Table 1.
Data set sources and specifications.
| Data set | Source | Time range | Size (number of articles) | Type | |||
|
|
|
|
Noncredible news | True news | Total |
|
|
| CoAIDa | Tweets | Until May 1, 2020 | 572 |
1324 | 1896 | COVID-19–specific | |
| FNNb | PolitiFact | N/Ac | 472 | 797 | 1270 | General news | |
| FNN | Gossip Cop | N/A | 16,818 | 5335 | 22,153 | General news | |
| Validation data set 1d | Poynter.org (noncredible news); Washington Post, Associated Press, Politico (true news) | July 20, 2020, to August 8, 2020 | 3874 | 3177 | 7051 | COVID-19–specific | |
| Validation data set 2d | Poynter.org (noncredible news); BBC, AXIOS, CBS News, The Globe and Mail (true news) | January 20, 2020, to June 15, 2022 | 14,398 | 16,232 | 30,630 | COVID-19–specific | |
aOnly the 05-01-2020 folder of the CoAID data set was used.
bFNN: FakeNewsNet.
cN/A: not applicable.
dScraped with the query term “COVID-19.”