Gaining insight with recursive partitioning of generalized linear models

Recursive partitioning algorithms separate a feature space into a set of disjoint rectangles. Then, usually, a constant in every partition is fitted. While this is a simple and intuitive approach, it may still lack interpretability as to how a specific relationship between dependent and independent variables may look. Or it may be that a certain model is assumed or of interest and there is a number of candidate variables that may non-linearly give rise to different model parameter values. We present an approach that combines generalized linear models (GLM) with recursive partitioning that offers enhanced interpretability of classical trees as well as providing an explorative way to assess a candidate variable's influence on a parametric model. This method conducts recursive partitioning of a GLM by (1) fitting the model to the data set, (2) testing for parameter instability over a set of partitioning variables, (3) splitting the data set with respect to the variable associated with the highest instability. The outcome is a tree where each terminal node is associated with a GLM. We will show the method's versatility and suitability to gain additional insight into the relationship of dependent and independent variables by two examples, modelling voting behaviour and a failure model for debt amortization, and compare it to alternative approaches.     

Multiracial Identities and Employee Absenteeism: Testing Contact and Depletion Hypotheses

Employee absenteeism costs the U.S. economy billions of dollars each year. Recent research finds that workplace characteristics are important predictors, yet variation remains by racial identity. Less is known about the reasons multiracial workers miss work. With data from the 2014 General Social Survey, we test a contact hypothesis and a depletion hypothesis. The findings support the depletion hypothesis. Multiracial workers are likely feeling cognitively taxed from comparatively more interracial interaction in the workplace, resulting in their higher rate of employee absenteeism. Given that the number of multiracial individuals living and working in the United States is expected to grow substantially over the next decade, understanding the factors that explain their work behavior may be beneficial to all social groups.     

Can We Improve the Measurement of Attitudes Towards the Welfare State? A Constructive Critique of Survey Instruments with Evidence from Focus Groups

There is a large and growing literature on welfare state attitudes, most of which is built on random-sample population surveys with standardised closed-question items. This article criticises the existing survey instruments, especially those that are used within the International Social Survey Programme, in a novel approach with focus group data from Germany. The article demonstrates: firstly, these instruments underestimate the inconsistency (the degree to which attitudes logically contradict each other), the uncertainty (the degree to which individuals are unsure about what to think), the ambivalence (the simultaneous occurrence of positive and negative reactions) and non-attitudes towards welfare state activities that common people have. Secondly, the meaning of these items to respondents seems to vary to such an extent that inference based on such measures is questionable. Finally, the article concludes by suggesting some survey instruments that alleviate these measurement problems.     

Older People's “Voices”—On Paper: Obstacles to Influence in Welfare States—A Case Study of Sweden

The official rhetoric of welfare states unconditionally pays tribute to older people's right to express dissatisfaction. In practice, users of older services in welfare states may be deprived of their “exit” options and face considerable constraints when it comes to raising their “voices.” For example, when older people in nursing homes would like to lodge a complaint, they may well be referred to the very staff members they depend on in their everyday lives. This article analyzes a national case study in which these contradictory tendencies are especially explicit: formal influence channels for older people in Sweden. Using data from structured interviews with 100 representatives of Swedish municipalities and drawing on Hirschman's (1970) Hirschman, A. 1970. Exit, voice, and loyalty: Responses to decline in firms, organisations, and states, Cambridge, MA: Harvard University Press.  [Google Scholar] theory on exit and voice, the article analyzes obstacles to older service users' influence in Sweden and develops explanations for these obstacles in terms of social contexts.     

The influence of top management corporate networks on CEO succession

We analyze empirically how supervisory board members with multiple directorships affect the decision to hire an inside or outside CEO successor. While a growing number of both theoretical and empirical studies analyze the influence of corporate performance and size or the ownership structure on this decision, the role of multiple board memberships within the CEO recruitment process has been widely neglected so far. The present study is based on panel data of the largest German companies covering the period from 1996 to 2008. Applying competing risk estimations we find a weak and positive association between the number of external directorships of the supervisory board members and internal CEO replacements. Distinguishing between different groups of external board positions, we find that external executives on the supervisory board increase the likelihood of external CEO replacements. In line with empirical findings for the US we argue that external executives improve the assessment of potential CEO successors leading to more outside CEO replacements. In contrast, we find evidence that external supervisory board mandates of the supervisory board members cause more internal CEO replacements. This finding indicates a substitution of external expertise of the executives by multiple supervisory board mandates but could also reflect a reduction of the monitoring intensity of the supervisory board.     

Cross-national policy learning in health system reform: The case of Diagnosis Related Groups

Observations of policy convergence and the cross-national diffusion of ideas, knowledge and policies have raised the question about the ways countries might learn from their peers. This article examines the role of cross-national learning with regard to Diagnosis Related Groups (DRGs). We review the spread of this policy instrument and analyse the implementation of DRGs in three late-adopting countries: Germany, Switzerland and the Netherlands. The three cases show that the implementation of this policy instrument required intense studies, cooperation with stakeholders and adjustment to country-specific needs. The countries learned from foreign experience, but it was only with the introduction of a regulatory framework for competition between sickness funds that DRGs came fully onto the political agenda. While Germany and Switzerland drew upon foreign DRG models, the Netherlands developed an alternative system to classify patients according to case-mix.