend to listings of which posts are being rated up or down?

In some cases, it's perfectly clear which posts are being rated up or down simply by looking on the bottom left of the postbit - the four posts of mine that are currently at 24 points, for instance. It may not be public knowledge how many up and downs they received, but of course they did receive at least one down vote or they would not be rated under 25. Besides, If we use the Western tradition analogy Godot rightly suggested as reason to keep the ballot secret, then the results of the ballot are not only publishable, but publishing them is indeed the entire point.

I tend to think the analogy falters on that point, because of course, we aren't electing anyone; there is no race with winners and runoffs in this voting system. I could see how such data might be useful in determining the way voting is actually used here and therefore helpful when discussing how best to design the system to achieve our goals (see the conversation between squian and Godot re the way voting actually works as opposed to how we might wish it to work), but I'm not really sure how much the very limited information we have right now with 30 users on the board can be said to mean anything.

Moving outwards from there, would publishing general numbers (total votes up and down per day, for instance) be of use to the discussion? If so, in what way? If not, then why publish anything at all? If we don't publish anything, and our seeing the data is so fraught with bias as to render any conclusions drawn hopelessly corrupt, then why run the report in the first place?

I'd particularly like to hear from Scotty on this issue and Godot who was going to make Scotty's suggestion before the bastard beat him to it. What did y'all have in mind we do with the information when you recommended we collect voting data?

Okay. Part 1 finished. Off to start part 2, but I don't think I'll be able to finish it tonight. Thank you all for your input. This discussion is so very necessary and I really appreciate your participation.

[Godot]

I apologize for my delay in posting to this thread; I should have known better than to attempt to have a life 2 days after opening. I will try to address some of the major issues I see here, but will have to do it in parts, I'm afraid, as my bedtime soon approaches. I'll do my utmost to reply promptly. Having a real life is sorely over-rated. At least, that's what I tell people.

Regarding the matter of voting data access and publication, now that we know that the admins do indeed have access to who has voted what post up or down, I agree with Godot that we should respect the tradition of anonymous voting and not publish detailed information. vm and I have had a quick discussion about it (no time for more; so much to write), and we are in agreement that any voting data we collect from the database should be discussed, if at all, in terms of broad pattern analysis.This is doubleplusgood stuff here. I'm even more glad because I rather understated my objection in my first post on this thread.
Now I'm going to a prick and ask you what sort of analysis you will be performing. What do you mean by "broad pattern analysis"? What techniques will you use to analyse your data? Why those techniques? Are there better methods available to you? If so, what are they?
I think I'll stop there. I do believe some analysis would be handy at this juncture in order to address some of the other issues squian, Godot and Adam have raised about the way voting works in practice. But will the data that you are capable of collecting have the ability to do so? Or will it be far too broad? And specifically, you need to put in writing what questions you wish to have answered and how exactly you intend to massage the needed information out of your data in order to answer the questions you wish to ask. And you have to do all of this before you begin the analysis. Hell, to be done right, these questions should have been asked long before the data was being collected.

I disagree, however, that the Rule 3 - which prohibits violations of the current privacy policy - would prohibit disclosures of voting data. As Adam pointed out, the privacy policy has nothing at all to say about voting data, mainly because we had no idea if it even existed until Scotty's post on the subject.If I'm reading this correctly, you're saying that publishing the personally identifiable data that you can collect about the voting patterns of your members does not contravene your privacy rules simply because they aren't explicitly mentioned in the rules as written? That you were ignorant of being able to collect this data does not negate your ethical responsibility to hide this data from public view. "We didn't know it existed" is hardly a valid defense of any sort. Now that it has been pointed out to you, you are obligated to amend the rules to take this into consideration. If you choose not to amend the rules, you are still bound by ethical considerations not to share this information. Of course, since you've already stated that you will not be sharing this information, my continued bleating is something of a moot point. But as I said above, I rather understated my first post in this thread and am making up for it here. You have no idea just how vehemently opposed I am to having this data shared.
If anything, the paragraphs on aggregate data or IP information are more apt as they refer to information collected by the software based on usage as opposed to identifying information about the person behind the handle (which is what the example Godot quoted from Rule 3 refers to). I really have no idea to what you are referring here. My objection stems from the fact that the voting data contains personally identifiable information about the users of this board. Since you did not state up front before people started voting that you had the ability to publish the heretofore private information, you still have a personal ethical obligation to not share it.
Still, even those paragraphs don't apply, because voting data is quite different from, say, browser information. This lacuna will be rectified shortly. Thank you for pointing it out. I repeat, much of my objection shall be made moot shortly (I hope).

As to whether the admins should have access to the raw data, I'll have to ask you to clarify your objection, Godot. In your first post (http://www.freethought-forum.com/forum/showthread.php?p=275&postcount=18) on this thread you stated:



Later (http://www.freethought-forum.com/forum/showthread.php?p=436&postcount=22) you seemed to suggest that the proposed data collection and analysis is an intrinsic violation of voter anonymity. If I misread you, I'm sure you'll correctly me promptly. It's rather late and this is a very information-rich thread. There is a distinction to be made between access to raw data and clean data. The term raw data as I'm using it here refers to the untouched log files that show each and every aspect of what was copied. The clean data would be much the same apart from having the personally identifiable aspect removed. The easiest way to do so is to create a coding system known only to the admins whereby every voter name is translated into an alphanumeric code. That way, the admins can still keep track of what they need to keep track of without having to disclose any sensitive information to the rest of us.

There is also a mix-up between two concepts in those posts, thus leading to your confusion. I still maintain that the only people that should have access to this information (in full) should be the admins. But even then I have issues with it regrading sources of bias (which I will address further along in this post).
The second issue is that you do have a problem in that you have begun collecting data without informing the people that it was possible for you to do so. Is it unfair of me to harp on this since you didn't know about it until Scotty mentioned it? Yep, it really is. But that isn't going to stop me from driving the point home. You now have data available to you that you had no clue existed and it can help you sort out a number of problems with the system. But since it was collected unawares, you cannot ethically use it anyways. Neither you nor the subjects for your experiment knew it existed, and now that you've collected it, you have an ethical obligation to not use it. Of course, you can rectify this problem once you go live and we finish with the beta-testing.
Since I know this will have only hopelessly confused matters further, I'm going to stop at that point to field questions and/or insults.
Anyway, when I wrote the privacy policy I tried to cover every base I could think of to make clear that nothing is entirely private (hence the paragraphs about PMs on the forum and in the Chatroom), because the database in its entirety is available to the administrators. Since we didn't know the data existed until this discussion, we shall now endeavor to make it clear that voting is not, in fact, anonymous to the administrators, although it is and will remain anonymous to everyone else. Gee, it's starting to sound like a broken record in here...

Bias is human of course, but as long as no biased conclusions are posted and no administrative action is taken as a result of it, I'm not sure I see what problem our bias could cause. Stop it right there.
How do you know that no biased conclusions will be drawn?
How do you know that no biased administrative action will be taken?

The answer: you don't.

Each and every decision you make is imbued with some sort of bias or another. Period. To assume that your administrative decisions will be free from bias is laudable but naive. You're going to fuck up as much as the rest of us are, hopefully less so since there are two of you to act as a system of checks and balances for the other. But it's still going to happen.
Look at it from a data analysis point of view:
Say you have Person X who always votes (+) for a certain type of post, no matter who posts it and who votes (-) for every post written by a specific user. Now, you can draw sertain conclusions about this user based on this data and you can use this information that you've gathered to tinker with the voting system. However, if you knew when performing the analysis that Person X was in fact John Doe, then you would know something about John Doe that nobody else knows. And you would alter your opinion of this person, not to mention how you interact with them, as a result of knowing said information. While this knowledge may be of some use to you as an administrator, as a poster on the boards it is dangerous.
If your sole concern regarding administrator access to the raw data is that we will see maliciousness where there is none or see conspiracies to play the system where there is only a sincere desire to be an active participant in the community, then again, unless we talk shit about people with third parties, or seek to use administrative privileges against the members our biases delude us into believing are out to dick with us, the problem seems to me to be one of character, and not a forum issue at all. I am stunned that you seem to be able to reduce my concerns down to one singular point. I certainly hope I have expanded on my concerns in sufficient detail to allay this concern of yours. Nonetheless, I must once again insist that the entirety of my objection does not rest on issues of potential impropriety, but it does play a part. And yes, it most certainly is a forum issue simply by virtue of your posting on it as a regular user as well as in your administrative role. If you fail to see that, then this conversation will come to a rathe abrupt end. Do you recall the issues that some people had against certain aspects of II regarding the appearance of impropriety? Well now it's your turn to feel the heat. The exact same criticism applies to you, and it is equally invalid when applied to you too. But it is there nonetheless.
So, I suppose the next question is what if any data we should publish and to what ends. Who voted will not be posted. This is a firm position. How about which posts were voted on? Should the anonymity of a vote extend to listings of which posts are being rated up or down? Good questions these are, but the answers are readily apparent. We alrady have this data provided to us at the bottom of each post in each thread. Do you intend to hide this data? If so, why?
I do not think that having either of these bits of information visible contravenes any ehtical considerations in the least. Having a visible posting of a vote tally serves as an excellent community standard of conduct. if a post is universally decried as poor, then it should be made visible to lurkers, guests, and memebrs just what sort of cummunity standard is set forth.How else can we self-moderate our response if we don't have a baseline for comparison?
In some cases, it's perfectly clear which posts are being rated up or down simply by looking on the bottom left of the postbit - the four posts of mine that are currently at 24 points, for instance. It may not be public knowledge how many up and downs they received, but of course they did receive at least one down vote or they would not be rated under 25. I see. Are you suggesting that the entire voting ecord be made visible rather than just the aggregate score? Again, I have no ethical issue with doing so, but it would make for a cumbersome report. Perhaps just include the summary of the number of (+) votes, the number of (-) votes and the balance as a result. Same end product, much nicer appearance.Showing the raw data log would serve no purpose whatsoever. Besides, If we use the Western tradition analogy Godot rightly suggested as reason to keep the ballot secret, then the results of the ballot are not only publishable, but publishing them is indeed the entire point. You miss the point of the analogy. I had been discussing the the principle of voting and how it related to anonimity. You're perfectly free to extend that analogy to any point you see fit, but it certainly will not be arguing against me.
I tend to think the analogy falters on that point, because of course, we aren't electing anyone; there is no race with winners and runoffs in this voting system. Good thing I never made that argument. Thus, the anaolgy still stands as valid. I could see how such data might be useful in determining the way voting is actually used here and therefore helpful when discussing how best to design the system to achieve our goals (see the conversation between squian and Godot re the way voting actually works as opposed to how we might wish it to work), but I'm not really sure how much the very limited information we have right now with 30 users on the board can be said to mean anything. I can confidently state right here, right now, that any data you collect from a sample of approximately 30 people in the very limited time you have allotted for beta-testing will be abso-fucking-lutely useless. The only conclusion that you will be able to draw is that the voting system is free of bugs. The absolute paucity of data that you will have will mean that you can do nothing apart from tally up a meagre frequency chart, cross-referenced by poster, forum, and time of day. i'd even suggest via gender, but this is the internet after all. Collect the data for at least a week or even a month if you want the chance to see something useful (of course, this doesn't even get into the issues surrounding validity or accuracy, but hey, I've probably pissed offenough people with this post to bother with that level of detail).
Moving outwards from there, would publishing general numbers (total votes up and down per day, for instance) be of use to the discussion? If so, in what way? If not, then why publish anything at all? What is it you hope to gain by even collecting this data? I have some ideas, but I am going to postpone them until I hear what others have to say. If we don't publish anything, and our seeing the data is so fraught with bias as to render any conclusions drawn hopelessly corrupt, then why run the report in the first place? I do hope this was in jest more than anything else. The data you collect can be valuable for a number of reasons and used in a variety of ways, most of which centre around ensuring that the voting system is as sensitive as you wish for it to be.
I'd particularly like to hear from Scotty on this issue and Godot who was going to make Scotty's suggestion before the bastard beat him to it. What did y'all have in mind we do with the information when you recommended we collect voting data? Like I said, I'm going to hold off until after the impending lynch mob pays me a visit for the vaster part of this post above. I want to know what you and the others have in mind before I have my say.
Okay. Part 1 finished. Off to start part 2, but I don't think I'll be able to finish it tonight. Thank you all for your input. This discussion is so very necessary and I really appreciate your participation. I still think you're going to be very pissed at me for this. ;)

[viscousmemories]

Thanks for the comments, Godot. It's past my bedtime so this is in no way meant to be a full response to your post, but I do want to make one point: The voting data that we've collected thus far is not useless. We know that of 30 people X number voted X times on X number of posts by X user, etc.

We may not be able to come to any significant conclusions about the voting system based on such trivial data, but we have still learned some things. We know that at least a certain percentage of our members found and used the buttons, for example. We know that some people used them more than others, etc. So while the data might not answer any important questions, it has at least raised some and illuminated the inadequacy of our answer to others.

I'm sure livius will want to reply to your response in much more detail tomorrow. I just ask that you be patient and realize that we are working through these issues and responding as quickly and thoroughly as we can.

[Hugo Holbling]

In analysing the data, i think there should be less concern at Type II errors and more for Type I. I would like to know what kind of demarcation criteria will be used to avoid confusing the two.

[Godot]

Thanks for the comments, Godot. It's past my bedtime so this is in no way meant to be a full response to your post, but I do want to make one point: The voting data that we've collected thus far is not useless. We know that of 30 people X number voted X times on X number of posts by X user, etc. But these are simply statements of fact and hardly rely on the nature of the data itself. At this crude a level, we are certain that the tool is functioning correctly. And that's it.

After a moment's thought on the matter, I realised that you've just said that you've already had a look at the data. Not only that, that you've done some rather crude ad-hoc analysis of the data. Did you happen to miss the part of my massive post where I quite clearly said that you do not have the right to do so? Ethically, you do not have the ability to analyse this data! I cannot emphasise this point severely enough!

I also see that you now have put a banner on the front page warning people that we're still in beta-testing mode in case theyt hadn't seen it before. I sincerely applaud the effort, but it really only applies to new registrees and all posts/votes made after the time that the banner was put in place. It is an ex post facto rule, and while it certainly is a step in the right direction, it feels like closing the barn door after the cows have escaped. Because of the nature of this rule, you still cannot use the data collected prior to this disclosure.

We may not be able to come to any significant conclusions about the voting system based on such trivial data, but we have still learned some things. We know that at least a certain percentage of our members found and used the buttons, for example. We know that some people used them more than others, etc. So while the data might not answer any important questions, it has at least raised some and illuminated the inadequacy of our answer to others. You really need to be careful about using the word significant. When we're having a discussion involving numbers and statistics, significance takes on a whole new world of meaning. Using the word improperly can get you severly castigated in the wrong company. I'll let you off with a warning this time. ;)
But as Hugo mentions, this really only increases the Type I error rate in the analysis of your data. Since I know most people around here don't have a bloody clue what the hell a Type I or a Type II error even is, click here (http://www.sportsci.org/resource/stats/errors.html) for a brilliantly written description.

I'm sure livius will want to reply to your response in much more detail tomorrow. I just ask that you be patient and realize that we are working through these issues and responding as quickly and thoroughly as we can.I understand totally, and I'm sure liv will have a go at me in the morning. Other than that, get yourself to bed! You have a big day tommorrow of us beating the shit out of you and making unreasonable demands of you.

[livius drusus]

Now I'm going to a prick and ask you what sort of analysis you will be performing. What do you mean by "broad pattern analysis"?

What kinds of posts get voted up or down, mainly, but please keep in mind that I have not seen this data which Tom only confirmed was accessible 2 minutes before I dragged my sorry ass off to see "I, Robot" and make a killer dinner of broiled marinated Salmon with caramelized onions and baked sweet potato fries.

What techniques will you use to analyse your data?

What techniques are available?

Why those techniques? Are there better methods available to you? If so, what are they?

Good questions. Have any answers?

I think I'll stop there. But will the data that you are capable of collecting have the ability to do so? Or will it be far too broad?

I think Tom has addressed that in his reply (at least in part).

And specifically, you need to put in writing what questions you wish to have answered and how exactly you intend to massage the needed information out of your data in order to answer the questions you wish to ask. And you have to do all of this before you begin the analysis. Hell, to be done right, these questions should have been asked long before the data was being collected.

That makes sense. What kinds of questions did you have in mind when you suggested we collect such data?

If I'm reading this correctly, you're saying that publishing the personally identifiable data that you can collect about the voting patterns of your members does not contravene your privacy rules simply because they aren't explicitly mentioned in the rules as written?

Personally identifiable as currently written in our privacy policy refers to rl information about the user, so yes, from my perspective voting data is not in that same category. It seems to me like it's very much in its own category.

That you were ignorant of being able to collect this data does not negate your ethical responsibility to hide this data from public view.

Indeed.

"We didn't know it existed" is hardly a valid defense of any sort. Now that it has been pointed out to you, you are obligated to amend the rules to take this into consideration.

Indeed.

If you choose not to amend the rules, you are still bound by ethical considerations not to share this information. Of course, since you've already stated that you will not be sharing this information, my continued bleating is something of a moot point.

Indeed.

But as I said above, I rather understated my first post in this thread and am making up for it here. You have no idea just how vehemently opposed I am to having this data shared.

I think I may have an inkling now. In any case, I quite share your vehemence on this score.

Since you did not state up front before people started voting that you had the ability to publish the heretofore private information, you still have a personal ethical obligation to not share it.
I repeat, much of my objection shall be made moot shortly (I hope).

It shall.

There is a distinction to be made between access to raw data and clean data. The term raw data as I'm using it here refers to the untouched log files that show each and every aspect of what was copied. The clean data would be much the same apart from having the personally identifiable aspect removed. The easiest way to do so is to create a coding system known only to the admins whereby every voter name is translated into an alphanumeric code. That way, the admins can still keep track of what they need to keep track of without having to disclose any sensitive information to the rest of us.

Oh, I see. That's an excellent idea and will see about putting it into practice as soon as the person who actually knows how wakes up. If Tom doesn't know how to set it up, I'm sure Brian will be able to help.

The second issue is that you do have a problem in that you have begun collecting data without informing the people that it was possible for you to do so. <snip>
But since it was collected unawares, you cannot ethically use it anyways. Neither you nor the subjects for your experiment knew it existed, and now that you've collected it, you have an ethical obligation to not use it.

Makes sense to me.

Of course, you can rectify this problem once you go live and we finish with the beta-testing.

Yessir.

Stop it right there.
How do you know that no biased conclusions will be drawn?

I didn't say no biased conclusions will be drawn, but rather that they would not be published or shared with third parties.

How do you know that no biased administrative action will be taken?

The answer: you don't.

Well, we have pretty strict rules regarding what sort of things require explicit administrative intervention, but if you count our participation on the board as administrative action, you're right.

Look at it from a data analysis point of view:
Say you have Person X who always votes (+) for a certain type of post, no matter who posts it and who votes (-) for every post written by a specific user. Now, you can draw sertain conclusions about this user based on this data and you can use this information that you've gathered to tinker with the voting system. However, if you knew when performing the analysis that Person X was in fact John Doe, then you would know something about John Doe that nobody else knows. And you would alter your opinion of this person, not to mention how you interact with them, as a result of knowing said information. While this knowledge may be of some use to you as an administrator, as a poster on the boards it is dangerous.

True that. Particularly now when there are so few people.

I am stunned that you seem to be able to reduce my concerns down to one singular point.

I said "if" because that was the issue I envisioned when I read your post last night. I had no doubt you'd flesh out the nature of your objection more fully in reply and so you have.

I certainly hope I have expanded on my concerns in sufficient detail to allay this concern of yours.

I think so.

Do you recall the issues that some people had against certain aspects of II regarding the appearance of impropriety? Well now it's your turn to feel the heat. The exact same criticism applies to you, and it is equally invalid when applied to you too. But it is there nonetheless.

It's invalid? I think it's quite valid. In fact, I wish you had raised it back when you first suggested we collect the data.

Good questions these are, but the answers are readily apparent. We alrady have this data provided to us at the bottom of each post in each thread. Do you intend to hide this data? If so, why?

You don't have the detailed data provided at the bottom of each post; you have a final tally, and no that will not be hidden.

I do not think that having either of these bits of information visible contravenes any ehtical considerations in the least. Having a visible posting of a vote tally serves as an excellent community standard of conduct. if a post is universally decried as poor, then it should be made visible to lurkers, guests, and memebrs just what sort of cummunity standard is set forth.How else can we self-moderate our response if we don't have a baseline for comparison?

Yup.

I see. Are you suggesting that the entire voting ecord be made visible rather than just the aggregate score? Again, I have no ethical issue with doing so, but it would make for a cumbersome report. Perhaps just include the summary of the number of (+) votes, the number of (-) votes and the balance as a result.

That looks very good to me.

You miss the point of the analogy.

No, actually, just extending it to help me think things through.

I had been discussing the the principle of voting and how it related to anonimity. You're perfectly free to extend that analogy to any point you see fit, but it certainly will not be arguing against me.

Correct.

Good thing I never made that argument. Thus, the anaolgy still stands as valid.

Good thing I never said it was invalid, then. Quite the opposite, in fact, since I have agreed (like a broken record, no less) with your original point about anonymous voting.

I can confidently state right here, right now, that any data you collect from a sample of approximately 30 people in the very limited time you have allotted for beta-testing will be abso-fucking-lutely useless. The only conclusion that you will be able to draw is that the voting system is free of bugs.

Fair enough. The data might give us some ideas when addressing the issues lisarea first raised and you, squian and Adam also pursued regarding the structure of the post voting system; so in that sense it is not absolutely useless even if conclusions about the data itself are statistically impossible.

Collect the data for at least a week or even a month if you want the chance to see something useful

Okay.

(of course, this doesn't even get into the issues surrounding validity or accuracy, but hey, I've probably pissed offenough people with this post to bother with that level of detail).

You've already bothered above, I think.

What is it you hope to gain by even collecting this data? I have some ideas, but I am going to postpone them until I hear what others have to say.

I believe Scotty originally suggested collecting data as a reply to lisarea's concern about one person systematically voting down another to dick with their reputation. My hopes for the data at this point is that it will illustrate the flaws in the current bare-bones setup of our voting system so that we can come up with as fair a system as possible before we go live.

Still, I have virtually no knowledge of statistics or the possibilities. That's why we need input from people who know about things like raw data v. clean. People much like yourself, for instance.

Like I said, I'm going to hold off until after the impending lynch mob pays me a visit for the vaster part of this post above. I want to know what you and the others have in mind before I have my say.

I hope I didn't disappoint you what with all my agreeing. Doesn't look the lynch mob has had its coffee and donuts yet.

I still think you're going to be very pissed at me for this. ;)

Anything that keeps you away from Tetris Arcade is just raindrops on roses and whiskers on kittens for me. :)

Thank you, Godot. Your perspective on these issues has been illuminating. I hope you won't wait too long for that lynch mob before you expand on your thoughts.

[viscousmemories]

But these are simply statements of fact and hardly rely on the nature of the data itself. At this crude a level, we are certain that the tool is functioning correctly. And that's it.

After a moment's thought on the matter, I realised that you've just said that you've already had a look at the data. Not only that, that you've done some rather crude ad-hoc analysis of the data. Did you happen to miss the part of my massive post where I quite clearly said that you do not have the right to do so? Ethically, you do not have the ability to analyse this data! I cannot emphasise this point severely enough!
Nope, I didn't miss it. Unfortunately it came after I had already compiled and examined the data. As you might imagine, the simple act of ensuring that the data gathering method was effective on such a small sample made me able to get a broad view of things with hardly any effort. What can I do about it now, short of not analyzing or sharing the analysis of the data on a going forward basis until we resolve how best to manage this?

I also see that you now have put a banner on the front page warning people that we're still in beta-testing mode in case theyt hadn't seen it before. I sincerely applaud the effort, but it really only applies to new registrees and all posts/votes made after the time that the banner was put in place. It is an ex post facto rule, and while it certainly is a step in the right direction, it feels like closing the barn door after the cows have escaped. Because of the nature of this rule, you still cannot use the data collected prior to this disclosure.
I put that banner up not to justify using the data - as I agree that it would be insufficient - but to notify the couple of people who joined without having received an invitation of our intentions during this time.

You really need to be careful about using the word significant. When we're having a discussion involving numbers and statistics, significance takes on a whole new world of meaning. Using the word improperly can get you severly castigated in the wrong company. I'll let you off with a warning this time. ;)
My apologies. As I'm sure you've deduced I am not well versed in statistics.

But as Hugo mentions, this really only increases the Type I error rate in the analysis of your data. Since I know most people around here don't have a bloody clue what the hell a Type I or a Type II error even is, click here (http://www.sportsci.org/resource/stats/errors.html) for a brilliantly written description.
Thank you for making an effort to explain your criticism. I'll have a look at that link and see if it helps me understand where you're coming from.

[livius drusus]

In conjunction with Godot's suggestion that people should work uphill to achieve a positive reputation, voting could be a feature that is granted to those with a good reputation.

We have considered setting up a promotions system which would delegate voting power to posters based on reputation. We haven't implemented it at this stage because we wanted to strip the system down to its essentials in order to explore all the possibilities, but I think it's pretty clear that the one man-one vote on registration default is far from ideal.

Of course, that leaves the question of how much posting and reputation should be necessary. In other words, how high is the barrier to entry? Any number we pick is going to be completely arbitrary at this point. Just thinking about my own ability to contribute, I doubt I can post more than 1 good post a week. If I could do that consistently for 4 weeks, I would hope that I could earn the right to vote.

When you say "good post", do you mean one that has been voted up or not voted down or by some non-voting standard? As the software is currently configured, we can create a promotions system based on reputation level, days since registration, number of posts, or any combination thereof.

[livius drusus]

I was also thinking...is there any way to limit the impact that any given individual can have on any other given individual's reputation? In other words, can we fix it so that, no matter how much ass I might think Godot kicks, there's no way I can personally cause his reputation to skyrocket? Perhaps limit the number of times a given user can vote for another given user's posts?

I like this idea a lot. The default vB reputation system allows such limits, but the way the vote system is currently configured does not. I'll ask Brian if we can add such a feature.

Or, better yet, allow us to vote for any number of User X's posts, but only factor a certain number of those votes in when determining reputation?

Hmmm... I have no idea how that would work, exactly. Were you thinking something like a ceiling beyond which the votes would not be factored into reputation until x amount of time had passed?

Lawds this is complex. Thank you for helping unravel all these details.

[livius drusus]

Another idea comes from a forum that liv recently linked me to. There, they had a forum for all the fluffy sillyness you could require and none of the posts therein counted towards your post tally. If the voting system could be worked into the equation in such a place, then we'd be able to create one level of filter to encourage participation in the higher areas of the board.

Worked into the equation in what way? Were you thinking of some kind of vote weighting system, perhaps? I'm not sure there are higher areas on this board, truth be told; more like areas of interest.

[viscousmemories]

Here's a possible alternative to the current voting system that just came to mind:

What if we were to limit the number of votes (up and down) each person is allowed, and increase that number based on the number of up-votes he/she has received, and reduce the amount by the number of down-votes. The benefit of this system would be that only people whose posts are appreciated by a substantial percentage of the community participants would have any significant (if that's not a misuse of the word in this context) ability to rate the contribution of others.

This way even a troll with a couple of friends constantly supplying him with up-votes would still find his ability to impact the community disabled by the inevitable down-votes he would get from many others. And by extension, if his friends were spending all their votes trying to boost the troll up, they would effectively lose their ability to use those votes to impact the community in any other way. And of course if the troll wanted to return the favor by up-voting the friends who were feeding him, he would have to do so at the cost of votes he would otherwise be using to degrade someone else.

In other words, person A votes up a post of person B. Now person B has one vote to give. He can choose to use that vote to vote someone else down, or to vote person A up, but not both. So if person A is a troll and person B is a troll-friend, the troll-friend can only help the troll or hurt someone else. But he can't do both. And the only way he can get another vote is for someone to vote him up.

Of course everyone would have to start with a certain number of votes, or no voting could take place to begin with. We could possibly base the number of starting votes on length of registration or some other factor, or even have a set number per hour, day, postcount, etc.

[livius drusus]

I have amended the Privacy Policy to include the following:

Post Voting Data
Post voting is anonymous to the general membership. The current tally on any given post is viewable on the bottom left of the post, but information on which member voted which way for what post is not. The information is stored in the database, however, and is therefore accessible to the Administrators. We are still in the process of determining how and if we will examine this data and for what purpose.

Will that work for now?

[Hugo Holbling]

Is there a reason why other concerns have been addressed but not mine?

[livius drusus]

Is there a reason why other concerns have been addressed but not mine?

My reason is that I thought Tom's reply to the post Godot made expanding on your question was a response to your concerns as well. However, I'll be glad to answer your concerns more fully. Please let me know if there's anything I've overlooked.

In analysing the data, i think there should be less concern at Type II errors and more for Type I. I would like to know what kind of demarcation criteria will be used to avoid confusing the two.

I am not familiar with statistical errors of any type. I have read the link Godot kindly provided by way of explanation, but unfortunately found it very difficult to comprehend.

At this point, I think it best to place the issue of data analysis firmly on the backburner. We know from yesterday's query that it can be collected if necessary. Beyond that, it remains to be seen if we have any reason to collect it at all, nevermind analyze it.

Should a compelling reason arise, then I think we will have to seriously consider hiring professional statisticians to overcome the difficulties you and Godot have underscored. One thing is clear: neither of the current admins are qualified to do statistical analysis.

Thank you for your participation in this discussion.

[Hugo Holbling]

Godot and i have both studied statistics and modelling methodologies at graduate level. I hardly think there is any need for professional help.

[Godot]

Worked into the equation in what way? Were you thinking of some kind of vote weighting system, perhaps? I'm not sure there are higher areas on this board, truth be told; more like areas of interest.
Well, I for one feel that the introductory threads have a lesser value on the boards compared to a more higher level of discussion and debate. But then again, that's likely just me.

It laregely would depend on what sort of environment you wish to foster on the board. If you're aiming to have a fair bit of serious discussion, you might prefer that posts of that nature are in a different category than the fluff ones. To that end, perhaps removing voting criteria from a particular set of posts (I mean, how the hell can you vote (+) or (-) for a hello post?) so they do not even factor into the decision.

[livius drusus]

Godot and i have both studied statistics and modelling methodologies at graduate level. I hardly think there is any need for professional help.
Thank you for your offer to help us make our way through the statistical minefield, but as I said above, I think it best to address the issue of how to properly analyze voting data when we have determined that there is compelling reason to collect it in the first place.

[Godot]

Here's a possible alternative to the current voting system that just came to mind:

What if we were to limit the number of votes (up and down) each person is allowed, and increase that number based on the number of up-votes he/she has received, and reduce the amount by the number of down-votes. The benefit of this system would be that only people whose posts are appreciated by a substantial percentage of the community participants would have any significant (if that's not a misuse of the word in this context) ability to rate the contribution of others.

This way even a troll with a couple of friends constantly supplying him with up-votes would still find his ability to impact the community disabled by the inevitable down-votes he would get from many others. And by extension, if his friends were spending all their votes trying to boost the troll up, they would effectively lose their ability to use those votes to impact the community in any other way. And of course if the troll wanted to return the favor by up-voting the friends who were feeding him, he would have to do so at the cost of votes he would otherwise be using to degrade someone else.

In other words, person A votes up a post of person B. Now person B has one vote to give. He can choose to use that vote to vote someone else down, or to vote person A up, but not both. So if person A is a troll and person B is a troll-friend, the troll-friend can only help the troll or hurt someone else. But he can't do both. And the only way he can get another vote is for someone to vote him up.

Of course everyone would have to start with a certain number of votes, or no voting could take place to begin with. We could possibly base the number of starting votes on length of registration or some other factor, or even have a set number per hour, day, postcount, etc.
That sounds mind-bogglingly complicated, Tom. But it serves for a start. ;)

[Godot]

I have amended the Privacy Policy to include the following:



Will that work for now?
An excellent start.

[viscousmemories]

That sounds mind-bogglingly complicated, Tom. But it serves for a start. ;)
I'm glad you approve. Is it really that complicated, though? I thought it very simple and straightforward. Of course it is just an off-the-top-of-my-head rough draft, but let me see if I can explain it in simpler terms and with a little modification based on further thought.
Each new member gets X number of votes in a pool - with which he can vote on other member's posts - when he meets a certain requirement (for example a certain postcount, reputation, etc.)
For every (+) vote on one of his posts, a vote is added to his pool.
For every (-) vote on one of his posts, a vote is subtracted from his pool.
That would be the gist of it. How many votes the member would start with, whether they would be replenished, if they would go into negative numbers and effectively lose all rights to vote, etc. yet to be determined. Does that help clarify it a bit? Do you see any inherent flaws in the idea?

[livius drusus]

Well, I for one feel that the introductory threads have a lesser value on the boards compared to a more higher level of discussion and debate. But then again, that's likely just me.

I don't think socializing is of lesser value than other sorts of discussion or debate. It might not be as personally interesting to me, but I suspect I'm somewhat in the minority on that score.

It laregely would depend on what sort of environment you wish to foster on the board. If you're aiming to have a fair bit of serious discussion, you might prefer that posts of that nature are in a different category than the fluff ones. To that end, perhaps removing voting criteria from a particular set of posts (I mean, how the hell can you vote (+) or (-) for a hello post?) so they do not even factor into the decision.

Since the voting and filtering system are tied together, posters in the social forum would not be able to hide or ignore abusive posts using their default filters if voting were disabled. They would still have some measure of control via the other custom features, but the filter system would go along with the voting system.

Social fora tend to be very large and populous; I think it would rather defeat the point of the custom filter system to render it inoperable in the areas with highest traffic.

[Godot]

The biggest problem I see involves the motivations behind voting. If I make a universally unpopular post (despite the points therein being valid) and the community votes me down because of it, why am I punished? What is stopping you from going through my history and voting me down on each of them because you're mad at me right now? I shouldn't have my voice quelled as a result of that sort of intervention.

Case in point: my above post where I replied "an excellent start" to liv's rule amendment has been voted down already. How the fuck is that possible unless someone is being malicious? It's bullshit.

[livius drusus]

The biggest problem I see involves the motivations behind voting. If I make a universally unpopular post (despite the points therein being valid) and the community votes me down because of it, why am I punished? What is stopping you from going through my history and voting me down on each of them because you're mad at me right now? I shouldn't have my voice quelled as a result of that sort of intervention.

Case in point: my above post where I replied "an excellent start" to liv's rule amendment has been voted down already. How the fuck is that possible unless someone is being malicious? It's bullshit.
Pettiness is indeed a problem with the current setup, which is on the major reasons it is so clearly untenable. I think Tom's idea goes a long way towards correcting those kinds of inequities.

[viscousmemories]

The biggest problem I see involves the motivations behind voting. If I make a universally unpopular post (despite the points therein being valid) and the community votes me down because of it, why am I punished?
You are punished because in a self-moderating community, the community decides what is and is not quality content. Just because your argument is valid and your conclusions sound does not mean your contribution will be appreciated by the other members of the community.

What is stopping you from going through my history and voting me down on each of them because you're mad at me right now? I shouldn't have my voice quelled as a result of that sort of intervention.
Indeed avoidance of that would be the whole point of my proposed solution.

Case in point: my above post where I replied "an excellent start" to liv's rule amendment has been voted down already. How the fuck is that possible unless someone is being malicious? It's bullshit.
I couldn't agree more, and I've seen quite a number of posts here voted down in the last 48 hours for no apparent reason. Again, that's why I think it's imperative that we reconsider our current system.

[Godot]

Indeed avoidance of that would be the whole point of my proposed solution.
Unfortunately, your proposal is just as open to the same abuse, albiet in a slightly reduced capacity.
It also has the unintended side-effect of perpetuating a clique. Those individuals with unpopular opinions will be voted down into oblivion simply by virtue of holding those opinions. If I hold a popular point of view, I get more votes and can perpetuate the cycle. Freethought it ain't.

[Adam]

The biggest problem I see involves the motivations behind voting. If I make a universally unpopular post (despite the points therein being valid) and the community votes me down because of it, why am I punished? What is stopping you from going through my history and voting me down on each of them because you're mad at me right now? I shouldn't have my voice quelled as a result of that sort of intervention.

Case in point: my above post where I replied "an excellent start" to liv's rule amendment has been voted down already. How the fuck is that possible unless someone is being malicious? It's bullshit.

The situation where someone might go through your history and vote all your posts down in a fit of pique was what I was attempting to address when I suggested that it might be useful to limit the impact that any given individual can have on any other individual's reputation. I still think it's a good idea, assuming that there's a way to handle it with a vB hack.

As far as voting down a post full of valid, yet unpopular points, that's certainly a possibility, but I'd like to hope that we'll manage to keep enough thoughtful and levelheaded participants around that your valid points would earn you positive votes to balance out the negatives from those who simply don't like your ideas. Granted, I'd like to hope that, but I haven't been voting many posts up yet, despite having read a number that were full of valid points. Call me part of the problem...

[Godot]

I thought I'd address these bits in a separate post.
You are punished because in a self-moderating community, the community decides what is and is not quality content. Just because your argument is valid and your conclusions sound does not mean your contribution will be appreciated by the other members of the community.
-and-
I couldn't agree more, and I've seen quite a number of posts here voted down in the last 48 hours for no apparent reason. Again, that's why I think it's imperative that we reconsider our current system.

These two statements are at odds with one another. In the former, you are saying that the community decides what is and is not judged as appropriate. If you insist that this is to be the case, then any need for reconsideration is irrelevant. If the community is to decide, then let the community decide. Their decisions may be erratic and irrational, but it is what you were aiming for.
The "no apparent reason" is moot when applied to the population dynamics of a community structure, which are inclusive by nature.

[Adam]

In analysing the data, i think there should be less concern at Type II errors and more for Type I. I would like to know what kind of demarcation criteria will be used to avoid confusing the two.

I'm confused as to what you mean (I guess the confusion actually stems from my not being 100% sure what the intent of this exercise is). What sorts of Type I errors are you worried about? Things like an apparent pattern where I always vote against, say, liv's posts, when no such intent actually exists on my part?

[viscousmemories]

These two statements are at odds with one another. In the former, you are saying that the community decides what is and is not judged as appropriate. If you insist that this is to be the case, then any need for reconsideration is irrelevant. If the community is to decide, then let the community decide. Their decisions may be erratic and irrational, but it is what you were aiming for.

The "no apparent reason" is moot when applied to the population dynamics of a community structure, which are inclusive by nature.
I don't think the comments were at odds with each other. We have no problem with some decisions of the community being erratic and irrational. Human beings, emotional creatures that most are, have that tendency. However we also expect a fair amount of careful, reasoned assessments. Our goal therefore would be to minimize the former and maximize the latter, without enabling a single member to have too great an influence on another member or the community as a whole.

As I explained above, one way we think we can accomplish this is by giving a member a limited number of votes he/she can cast equivelant to the number of positive votes his/her posts have received. Another idea is to implement a cap, as Adam suggested, on the number of a particular members' posts another member can vote on in a given timeframe.

Any other ideas you or anyone else has about how else we might encourage positive use of the system and discourage abuse would be more than welcome. Given the paucity of online forums of this type that have attempted such, we do not expect the solution to come easily.

[Petra]

I haven't read this whole thread - been a bit caught up with events in the Bay of Plenty - but how easy would it be to have a system like CF has - at CF, you can leave a comment with your reputation vote, and even if you didn't, the recipient can see who gave them the rep point and if they leave a comment, why they gave them that rep point.

Here's a screen shot of mine to illustrate what I mean.

[viscousmemories]

I haven't read this whole thread - been a bit caught up with events in the Bay of Plenty - but how easy would it be to have a system like CF has - at CF, you can leave a comment with your reputation vote, and even if you didn't, the recipient can see who gave them the rep point and if they leave a comment, why they gave them that rep point.

What you're describing is actually the default reputation system built in to vBulletin, which we aren't using as it was designed. We could configure the software to allow people to leave comments when voting on a post - that might be a feature that people would like to have - and we could also make it so that the member could see who voted on their post.

Unfortunately there are predators out there who probably wouldn't mind leaving a nasty comment or two, and who wouldn't be deterred by their victim knowing it was them throwing the stones. On the contrary, I suspect some people would actually prefer that their victim be aware of exactly who is hurting them. I'm afraid the only real solution is to disempower such people by limiting their ability to vote.

[Petra]

Yeah, I think that is why CF changed their rep system so that you could only give a positive rep point as the nasty comments and assholish behaviour were rife.

Are only members allowed to vote, or are guests allowed to vote, too?

[viscousmemories]

Yeah, I think that is why CF changed their rep system so that you could only give a positive rep point as the nasty comments and assholish behaviour were rife.

Are only members allowed to vote, or are guests allowed to vote, too?
Members only, fortunately. Earlier today a bug in the "Who's Online" page revealed a guest attempting to downvote one of my posts. It didn't work, but at least it alerted me to the bug and I fixed it. There are some real winners out there to be sure. :)

[Petra]

Well, I am unsurprised but greatly disappointed that people feel so fucking petty as to go on a silent point lowering crusade.

Obviously this is a person of zero integrity with a pathetic little chip on their shoulder. A wanker of underwhelming proportions.


Edit: As an admin are you able to see logs showing who does this juvenile nonsense?

[Hugo Holbling]

What sorts of Type I errors are you worried about?

Thanks for asking, Adam. In this instance, it seems that a Type I error would be not seeing systemic abuse where it does in fact exist, whereas a Type II error would be to see it where it does not. Although people have noted that the former is something they wish to avoid, the latter can be just as much a problem. If we had all had access to the voting data, it would be far too easy to make a Type II error and conclude that vendettas are being played out. For example, the anonymous voter might just have been a member who had logged out.

The consequences of wrongly assuming systemic abuse would be just as unfortunate for this place.

[viscousmemories]

To reiterate livius' earlier point, we have at this point in time decided that we will not be compiling or attempting to analyze the vote data. As such it probably isn't a very productive use of our time to continue discussing a hypothetical situation wherein we do so.

A couple of alterations to the existing vote system have been proposed, including limiting each members' pool of votes to the number of positive votes their posts have received, and putting a cap on the number of votes an individual is allowed to cast on the posts of another specific individual.

Thoughts?

[Godot]

The criticism being offered in this thread is hardly being well-received by the community. This is evidenced largely by the cumulative tally on many posts, across the board.

As such, since my contributions do not appear to be welcome here (while the words offered in support say one thing, the silent majority says otherwise), I am removing myself from all critical discussion. Good luck.

[viscousmemories]

The criticism being offered in this thread is hardly being well-received by the community. This is evidenced largely by the cumulative tally on many posts, across the board.
Are you sure you can make that assumption? What evidence is there that any of the votes that have been cast on this forum have a thing to do with this particular discussion?

As such, since my contributions do not appear to be welcome here (while the words offered in support say one thing, the silent majority says otherwise), I am removing myself from all critical discussion. Good luck.
Well, I disagree that there is any conclusive evidence of such an opinion, but if you would rather not participate that is of course your decision to make. I'm sorry to see you go, though. I'm sure you could contribute a lot to this forum if you were so inclined, and we could use all the support we can get.

[viscousmemories]

Edit: As an admin are you able to see logs showing who does this juvenile nonsense?
My apologies, I missed this edit the first time around.

We discussed this possibility earlier in this thread. Yes, it is possible for me to run a query that pulls all of the vote information from the database. However Godot made a very good argument for why it would be unethical to do so (at least without fair warning to and consent of the voting members) so I have not run that query since my first effort to work it out.

But even if we decide at a future time that compiling and analyzing the vote data would be beneficial, we would not only make sure that all the members are aware and consenting, but we would very likely "scramble" the results of the query, so to speak, to obfuscate any personally identifying information. As Godot also pointed out early on (unless I misunderstood) people vote for all sorts of different reasons so it wouldn't be particularly useful to consider the voter or post author in the data analysis.

[Hugo Holbling]

A couple of alterations to the existing vote system have been proposed, including limiting each members' pool of votes to the number of positive votes their posts have received, and putting a cap on the number of votes an individual is allowed to cast on the posts of another specific individual.

Thoughts?

In the first instance, a clique may abuse the system just as easily as currently. In the second, why should the freedom to cast an opinion be restricted? The basic problem here is epistemic: how can we know why someone is voting for or against posts or members, short of asking them? It may be that someone is abusing the system by voting me down, say, but they may just dislike my tone, approach, ideas or a myriad other things. If a particular person votes me down at every opportunity, it doesn't follow that they are abusing the system; indeed, it's far more likely that i suck and am being told as much. How would we estimate the probability of systemic abuse, in any case? Speaking as a mathematician, i don't see any easy answers.

As i explained previously, there are two possible errors we can make: on the one hand, not doing anything about abuse of the system when it is genuinely happening; and, on the other, supposing abuse exists when it in fact does not. The reason i asked for demarcation criteria is because it seems to me that this problem is insoluble: we simply cannot know, and even if i - or anyone else - was to come up with a statistical model to provide confidence intervals, it's hard to see how they could have any value here or form the basis of any action against a member.

[Dingfod]

I'm not inclined to use the voting system except to make negative marks for "bad form", but I might if I think a succinct post that nails the topic merits a positive vote. Just to vote based on whether you agree or not turns each vote total into an opinion poll rather than an indicator of post quality. I cannot speak for others though, they may view it differently. So, if I see a negative vote total when I see it differently, I might vote positive just to offset someone with a personal vendetta, kind of like my mother voted straight Democrat ticket to offset my grandmother's straight Republican vote.


Warren

[LadyShea]

Hugo's last post makes a lot of sense, specifically

how can we know why someone is voting for or against posts or members, short of asking them? It may be that someone is abusing the system by voting me down, say, but they may just dislike my tone, approach, ideas or a myriad other things.

on the one hand, not doing anything about abuse of the system when it is genuinely happening; and, on the other, supposing abuse exists when it in fact does not.

I think we may need to just see what happens for a bit, see if the majority of people can be mature about it, rather than assume widespread abuse right off the bat.

I, personally, do not plan to filter anything based on points or reputation, because that would be allowing others to determine what is worthwhile reading or not for me, and that makes no sense.

[Dingfod]

I, personally, do not plan to filter anything based on points or reputation, because that would be allowing others to determine what is worthwhile reading or not for me, and that makes no sense.Nor am I, but it won't stop me from voting.


Warren

[Scotty]

Whew, got through all of that.

What an interesting discussion, my only desire was to see if the data existed to make a report to give information so that you could tell if somebody was abusing the system.

Obviously figuring out statistical models isn't something I know well, but I guess I was slightly surprised to find that you didn't know the information was in the database (it just seemed obvious to me that it had to be available in some form).

I would think that a proper report, with enough data points should indicate some problem, or at least a place to look for an issue (a trend). The problem with this is what Hugo was mentioning and what I was thinking, is once you see the potential problem, you would need to ask that person what they were doing (not that you would get an answer of course if they were just being annoyed with a poster at the moment, or were trolling).

I also noticed that seeing the post vote on the bottom did lead to one thing. Godot seemed very concerned about how this information would be used and how it would affect people. Well, he got affected by seeing his vote count go down when we was just being concerned. What does that lead him to think?

Just because somebody seems to have a bad tone doesn't mean the information in the post is invalid. Does it mean you should vote it down? Does it mean you know WHY it was voted down? Does that matter? I would think so. If my tone is poor and I get voted down, but my information is valid, how do I know which it is? You can't always get that information across in a post.

How about even more voting? :D You could have "valid point" voting and "emotional content" voting (I am just THROWING this out because I just thought of it). That way, a person could understand more why the post wasn't to somebodies liking.

As for reporting on the data, maybe if the system is laid out well, in other words, you keep initial errors to a minimum (or not errors, but "post vote slutting") then maybe you would never need to look at the data, thus keep most or all of it private. Plus, if you split the voting by emotional response and valid point (if people used it) then it would give more meaning to the poster.

A way of leaving a comment in the original vb code in the "emotional content" vote?

Hmm, just dumping out here again:
Each registered user has 1 vote for each other user depending on post count per day (or time period). For every 50 posts (1-49 gets you one vote, or maybe 0-49 so if you just lurk you get one per person per day) you get one extra vote per time period (an undisclosed period to the user, so they don't want and vote a bunch, then do it again a minute later), minus your negative score (if you have any) plus your positive score if you have any. So, 50 posts + 50 positive votes +0 negative votes gets you 2 votes per day per person (but only one vote/post)
If you then split votes between "content driven" and "emotionally driven" they could have different weights. Or, the emotional one could just be an indicator for you in how your posts come accross and have no bearing on the number of votes. Or not ;)

This could be the privilege you give to long time posters, or it could just be a darn popularity contest again! <sigh> I was trying to come up with something...

Anyway, just a thought, then maybe you wouldn't have to look at the data at all, unless somebody was able to log in as an old-time user and vote somebody down big time, or a long-time popular poster didn't like a newcomer and voted them into oblivion. This then, would require a report.

Ah well.

-Scott

[Hugo Holbling]

That was an excellent post, Scotty. Here are a few questions for you:

I would think that a proper report, with enough data points should indicate some problem, or at least a place to look for an issue (a trend). The problem with this is what Hugo was mentioning and what I was thinking, is once you see the potential problem, you would need to ask that person what they were doing...

Of course, but what if i voted you down on every occasion? It doesn't follow that i'm doing so deliberately and we run up against a famous problem in philosophy if we ask when it becomes reasonable to suppose that i am. I'm asking if there's any way to determine when intervention to stop such behaviour becomes justified.

If my tone is poor and I get voted down, but my information is valid, how do I know which it is? You can't always get that information across in a post.

No, nor can you get it across at all with this voting system. It's this unfortunate circumstance that means you can never know why a post or series of posts was voted down and hence make it the basis for action.

As for reporting on the data, maybe if the system is laid out well, in other words, you keep initial errors to a minimum (or not errors, but "post vote slutting") then maybe you would never need to look at the data, thus keep most or all of it private.

Sure, but it would remain raw data. The issue is with the interpretation, but that's what is far from clear.

Hmm, just dumping out here again...

I appreciate your efforts, but why should i be restricted in - for example - how many times i can vote in favour of your posts, or anyone else's? If there happens to be a campaign against me, say, then i get less power to influence the community in spite of how hard i try to be positive. Even if all other things are equal, why should i not be allowed to, say, vote down multiple porn threads if i happen to disapprove of such things? If the community is supposed to decide such things then why restrict their chance to vote? A troll may have a valid opinion, after all.

I was trying to come up with something...

Well, look at LadyShea's nice point:

I, personally, do not plan to filter anything based on points or reputation, because that would be allowing others to determine what is worthwhile reading or not for me, and that makes no sense.

If you don't like my posts, say, then surely it would be a more effective strategy to put me on ignore? Even if many people dislike me, it will take quite an effort to get me below a points threshold and even then, as LadyShea hints at, how many people are actually going to filter their posts in any case? I guess i don't understand why voting is required at all.

[Scotty]

Of course, but what if i voted you down on every occasion? It doesn't follow that i'm doing so deliberately and we run up against a famous problem in philosophy if we ask when it becomes reasonable to suppose that i am. I'm asking if there's a