Voting on our problems

Voting on our problems

Voting is one of the most frequent patterns we see online. Companies like Amazon, StackExchange, Yelp and others use it as a means to increase user engagement and to collect data that could give them insights about our preferences and how they can be used to foster additional growth. Stars, up/down arrows, likes and plus signs, checkmarks and others draw our attention and encourage us to participate. Through our votes we exchange information about the quality of cars, experiences at restaurants or hotels, the quality of service at various companies among other things. Our votes and the corresponding data we create are often useful to people in a similar position who need more data to make better decisions. Voting is also the instrument through which we periodically elect our government.

Whether Coca-Cola or Pepsi taste better has been shown that it can be determined through a blind test—one that requires a form of voting at the end—without which both drinks would still be perceived as equal. This means that voting can be of great help when we need to evaluate very subtle nuances. If we had ten bottles of wine, it would have been hard to judge which one would be of highest quality, with the chance that various people have a different opinon. Without keeping track of some scores—essentially voting,— we would be unable to sort the bottles in their order of perceived quality.

Voting is subjective, but the more people participate in it, the more valuable it becomes. The collective knowledge about the value of a book is reflected in the number of votes and reviews it has received on Amazon (for instance). As soon as someone rates a less popular title higher than seemingly better ones, the community will respond accordingly and improve the rank of the better titles as well. While this seems fair on the surface, it may increase the perceived difference between the quality of an existing book and new ones, which can discourage new authors from publishing their work.

Without some form of voting on Wikipedia's article quality, the encyclopedia would be chaotic and hard to trust. If something wrong is discovered, a volunteer can easily fix it or inform a more compentent person that they can do it. Thus, voting helps to sustain high quality. But on blogs or Q&A websites, people may seem to be more interested in getting more votes for their comments and less in providing more value per answer. Someone may ask for votes, putting pressure on others, where they feel that a vote isn't an appropriate response yet. Votes should never be forced.

Different people have different worldviews, which is why their votes could never be "accurate". If they knew what others have chosen to vote on, this would affect their own choice (especially in the case of long queues in front of the urns). We are very sensitive to the biases of the people around us.

At work, knowing the preferences of our colleagues can reduce tension within the team. Everyone makes an internal choice what they will value most and what they can still tolerate. This compromise between options is a vote we are involved in every day, not realizing it in the speed of small action.

When we choose among many options, it becomes hard to evaluate whether we have considered them all. Some of us may look only at one subset, while others at another. More options suggest that the votes will be scattered and that each one will be given less weight and support. Normally, we should try to limit the number of options. When a single person is allowed to vote multiple times, as in the case of selecting five options out of ten, while others do the same independently, this could lead to some interesting results. It should be impossible for a single person to vote periodically on the same publicly available option. But in private, such periodic voting on our daily performance could form the evaluation criteria by which we learn to track our progress.

For each option we vote on there are some choices to be made. We can either increase the vote count (upvote) or decrease it (downvote). After this we should no longer be able to “give” a vote; we should only be able to retract it—a fundamental, often neglected part in democractic voting. As soon as we consider that a vote we have given is doing us more harm than good, we should be able to retract it easily (and eventually consider another option) without having to wait needlessly long (say 4 years). A vote is earned every day and from everyone, so it shouldn’t become a mechanism for long-term keeping of the status quo. Anywhere where votes are involved, there will be attempts to manipulate them: by voting multiple times; buying votes; bringing people to vote, which shouldn’t be allowed to; even hacking websites and changing the counters. It can be very hard to guarantee the integrity of the data, which is why we aren’t seeing many attempts to build electronic voting systems online. Many people claim that doing this would stop the buying of votes while making the voting process more fair, but they are simply not aware of the risks that this brings with itself. Any stronger massive attack and all votes will be compromised.

One example where I thought that voting could be useful is in seeing which of the world’s problems people think are most important to solve, at any given time. If the air at a particular place and time is full of harmful microparticles, an affected person should be able to vote for pollution, since it is their biggest problem today. But tomorrow the type of the problem may change, so they should be able to decide by their own criteria whether the new problem is more severe than the old one, and if so, to retract a vote they gave previously and use it for the new problem. Eventually this could provide a pulse of our problems in the way we perceive them “from first hand”, not necessarily in the way people living thousands of kilometers away want us to believe they are. Even if the list of problems may seem incomplete, it is probably a good start. If you have a problem that doesn't fit into any of these categories, feel free to email me so I can include it.

A paragraph of caution: A real working system would require more secure voting mechanisms to validate identity and to prevent people from voting multiple times by changing their accounts and IP addresses. Authentication would be just a basic requirement to keep automated bots aside and to guarantee more accurate results. Even authentication wouldn't be enough to guarantee accurate results and accuracy may often seem misleading. Ron Rivest has published some papers on voting where he discussed some of the problems and the need for data encryption. Keeping track of who voted on what (and possibly when) can create lots of data and it’s best that we create the right scheme from the start. In order to not hurt the the neutrality of the page on which people vote, the number of votes so far and the rank of the problems should never be revealed. Otherwise this can lead to fewer people participating, which is a bad outcome. But this information is still revealed here to signify that we are all connected and that our choices influence each other in surprising ways, which will hopefully encourage a better discussion about the kind of world we would like to live in. Since the votes are presented as numbers, there is a danger that the voting itself will be perceived as a game, being far from reality. It is also easy to set inappropriate contrast to the small, but powerful attention-grabbing arrow signs, which can quickly distract from the text. What I wanted to have here is fast feedback without refreshing the page. It would have been nice to have an immediate "elevator effect", animating how a particular option changes its rank by going up or down. Finding the end state where the animation should stop, shifting the other options accordingly to make room for it (with some opacity and easing change), synchronously moving the scrollbar with the option presented some problems which I didn't foresee. The case where someone clicks the up arrow of the bottom-most option, while no or few options of those above it have received votes, would move that option across multiple screens, and although the feedback itself seems right, the duration of it becomes a problem then, or if not everything would need to move too fast, with less chance to be perceived as smooth. For this reason, I have chosen to remove the JavaScript for now.

The votes here aren’t meant to be representative, but if you vote, it will be appreciated. The mere creation of this list doesn’t mean that I have the capacity to deal with the problems. But I hope that this list will help you to get more clarity of the problems you feel are important to solve while evaluating them in a wider context. Thank you for your patience and votes.

bit.ly/1zLkTiE