01/29/20 – NAN LI – Beyond Mechanical Turk: An Analysis of Paid Crowd Work Platforms

Summary:

Since the AMT has long been the research focus for the study of crowdsourcing. The author compared a series of platforms and their diverse capabilities to enrich future research directions. The aim of his work is to encourage more investigation and research studies from different sources instead of focusing only on AMT. The author reviewed the related work, pointed out the AMT limitation and problems that have been criticized, then defined the criteria to assess the platform. Mainly in the paper, the author performed a detailed analysis and comparison among seven alternative crowd work platforms. Through the analysis and comparison with AMT, the author identified the same approaches that each platform has been implemented, such as the peer assessment, qualification tests, leaderboards, etc. Besides, the author also found the difference such as the use of automated methods, task availability on mobiles, ethnic worker treatment, etc.

Reflections:

I think the most serious problem in the research study is the limitation of the scope of the investigation. Only a variety of usage of resources can make the research results credible and highly applicable. Thus, I think it is very meaningful for the author to make this comparison and investigation. This is what research should so, always explore, always compare and keep critical. Besides, with the increase in popularity and the increase in the number of users, AMT should pay more attention to improving its own platform, rather than keep staying at a level that can meet current needs. Especially now that the same type of platforms are gradually increasing and developing, they are attracting more users by improving a better management service or develop reasonable regulation and protection mechanisms. Although, until now, AMT is the biggest and most developed platform, AMT still needs to learn from other platforms’ advantages.

In spite of that, other platforms should also keep update and try to create novelty features to attract more network users. A good way to improve their own platform is to always consider what the user’s requirement is. Hot topics which include ethical issues and labor protection issues should be considered. Besides, how to make good use of these platforms to a great extent to improve their product quality is also worth considering.

A short path towards improvement is through discussion. This discussion should include the company, the client, the product development team and even the researcher. As for the companies, they should always ask feedback from their network users. This is a baseline for them to improve not only their platform and user experience but also their product. Also, companies should discuss with each other and even though with the researchers to think about solutions to current problems in their platforms. Even though we cannot predict how things will go. Will these platforms last long? How many of these works will be available with the development of technology? I still believe it is worth to work hard to make these platforms better.

Questions

  • Why researchers are more willing to do research on AMT?
  • Is there any solution to pursue researchers do research on other platforms?
  • For the other platforms besides AMT, what is the main reason for their users to choose them instead of AMT?

2 thoughts on “01/29/20 – NAN LI – Beyond Mechanical Turk: An Analysis of Paid Crowd Work Platforms

  1. I think the main reason for the researchers more willing to do research on AMT is that most of the workers and the companies are using AMT and researches on AMT will raise more attention and be more valuable. For this reason, I have no idea about how to let researchers pay attention to other platforms. However, some researchers may pay attention to other platforms because of this paper, which is a good trend. There is a lot reasons for companies to choose other platforms, for example, lower service fee, workers using particular languages, user-friendly interfaces, more APIs for customization and so on.

  2. I do agree that the investigation conducted by the authors is very useful. I also agree with your point that ‘AMT should pay more attention to improving its own platform’. I would have expected that AMT would have attempted to bring about reforms and aim to overcome their shortcomings with time – which is why I was quite surprised to learn about some of the limitations of AMT that have been mentioned in this paper.

Leave a Reply