Difference between revisions of "Collective Action and Decision-making"

From Technologies of Politics and Control
Jump to navigation Jump to search
Line 35: Line 35:
  
 
The articles that we had to read for this week were very interesting to me. I really enjoyed how the first article focused on statistics regarding various experiments and “Who wants to be a millionaire”, a program which I personally loved watching. Even though I respect and find very interesting the point made regarding the percentages being higher and closer to the correct answer for groups and less accurate for individuals, I don’t agree. Math and Statistics aren’t my field and I must admit that I’m not very good at any of them but I feel off the top of my head that it is quite obvious that a group would obtain a higher and more accurate score than an individual because the general population or “average Joe” is likely to get fairly close to the right answer which can be higher or lower but of course adding all the higher scores to the lower ones, I find it to be mathematically obvious that we shall obtain an approximately correct average score. Therefore this being said I find the “Condorcet Jury Theorem”, mentioned in the third article to pretty much respect my personal opinion on the subject. In conclusion I very much enjoyed these articles and I find that in some way these theories emphasizing on group work and force are exactly what Democracy is about. [[User:Emanuele|Emanuele]] 18:12, 19 March 2012 (UTC)
 
The articles that we had to read for this week were very interesting to me. I really enjoyed how the first article focused on statistics regarding various experiments and “Who wants to be a millionaire”, a program which I personally loved watching. Even though I respect and find very interesting the point made regarding the percentages being higher and closer to the correct answer for groups and less accurate for individuals, I don’t agree. Math and Statistics aren’t my field and I must admit that I’m not very good at any of them but I feel off the top of my head that it is quite obvious that a group would obtain a higher and more accurate score than an individual because the general population or “average Joe” is likely to get fairly close to the right answer which can be higher or lower but of course adding all the higher scores to the lower ones, I find it to be mathematically obvious that we shall obtain an approximately correct average score. Therefore this being said I find the “Condorcet Jury Theorem”, mentioned in the third article to pretty much respect my personal opinion on the subject. In conclusion I very much enjoyed these articles and I find that in some way these theories emphasizing on group work and force are exactly what Democracy is about. [[User:Emanuele|Emanuele]] 18:12, 19 March 2012 (UTC)
 +
  
 
Collective action affects us everyday (or at least those of us that read news, shop, or blog online). For instance, if you use Reddit, most likely you're reading posts on the front page that were up-voted--a form of collective action. Zuckerman's article also points out "Amazon's collaborative filtering recs and Google's page rank algorithm." This had a huge effect on business (as we read in the long tail article, for example).
 
Collective action affects us everyday (or at least those of us that read news, shop, or blog online). For instance, if you use Reddit, most likely you're reading posts on the front page that were up-voted--a form of collective action. Zuckerman's article also points out "Amazon's collaborative filtering recs and Google's page rank algorithm." This had a huge effect on business (as we read in the long tail article, for example).

Revision as of 18:25, 20 March 2012

March 20

Mass collaboration and the aggregation of information enable potentially profound changes in business and politics. In this class, we will compare and contrast the transformations in economic life and collective decision-making processes brought on the information revolution. The discussions will also explore the role of open information systems on business and the scope for greater transparency and participation in government, politics and public life.


Assignments

Assignment 3 due


Readings

Optional Readings

  • Divided They Blog - a paper showing trackbacks between political blogs, mentioned by Ethan Zuckerman in his review of Cass Sunstein's Infotopia


Class Discussion

March 20: Collective Action and Decision-making Just Johnny 17:11, 15 February 2012 (UTC)


Even though group intelligence is more difficult to measure than individual intelligence, I do believe that a crowd can outsmart a genius. Imagine for example, if a very intelligent physicist is isolated and only surrounded by other physicists, he/she maybe outstanding in this field, but is limited to what he/she can do. However, when allowed to collaborate with cell and molecular biologists, chemists, mechanical and electrical engineers, medical doctors, veterinarians, etc… a physicist learns to conduct cell mechanics, biophysics, molecular and biochemical experiments, and has the potential to solve health problems, such as coming up with a drug to relax airway smooth muscle cells during an asthma attack. When a group of diverse individuals collaborate, they can solve problems that they otherwise cannot solve individually. The internet has allowed this collaboration to increase globally through technologies such as Skype and E-mail. In some instances, group intelligence depends on its structure and dynamics. For example, Megan Garber, from Nieman Lab, reported that MIT researchers found that “[g]roup intelligence is correlated…with emotional intelligence, http://www.niemanlab.org/2011/05/mit-management-professor-tom-malone-on-collective-intelligence-and-the-genetic-structure-of-groups/. The researchers concluded that a group is more intelligent and is more likely to solve difficult problems when there are more women in it. Simply placing very smart individuals together in a group does not make a group smarter. When I think of a very intelligent individual or genius, I think of my lab principal investigator, who has the ability to lead his lab members and make important decisions. However, he would have not made an informed decision without hearing the lab members deliberate. Qdang 15:52, 19 March 2012 (UTC)


I like the connection between the reality of crowd intelligence in "Who Wants to Be a Millionaire" or marble-jar guessing and the concept of blog aggregates and online communities. It seems like this is a good argument against the dangers of cocooning at some levels. While a given blog/website community is likely there because they all subscribe to a certain set of interests or views, at least you know that if you're going to a big one you're probably getting the very best and most cohesive expression of those concepts. It may not make them right, but it adds value to them as a tool for educating yourself. Aggregates also will provide a balance to the problem that 1 or 2 of every group of 50 people will actually be more accurate than the group at guessing the number of marbles... but only in that one specific trial. If you follow one blogger religiously you are susceptible to their blind spots and moments where they were simply wrong. If you follow a collection of sites and blogs you will be exposed to the correct answers to most questions; whether you realize which is the correct answer is up to you haha. AlexLE 14:04, 19 March 2012 (UTC)


The articles that we had to read for this week were very interesting to me. I really enjoyed how the first article focused on statistics regarding various experiments and “Who wants to be a millionaire”, a program which I personally loved watching. Even though I respect and find very interesting the point made regarding the percentages being higher and closer to the correct answer for groups and less accurate for individuals, I don’t agree. Math and Statistics aren’t my field and I must admit that I’m not very good at any of them but I feel off the top of my head that it is quite obvious that a group would obtain a higher and more accurate score than an individual because the general population or “average Joe” is likely to get fairly close to the right answer which can be higher or lower but of course adding all the higher scores to the lower ones, I find it to be mathematically obvious that we shall obtain an approximately correct average score. Therefore this being said I find the “Condorcet Jury Theorem”, mentioned in the third article to pretty much respect my personal opinion on the subject. In conclusion I very much enjoyed these articles and I find that in some way these theories emphasizing on group work and force are exactly what Democracy is about. Emanuele 18:12, 19 March 2012 (UTC)


Collective action affects us everyday (or at least those of us that read news, shop, or blog online). For instance, if you use Reddit, most likely you're reading posts on the front page that were up-voted--a form of collective action. Zuckerman's article also points out "Amazon's collaborative filtering recs and Google's page rank algorithm." This had a huge effect on business (as we read in the long tail article, for example).

Sunstein has a valid point with ideological cocoons, but does seem flawed. I agree with Zuckerman and the others who labeled his idea as "alarmist." While it's entirely possible (as "Divided they Blog" suggests) for people to seek out news and sites with similar ideologies and have their beliefs continually reinforced, that is not necessarily the way most people 'read the news.' Greater exposure to new ideas or newspapers/news from far away, like the readings said, is a benefit that far outweighs the risk of people forming an ideological cocoon. Aberg 18:24, 20 March 2012 (UTC)


I realize we'll probably dig deeper into the topics of collective action and decision-making in the "Internet and Democracy" classes but I'd love to have a longer reading list on this topic. If anyone has any additional recommendations, please share. Thanks! Aditkowsky 13:17, 20 March 2012 (UTC)

Links