Table 1. TV episode information and pre-processing details.
| Show | Network | Network type | Show type | Average Twitter volumea | Average TV viewersa | Twitter lagb | Omit startc | Omit endd |
|---|---|---|---|---|---|---|---|---|
| Big Brother Ep. 1 | CBS | Broadc. | Reality | 979.0 | 3,345,932 | 2 | 3 | 2 |
| Big Brother Ep. 2 | CBS | Broadc. | Reality | 502.3 | 4,109,862 | 0 | 2 | 1 |
| Gang Related | Fox | Broadc. | Drama | 66.2 | 1,288,583 | 2 | 0 | 0 |
| MasterChef | Fox | Broadc. | Reality | 101.7 | 2,825,495 | 1 | 3 | 2 |
| Naked and Afraid | Disc. | Cable | Reality | 58.1 | 1,205,048 | 0 | 3 | 2 |
| NY Med | ABC | Broadc. | DS | 41.4 | 4,025,144 | 1 | 0 | 0 |
| Reckless | CBS | Broadc. | Drama | 30.2 | 2,897,098 | 2 | 2 | 2 |
| Suits | USA | Cable | Drama | 120.4 | 1,438,171 | 1 | 2 | 2 |
| Taxi Brooklyn | NBC | Broadc. | Drama | 22.0 | 3,324,714 | 1 | 2 | 3 |
Broadc.–broadcast, Ep.–episode, Disc.–Discovery, DS–documentary series.
a For each episode, numbers of tweets and TV viewers were calculated for each minute of the show and then averaged across the show duration.
b Twitter lag (in minutes) was used to align minute-by-minute Twitter volume values with EEG and TV viewership values, correcting for the lag between what people see on the screen and when they type and post tweets online. For example, a lag of 2 minutes means that tweets from minute (t+2) are aligned with EEG and viewership values for minute t.
c Omit start (in minutes) reflects the number of minutes at the beginning of each show that were omitted from analyses due to outlier Twitter or viewership values (audiences tuning in late or tweeting about the show in general rather than about the content of the specific episode).
d Omit end (in minutes) reflects the number of minutes at the end of each show that were omitted from analyses due to outlier Twitter or Viewership values (audiences for the next show tuning in early or tweeting about the upcoming episode rather than about the content of the current episode).