
Language’s future-oriented framing 43
homogeneous country (−1.58), and India (1.33), Luxembourg (1.22), Belgium
(1.04), and Switzerland (0.98) being the most heterogeneous countries.
Given our theoretical interest in a cross-national comparative analysis, we
used a random-eects model estimation. We considered a country fixed-eects
model, but it was not applicable to our setting because it only used within-
country dierences, essentially discarding information on dierences between
countries, which is vital given our theoretical interest in cross-national com-
parison. The next step was deciding whether to use a pooled ordinary least
square