Complex tasks and motivating crowdworkers – a natural experiment

How to cite

Adena, Maja; Harke, Julian; Janowska-Widomska, Patrycja. 2020. "Complex tasks and motivating crowdworkers – a natural experiment." AEA RCT Registry. AEARCTR-0006744. https://doi.org/10.1257/rct.6744-1.3

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Maja Adena* , Julian Harke , Patrycja Janowska-Widomska

* Primary investigator

Registry ID
AEARCTR-0006744
DOI
https://doi.org/10.1257/rct.6744-1.3
First registered
Last updated
Status
Completed
Trial dates
2020-12-03 – 2021-12-31
Countries
Germany (), United Kingdom of Great Britain and Northern Ireland ()
Registry record
http://www.socialscienceregistry.org/trials/6744

Abstract

Crowdsourcing has become more popular in recent years. However, not much is known about how to motivate crowdworkers except for the use of monetary incentives. This becomes even more important when they are conducting tasks that require high engagement but for which the objective quality is difficult to assess. Such tasks might include video, picture, or text classifications and ratings. Per piece payments may increase the speed but actually reduce the quality. In an online experiment, we study the use of (performance unrelated) recognition and positive versus negative framing regarding (performance-based) payment. We use different (more or less objective) performance measures as well as measures of (objective and subjective) crowdworkers’ motivation and (subjective) job satisfaction.

Experimental design

We follow a 3x2 between-subjects design.
The differences in the first dimension concern the inclusion of recognition: Control treatment (C) is the treatment without any specific recognition; treatment R includes four recognition phrases; and treatment A includes four appreciation phrases.
The second dimension consist of positive versus negative framing regarding the performance dependent payment.

Design details

Randomization method
Randomization by a computer.
Randomization unit
individuals
Clusters
no clusters
Observations
around 4,500 individuals
Arms
around 750 individuals per treatment
Primary outcomes
job satisfaction worker motivation performance see the attachment for details

Cited by

No citation data available for this registration yet.

Keywords