This corpus of political debates annotated with aspect-based sentiment and its analysis. The source corpus consists of two presidential debates of the 2016 US election. The first debate is annotated using two schemas,
1) adjective-noun markers
2) sentence-based without marker.
The third debate is annotated with 1) schema.

Dataset is further described in:

@inproceedings{gold2018corpus,
title={{Corpus of Aspect-based Sentiment in Political Debates}},
author={Gold, Darina and Bexte, Marie and Zesch, Torsten},
year={2018},
publisher={Verlag der {"O}sterreichischen Akademie der Wissenschaften},
pages={89--99},
location={Vienna, Austria},
booktitle={KONVENS}
}