Suchergebnisse

This article provides guidelines and illustrates practical steps necessary for an analysis of results from the multinomial logit model (MLM). The MLM is a popular model in the strategy literature because it allows researchers to examine strategic choices with multiple outcomes. However, there seem to be systematic issues with regard to how researchers interpret their results when using the MLM. In this study, I present a set of guidelines critical to analyzing and interpreting results from the MLM. The procedure involves intuitive graphical representations of predicted probabilities and marginal effects suitable for both interpretation and communication of results. The practical steps are illustrated through an application of the MLM to the choice of foreign market entry mode.

Denne artikel demonstrerer, hvordan en analyse og formidling af resultater fra en multilevelanalyse med fordel kan suppleres med grafiske illustrationer.Ved brug af data fra European Social Survey demonstreres det, hvordan man grafisk kan forenkle både fortolkning og formidling af ellers forholdsvist komplicerede resultater som eksempelvis Random Intercept-, Random Slope- og cross-level interaktionseffekter.Demonstrationen foretages gennem en analyse af sammenhængen mellem individuel social kapital og jobtilfredshed på tværs af 23 europæiske lande. Det undersøges hertil, hvordan denne sammenhæng varierer alt efter de specifikke kulturelle forhold, der er til stede i de enkelte lande.

Abstract
Equal treatment is a key feature of modern bureaucracy. However, several studies have shown that public organizations discriminate against ethnic and racial minorities to different degrees. Which organizational features explain differences in discrimination is largely unknown. This article proposes that organizational performance relates to an organization's likelihood of engaging in employment discrimination and argues that poor‐performing organizations tend to be less open to new ideas and that decision makers in such organizations are more prone to stereotyping behavior. The study combines a field experiment in which applications were sent to real job vacancies in 71 Danish public schools with administrative data on the schools. Bayesian analyses show that minority applicants generally faced discrimination but that they experienced a higher callback rate from better‐performing schools than from poorer‐performing schools. Implications for practice and research are discussed.