Table 1 Summary of the four baseline periods and the story period (in chronological order)

\(1\)

\(b_1\): Four years pre-story period

Since many fact-checks concerned the US (elections), this baseline period should represent a possible election period (a term lasts four years in the US).

\(b_{1}^o = \widehat{{\dot{T}}^{o}} - 4\ \text {years}\)

\(b_{1}^e = \widehat{{\dot{T}}^{e}} - 4\ \text {years}\)

\(2\)

\(b_2\): Two years pre-story period

This baseline period was used to account for seasonality, meaning that it would compensate for false positives that only occur due to seasonal effects.

\(b_{2}^o = \widehat{{\dot{T}}^{o}} - 2\ \text {years}\)

\(b_{2}^e = \widehat{{\dot{T}}^{e}} - 2\ \text {years}\)

\(3\)

\(b_3\): Close pre-story period

This baseline period was located as close as possible before the onset of the story (but, ideally, without overlapping). This reduces the likelihood of violating assumption iii, as ample changes in public discourse arguably occur less often in a shorter time period.

\(b_{3}^e = \widehat{{\dot{T}}^{o}} - 30\ \text {days}\)

\(b_{3}^o = b_{3}^e - \widehat{{\dot{T}}^{\Delta }}\)

–

Story period \(s\)

Time period where the story spread. Everything before and after this period is considered to be extratemporal.

Onset: \(\widehat{{\dot{T}}^{o}}\) End: \(\widehat{{\dot{T}}^{e}}\) Duration: \(\widehat{{\dot{T}}^{\Delta }}\)

\(4\)

\(b_4\): Post-story period

This baseline period was located after the (estimated) end of the story. It should compensate for very sudden changes in public discourse, which cannot be captured by pre-story periods.

\(b_{4}^o = \widehat{{\dot{T}}^{e}} + \frac{1}{4} \cdot \Delta ^{\rightarrow }\) \(b_{4}^e = b_{4}^o + \widehat{{\dot{T}}^{\Delta }}\)