This chapter deals outliers in public administration and policy studies. They argue that a commonly defined as observations inconsistent with general patterns in a data set, researchers often consider outliers to be undesirable because they do not fit well with hypothesized generalizations or because they exert disproportionate influence in commonly-used statistical estimation techniques. Less commonly, outliers are viewed as providing valuable information rather than being just unlucky realizations of random error. For public policy and public administration researchers, who look for ways to promote change that improves public well-being, identifiable differences between outliers and the more general pattern may contribute to innovative and desirable policy designs and formulations. This chapter introduces a three-phase method we call outlierism: 1) identifying relevant outliers; 2) identifying possible underlying mechanisms that give rise to the outliers; and 3) assessing accumulated knowledge about outliers for use in policy design. Moving from outliers to policy poses methodological challenges, including establishing a credible relationship between identifiable difference and the outcome of interest, determining if the difference is sui generis or something that could be replicated, and translating the differences into working policies. Outlierism seeks to address these challenges in a structured, systematic way that allows a better integration and exploitation of scientific knowledge to improve policy and administration.
Anat Gofen and David L. Weimer
Anat Gofen, Shelly Sella and Drorit Gassner
Which levels of analysis have been employed to explore the influences on, and the implications of, street-level implementation? This chapter suggests that micro, meso and macro levels of analysis are relevant to street-level bureaucracy research because street-level implementation is exercised by individuals, during interactions with clients, in different organizational settings, for the implementation of multiple policies of different professions, in different geographical areas. This review identifies the dominance of micro-level analysis, predominantly using the interaction with client and the individual worker as the unit of analysis. In contrast, meso-level analysis that might use the organization, the implemented policy or the profession as the unit of analysis, as well as macro-level analysis that might refer to the state or country as the unit of analysis are understudied. Consequently, current research rather overlooks street-level implementation variance across organizations, across professions and across states or countries as well as the interrelations between influences of different levels.