Skip to main content
. 2022 Jul 4;12:11236. doi: 10.1038/s41598-022-14579-y

Table 1.

Correlation of positive mood in the survey with text sentiment measures on both platforms (Der Standard and Twitter).

Der Standard (no shift) Twitter (Shift 1) Twitter (No shift)
LIWC+GS 0.93 [0.82,0.97] 0.90 [0.75,0.96] 0.71 [0.39,0.88]
LIWC 0.74 [0.44,0.89] 0.85 [0.65,0.94] 0.66 [0.31,0.85]
LIWC pos 0.81 [0.56,0.92] 0.80 [0.56,0.92] 0.60 [0.22,0.83]
LIWC neg 0.03 [−  0.42,0.46] −  0.74 [−  0.89, −  0.43] −  0.63 [−  0.84,−  0.26]
GS 0.91 [0.78,0.96] 0.91 [0.79,0.96] 0.73 [0.43,0.89]
GS pos 0.89 [0.75,0.96] 0.91 [0.79,0.97] 0.80 [0.54,0.92]
GS neg −  0.57 [−  0.81,−  0.18] −  0.39 [−  0.71,0.06] −  0.17 [−  0.57,0.3]

The table presents sentiment aggregates (positive minus negative emotions), as well as positive and negative components separately. LIWC+GS indicates the average across both sentiment analysis methods, all other lines present aggregates or components separately for each method. Shift 1 denotes a shift of one day, where survey values precede Twitter values.