Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
119 tokens/sec
GPT-4o
56 tokens/sec
Gemini 2.5 Pro Pro
43 tokens/sec
o3 Pro
6 tokens/sec
GPT-4.1 Pro
47 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Quality Control in Crowdsourcing: A Survey of Quality Attributes, Assessment Techniques and Assurance Actions (1801.02546v1)

Published 8 Jan 2018 in cs.HC

Abstract: Crowdsourcing enables one to leverage on the intelligence and wisdom of potentially large groups of individuals toward solving problems. Common problems approached with crowdsourcing are labeling images, translating or transcribing text, providing opinions or ideas, and similar - all tasks that computers are not good at or where they may even fail altogether. The introduction of humans into computations and/or everyday work, however, also poses critical, novel challenges in terms of quality control, as the crowd is typically composed of people with unknown and very diverse abilities, skills, interests, personal objectives and technological resources. This survey studies quality in the context of crowdsourcing along several dimensions, so as to define and characterize it and to understand the current state of the art. Specifically, this survey derives a quality model for crowdsourcing tasks, identifies the methods and techniques that can be used to assess the attributes of the model, and the actions and strategies that help prevent and mitigate quality problems. An analysis of how these features are supported by the state of the art further identifies open issues and informs an outlook on hot future research directions.

User Edit Pencil Streamline Icon: https://streamlinehq.com
Authors (5)
  1. Florian Daniel (8 papers)
  2. Pavel Kucherbaev (5 papers)
  3. Cinzia Cappiello (3 papers)
  4. Boualem Benatallah (36 papers)
  5. Mohammad Allahbakhsh (6 papers)
Citations (156)

Summary

Quality Control in Crowdsourcing: Insights from a Comprehensive Survey

The paper "Quality Control in Crowdsourcing: A Survey of Quality Attributes, Assessment Techniques and Assurance Actions" by Florian Daniel and colleagues dissects the multifaceted challenge of quality control in crowdsourcing environments. As crowdsourcing continues to be leveraged for various tasks such as image labeling and text translation, quality assurance becomes imperative due to the diverse skill sets and interests inherent in crowdsourcing participants. This survey addresses the entire spectrum of quality control activities in crowdsourcing, laying the groundwork for understanding the state of the art and earmarking future research directions.

Quality Model for Crowdsourcing

The authors propose a detailed quality model that contextualizes quality in crowdsourcing through multiple dimensions — Data, Task Description, User Interface, Incentives, Terms and Conditions, Task Performance, and People. Specific attributes such as accuracy, consistency, clarity, complexity, usability, extrinsic versus intrinsic incentives, privacy, and cost efficiency are meticulously delineated. A rigorous examination speaks to the breadth and depth these dimensions offer in encapsulating the concept of quality in crowdsourcing tasks.

Assessment Techniques

The paper explores diverse assessment methods categorized into Individual, Group, and Computational-based assessments. Rating, self-assessment, peer review, and ground truth comparisons are exemplified approaches reflecting the complexity involved in gauging crowdsourced outputs. The survey signifies that straightforward assessment approaches like rating are widely adopted, while advanced techniques such as fingerprinting and association analysis remain less prevalent.

Assurance Strategies

To bolster quality in crowdsourcing, a variety of assurance actions are identified. These strategies are explained vis-à-vis their application to attribute dimensions — ranging from data cleansing and aggregation of outputs to dynamic task allocation and incentivizing workers. The authors detail the importance of tailored rewards, iterative improvements, social transparency, and prompt feedback as vital actions in assuring quality. Emphasis is placed on both reactive processes like output filtering and proactive measures such as worker engagement techniques.

State of Practice and Future Research Directions

A comparative analysis of fourteen prominent crowdsourcing platforms illustrates the disparity between theoretical quality models and their practical implications. The paper highlights how most platforms prioritize accuracy and extrinsic incentives but fall short in addressing attributes like worker personality and interface usability comprehensively. Notably, research prototypes tend to explore coordination and task automation, while commercial platforms focus on broader, impactful strategies like collaborative work and team building.

The paper identifies substantial avenues for future exploration: domain-specific services, improved interface quality assessment, and regulation of crowdsourcing practices. For instance, fostering ethical standards and regulatory frameworks can potentially guide requesters and platforms in managing tasks effectively. Additionally, developing robust assessment and assurance frameworks can lead to sustainable and efficient crowd work practices.

Conclusion

The survey by Daniel et al. represents a seminal discussion on quality control in crowdsourcing, paving the path for both theoretical advancements and practical applications. The comprehensive nature of their research underlines the intricate balance required to maintain high standards in crowdsourcing initiatives, and it serves as a potent catalyst for further discourse in this burgeoning field. In advocating for more domain-specific and worker-centric models, this paper aspires to shape the future of crowdsourcing toward a more transparent and quality-driven ecosystem.

X Twitter Logo Streamline Icon: https://streamlinehq.com