A Data-Driven Analysis of Workers’ Earnings on Amazon Mechanical Turk
summary
Collected via the Chrome Workers plugin from 2,676 workers and 3.8 million HIT records, the study uses temporal clustering (D0/D1) to correct overlap and idle gaps and finds a median hourly wage around $2/h (naive 1.77, clustering up to 2.11), only 4% earn above $7.25, and per-requester median pay is $4.57/h. Unpaid work is large: returned tasks 31.8% of work time, rejected tasks 7.7%, and thousands of hours lost to gaps between tasks.
AI-written summary · about
chapters (13)
- 0:00Introduction to study
- 1:32Data collection method
- 3:02Calculating hourly wages
- 5:29Temporal clustering method
- 7:16Hourly wage distribution results
- 8:46Factors behind low wages
- 9:07Impact of unpaid work
- 10:55Returned and rejected tasks
- 12:01Time between tasks
- 13:11Requesters' role in wages
- 14:58Key takeaways summary
- 15:24Discussion and future work
- 17:36Q&A session begins
related talks
05mTowards Strata 2014
05mLukas Biewald interviewed at Web 2.0 Expo New York
09mUsing your technical skills with Amazon Ads
40mReady, set, data: Perfecting the art of talent matching with AI
24mPROCESSING 40 TB OF CODE FROM ~10 MILLION PROJECTS WITH A DEDICATED SERVER AND GO FOR $100
14mDesigning Videogames to Crowdsource Accelerometer Data Annotation for Activity Recognition Research
06mDBTB INT Alya Abbot r
38mUnleashing the serverless data pipelines