Methodology and epistemology of multilevel analysis (original) (raw)

Methodology and epistemolgy of multilevel analysis

The purpose of the multilevel approach is to understand individual behaviours taking account of the social context in which they occur. This book deals with concepts and methods underlying this approach. Its scope is therefore larger than solely statistical multilevel modelling, even though the latter enjoys a prominent place in the volume. As the subtitle says, the text considers, through examples drawn from different social sciences such as education, demography, epidemiology, human geography, economics, and so on, the challenges multilevel analysis permits us to answer and it also points out some limitations of these models. Simultaneously with spatial, organisational or institutional levels, it considers different time scales used in social sciences, and particularly the treatment of time in the history of economic thought. It also considers the more general philosophical and epistemological issues raised by their use: it shows that it no longer makes sense to choose between methodological holism and individualism, as multilevel analysis paves the way for a new approach in social sciences, studying how these different levels interconnect. This book is therefore of interest to a very wide audience of social scientists, statisticians and philosophers concerned with new issues raised by the multilevel approach, and more generally with explanation in the social sciences. Finally, it allows us to resume the theme of the first volume of this series on the explanatory power of models, offering a means of combining causal explanation and systemic explanation.

Methodology and Epistemology of Multilevel Analysis. Approaches from Different Social Sciences

Population (French Edition), 2003

Opposition between holism and individualism 2 How are the two approaches linked 3 A plurality of aggregation levels and a plurality of time scales 4 Towards a recomposition and a multilevel analysis 5 Outline of this volume Chapter 1 Multilevel modelling of educational data Harvey Goldstein 1 Fundamentals : units and levels 2 The basic multilevel model 3 Cross-classified models 4 The multiple membership model 5 Types of responses 6 Final thoughts about new insights Chapter 2 From the macro-micro opposition to multilevel analysis Daniel Courgeau 1 Introduction 2 The aggregate period approach 3 Cohort analysis 4 Event history analysis 5 Contextual and multilevel analysis 6 Conclusion Chapter 3 Potentialities and limitations of multilevel analysis in public health and epidemiology Ana V. Diez-Roux 1 Introduction 2 The presence of multiple levels : conceptual and methodological implications 3 Multilevel analysis 4 Multilevel analysis in public health and epidemiology 5 Challenges raised by the use of multilevel analysis in epidemiology 6 Limitations and complementary approaches Chapter 4 Exploring small area population structures with census data

Individuals and levels in the multilevel approach of behavioural analysis

Demography: Analysis end synthesis, 2006

Analysis of multistate data aggregated by socio-occupational group, geographical area, administrative unit, and so on, is based on the assumption that the behaviour of individuals is homogeneous in each of these sub-populations. If these conditions hold, it is sufficient simply to use the conventional demographic measures and rates and identify the relations that exist between them. For example, a study might relate the emigration rates for the various regions of a country to the unemployment rates, percentages of farmers, etc. observed in each region. But we must be careful not to infer that these relationships between aggregate-level values are also valid for the corresponding individual-level characteristics. That would be making the error of reasoning known as the ecological fallacy-a positive association between a region's emigration rate and the proportion of farmers in its population in fact tells us nothing about the migration probability of farmers, for whom the correlation may well be negative. Aggregate-level models can also be constructed in which fertility, mortality and interregional migrations are introduced simultaneously (see chapter 20 on r-level population models), under the additional assumption of independence between the phenomena being studied.

Multilevel theory and the underspecification of multilevel models

Quality and Quantity, 1992

This article emphasises the importance of the adequate specification of models of multilevel analysis in accordance with multilevel theories. Until recent times, multilevel theories tried only to explain the direct effect of group characteristics on an individual's characteristic. It seems to be more suited to adopt a more general theoretical approach, in which it is assumed that group characteristics affects individual processes. There a treshold effect and a process effect have to be distinguished. The propositions result in a model specification within the random coefficient model of multilevel analysis. The theory and model recommended are illustrated by means of data of Dar and Resh's (1986) study into social learning environment.

Multilevel Analysis in Critical Research

Annals of the International Communication Association, 1987

VER the past few years, the critical approach to communication has been increasingly recognized and used. In spite of the considerable differences that can be found within "the" critical approach, two common strands can be identified. First, the position of critique can be found in most critical approaches. Described most generally as "reflection on a system of constraints which are humanly produced" (Connerton, 1978, p. 18), the position of critique questions the Hnaturalness" of human experience. Critique seeks to undermine "the everyday attitude that identifies what one perceives with what is" (Appelbaum & Cotiner, 1979, p. 74) and brings to the surface the underlying processes and relationships that shape and govern the surface world. This concern with revealing deep structures and realities is combined in most critical approaches with a second common idea, namely, a conception of the social world that stresses the complexity and multiplicity of existing relationships. In emphasizing complex and multiple relationships, critical approaches seek to connect traditionally separate ideas and processes such as social structure and human agency, the material and the symbolic, and so forth. Grossberg (1984), in a recent review, has shown the many ways in which different critical approaches conceptualize the relationship between society and culture. Although the differences in some cases are quite substantial, all critical approaches maintain conceptually the idea of an inherent complexity in social relationships. Extremes of structural determinism and individualistic voluntarism are avoided typically by proposing some concept that stresses processes of mutual determination and influence thought to operate between human symbolic activity on the one hand and society or social structure on the other. Although ideas like the "duality of structure" (Giddens, 1979, 1984) or the "double articulation of social structure and human praxis" (Bhaskar,

Multi-level research into the social: An old wine in an old forgotten bottle?

2016

This essay provides a cursory discussion on the issue of how to conceptualise social phenomena in terms of the multiple levels that characterise it. It recognises the nested nature of different levels and the challenges of integrating or aligning different levels, whether it be from a qualitative of a quantitative perspective. In response it is proposed that a complexity approach, whereby the complexity of a situation is unfolded to reveal multiple levels offers a useful approach for handling multi-level perspectives. In doing so, it draws upon an old model which has perhaps been in the main, forgotten: the Viable System Model (VSM) developed by Stafford Beer Introduction It is not uncommon that research takes as its unit of analysis a social entity (e.g. individual) or assortment of similar entities (i.e. member of a specified population) in isolation or within ‘context’, and assumes a simple relationship between the entit-y(ies) and this context (Hitt, Beamish, Jackson & Mathieu, ...

Influential Cases in Multilevel Modeling: A Methodological Comment

American Sociological Review, 2010

A large number of cross-national survey datasets have become available in recent decades. Consequently, scholars frequently apply multilevel models to test hypotheses on both the individual and the country level. However, no currently available cross-national survey project covers more than 54 countries (GESIS 2009). Multilevel modeling therefore runs the risk that higher-level slope estimates (and the substantial conclusions drawn from