Comparison of STURLA classification results from current and
previous studies
Permutational t-tests with Bonferroni correction were used to test for
differences between cities in STURLA classes. The permutational t-test
was used because we test data representing the population rather than a
sample. The null hypothesis of the permutational t-test is that STURLA
class composition does not differ between the cities. Permutational
Pearson correlations were conducted to determine if the cities
distribution of STURLA classes were similar between cities. These tests
were conducted in R using the package “RVAideMemoire” (Hervé,
2020).