²³If this equation was used to estimate the relationship between urban pride and the level of stake holding the effect of the city itself would be subsumed within the error term _i and go unrecognised as such. By contrast, the random intercepts model (equation 3) allows this inter-city heterogeneity to be recognised.