Quantifying model selection uncertainty via bootstrapping and Akaike weights

Picking one ‘winner’ model for researching a certain phenomenon while discarding the rest implies confidence that may misrepresent evidence. Multimodel inference allows researchers to more accurately represent their uncertainty about which is ‘best’. But multimodel inference, with Akaike weights—weights reflecting relative probability of each candidate model—and bootstrapping, can also be used quantify selection uncertainty, in form empirical variation parameter estimates across models, minimizing bias from dubious assumptions. This paper describes this approach. Results simulation example and an study on impact perceived brand environmental responsibility customer loyalty illustrate provide support our proposed