A Pairwise version of Paul's Poor Man Poll.

I created a pairwise survey to help Paul decide which ideas seem most appealing to Permies. Here is the Link. Anyone can add additional ideas. You can vote as many times as you'd like too! https://www.allourideas.org/Permies

Have Fun!

Oh, that does look like fun.  How does it work?  Putting two randomly generated ideas against each other each turn?  

I hear whispers that there might be an apple poll in development.  I love apple polls especially because I don't have to do any extra page loads.  It all just sits in the thread.

Here's an example: https://permies.com/wiki/73205/Apple-Poll-Worst-Christmas-Song

If you go on the site there is a great explanation about how pairwise surveys work and how New York City used the tool for their Green City Plan. I preloaded the survey with the ideas that were listed in the last Dailyish email I got about the idea thread.

I tried it out, but got tired of the process because I can go through the poor man's poll and just thumbs up things a lot faster. And also without page loads.

It is cool to know that tool is out there, though.  I wonder what else it would work with?

Pairwise comparisons give you a different type of knowledge of a straight "thumbs up if you like it". It gives you data on relative preferences which can be highly informative.

Here is a demo application of it

Nomoremarking demo

Michael Cox wrote:Pairwise comparisons give you a different type of knowledge of a straight "thumbs up if you like it". It gives you data on relative preferences which can be highly informative.

Here is a demo application of it

Nomoremarking demo

Thanks for posting the demo, I love stuff like this.

So I tried the demo and I see some of what you are saying - and I still don't get the advantage in this case.  For one, the demo was quite different than making a preferential list.  For example, I can rather easily give you a list of my 10 favorite fruits...plants...movies...people.   And if someone else's list were presented to me to order relative to my tastes, that would probably be even easier because there would likely be a bunch of, say, fruit that aren't even on my personal list.  Easier than I can correctly order the color shades (which I have no personal interest in) of two tiles by memory.  So I'm still not sure the application in this scenario, but I would love to know more.

The way the poor man poll works is also black/white - you just thumbs up what you like, and as many things as you like.  You don't make a preferential list.  It's a form of upvoting things people are interested in. So to me, it seems simpler.

But I would like to understand how seeing someone's preferences in a different way (like the pairwise results) would work and possibly benefit Permies.com.  I'm very interested in polling methods.

To anyone reading this -  the poor man's poll  (upvote poll) in question is this one:  https://permies.com/t/172960/permaculture-projects/PTJ-projects-poor-man-poll

That is a great link Michael! The Judgement test was highly entertaining. I definitely thought I'd do better. LOL.

Kim, there are a couple videos that explain pairwise on the allourideas website, but let me address one thing in particular. You mentioned getting tired of the process. That is exactly one of the reason that pairwise was created. When social scientist are trying to gather information through surveys they often have to deal with incomplete surveys. Typically people will agree to do the survey but get tired, or just stop caring enough to finish, or get interrupted. This information typically has to be thrown away because it doesn't conform to the statistics being used to correlate the information.

So pairwise uses a method and statistics to gather information from people that want to give more info that typical and those that give just a little.

Also, upvoting on the forum may be simpler but it doesn't allow you to compare the votes in any meaningful way. You can see a difference in the number of votes but that difference is meaningless.

Here is an example: Suppose there are ten ideas on the forum thread and over the course of a week one idea is clearly preferred and has more votes; lets say it has 100 votes. Then a new idea is added to the list that is very popular and starts to get a lot of votes. And after two weeks this idea has 100 votes and the previous best idea has 150 votes. Which idea is actually preferred? Well, you have no way of knowing. So the poll is meaningless. Except as a way to count votes over time.

I understand that the page loading can be an issue. Unfortunately a forum is definitely going to be able to present more information an any given time.  

Andrew Cegielski wrote:

Also, upvoting on the forum may be simpler but it doesn't allow you to compare the votes in any meaningful way. You can see a difference in the number of votes but that difference is meaningless.

Here is an example: Suppose there are ten ideas on the forum thread and over the course of a week one idea is clearly preferred and has more votes; lets say it has 100 votes. Then a new idea is added to the list that is very popular and starts to get a lot of votes. And after two weeks this idea has 100 votes and the previous best idea has 150 votes. Which idea is actually preferred? Well, you have no way of knowing. So the poll is meaningless. Except as a way to count votes over time.

I think Paul's goal s to see which are the most popular class topics, and there can be multiple popular topics.  Like the Poor Man Poll is a way to narrow down the topics with the most interest from a big list. It seems to work well for that purpose.

Now if he was asking asking "Which one single class topic is the best?"  I can see how it wouldn't work for that motivation.

But theory doesn't really matter.  Let us know what you find out from your poll, and you will be able to see the results of Paul's poll. And if there is something really informative to be gleaned from yours, tell us!  That will be great to find out.