Introduction

In order to develop effective crowdsourcing aggregation methods for multiple choice question answering(MCQA) and evaluate them empirically, we developed and deployed a crowdsourced system for playing the “Who wants to be a millionaire?” quiz show. Note that, as question and answer texts are originally in Turkish you should use UTF8 format at all times to avoid encoding problems.

Citation

Harvard Aydin BI, Yilmaz YS, Demirbas M. A crowdsourced “Who wants to be a millionaire?” player. Concurrency Computat.: Pract. Exper. 2017;e4168. https://doi.org/10.1002/cpe.4168

Data

Over the period of 9 months, we collected over 3 GB of data using our CrowdMillionaire app. In our dataset, there are 1908 questions and 214,658 unique answers to those questions from CrowdMillionaire participants. In addition, we have more than 5 million offline answers for archived live questions. Our dataset includes detailed information on the game play. For example, our exhaustive timestamps show (1) how much time it took for a question to arrive to a participant, (2) when the question is actually presented to the participant on her device, and (3) when exactly the participant answered the question. We shared this dataset in order to advance the understanding of the MCQA dynamics, after we cleaned and anonymized the data.

Acknowlegement

Foto von Jason Leung auf Unsplash