Table 1.
Variables used in the models, their descriptions, and rationale for their inclusion in the models.
Variable names, descriptions and values are taken from Mammola et al. [32] unless when otherwise stated.
Fig 1.
Results of analyses with Google Trends search volumes (A–C), Wikipedia page views (D–F), and iNaturalist observations (G–I). Left plots show daily search volume levels depicted by locally estimated scatterplot smoothing (LOESS) fits (filled lines) and 95% confidence envelopes (shaded surfaces), where day 0 corresponds to the date of news story publication. Center plots show density of values of change in search volumes after the publication date of each news. Right plots show estimated parameters (and 95% confidence interval) for linear mixed models testing the relationship between the intercept change in search volume after the publication of each news story and different news-level predictors. Exact model estimates are provided in Table B in S2 Appendix. * indicates p < 0.05.
Fig 2.
Negative binomial generalized linear models testing if the monthly number of news published has an effect on the number of spider (order Araneae) observation uploads on iNaturalist (left) and Wikipedia page views (right). Exact model estimates are given in Table C in S2 Appendix.