Fig 1.
This diagram depicts the research model as a multilevel model with longitudinal country-level observations at level 1 and ecological country variables at level 2. Nested frames (grey dotted lines) indicate levels of sampling, boxes indicate variables, and arrows represent effects. Hypotheses are referenced in ellipses; the plus or minus sign in the grey or white ellipse indicate a positive or negative effect, respectively. A dashed ellipse (for H4) means that the effect is directionally indicative, but not statistically significant. Variables modelling the level-1 variance are not shown in the diagram, but are included in the statistical model.
Table 1.
Variables in the research model.
Fig 2.
Social contact avoidance, pandemic threat level, and government restrictions.
For eight countries from different geographies, this figure shows the extent of social contact avoidance (solid black line) from February 15, 2020 to January 10, 2021, alongside the pandemic threat level (dashed grey line), and government restrictions (dotted grey line). All three lines are on a separate scale; the line for the pandemic threat level is normed between zero and the country’s maximum threat level during the time period.
Fig 3.
Social contact avoidance during the COVID-19 pandemic.
Darker green colors signify countries with greater social contact avoidance during the COVID-19 pandemic. Lighter green colors stand for countries where social life went on more or less as usual. The measure used is the median plus one standard deviation of all residential mobility changes reported by Google for about 330 days from February 2020 to January 2021, against a baseline in January 2020. The data is available for 135 countries; no data is available for countries marked in light blue. (This is a chart created with Microsoft Excel®, based on data calculated by me in this present research; reproducing this map is thus considered to be fair use).
Table 2.
Correlation matrix.
Table 3.
Effects in the HLM contextual model.
Table 4.
Robustness of effects.
Fig 4.
Policy and practical implications.
This chart summarizes the study’s policy and practical implications. Because defensive behavior in a pandemic only makes sense at the group level, health-based communication needs to consider cultural and other contextual aspects.