Abstracts, by session.

1A: Health Statistics                                                             

Biostatistics and Biopolitics: Modelling Population Health via Risk Assessment

Susanne Bauer

This paper follows the trajectory of an epidemiological database – from data collecting and numerical modelling to the moment these data translate into a software tool used in preventive medicine. Large-scale epidemiological studies, investigating an increasingly broad range of potential ‘risk factors’, were initiated in Denmark in the 1960s and 1970s and are continued until present. In particular in the Nordic countries with population registries maintained by governmental agencies, statistical patterns of health and disease can be evaluated through practices of record-linkage, a feature referred to as ‘unique resource’ for epidemiologists. While record-linkage performs population categories as potential co-determinants of health and disease, population health – with a host of contested categories – emerges as an epistemic object that is monitored under an extended epidemiological gaze. Bringing together data from different sources, such as registries, screenings or questionnaires, under the mediation of quasi-experimental study designs, epidemiologic research employs techniques of multivariate statistics which could become routine only due to the increase in computing capacity. The proliferating multivariate models not only inform the knowledge production, but also reconfigure public health interventions and policies as well as understandings of health and disease on a broad biopolitical level. I will describe how this plays out in an epidemiological study in Copenhagen – being a research-driven, in part private and state-supported endeavour that eventually produced a clinical software tool entering doctor-patient encounters. As boundary objects between research, public health policy and clinical practice, epidemiological risk estimates have come to play an important role in how health and disease are conceptualised, negotiated and acted upon on societal and individual levels.

What ecological networks and market based health programs have in common

Stans van Egmond & Ragna Zeiss

In this article we discuss in two case studies the process of model making as a political process. Models that are used for this specific purpose can be seen as hybrids of scientific standards and policy demands, and are political in that they (try to) depoliticize the policy problems for which they are used. Our examples show that the construction of model parameters is a ‘political’ process in which scientists and policymakers have to negotiate their own and the other’s standards. We will reflect ‘boundary work’ as a ‘tool’ to analyze the processes, negotiations, and model constructions, and the role of econometric models as boundary objects between scientific theory and policy practices, by comparing both cases.

The role of statistics in defining the efficacy of medicines: the trial of an herbal remedy in South Africa.

Julie Laplante

Science’s golden standard to obtain evidence based knowledge about the benefits of pharmaceuticals today relies upon the practice of randomized clinical trials (RCTS).  The summit of scientific objectivity is meant to be found through this empirical practice strongly endorsed by biomedicine, which in turn is strongly endorsed by humanitarian medicine. As has been shown several times, standards of objectivity nevertheless change through space and time.  Statistics have for example only progressively taken a central role in the embedded protocols of RCTS, especially since the mid 1960s.  The results of epidemiological studies based on statistics today carry strong biopolitical meanings and have become essential in the enactment of dominating relations of power.  Such trials of pharmaceuticals are today conducted in diverse ‘traditional’ settings making obvious how this epistemology and inherent methodology is culturally specific.  Clinical trials of ‘traditional’ herbal remedies are also being conducted and bring about other challenges as to how to define the healing benefits of medicines.   In this communication I would like to specifically examine how statistics have become contingent to knowledge about the efficacy of medicines, a form of knowledge somewhat beyond science, the state and the people.

“The Society of Disease”: Medical science as co-construction of nature and society - Three accounts of epidemic disease from the 19th century.

Jan Groven Grande

In this paper I will describe and analyze three medical accounts of epidemic disease in Norway during the 19^th century. These epidemic accounts take place when the established history of medicine maintains that the traditional medical practise of curing was supplemented by the biological study of disease in laboratories and the statistical study of the incidence of disease in society. The first account is an official report to the authorities and to the medical community on a cholera epidemic in a rather densely populated city, the Norwegian capital Christiania. The second account is a scientific article written by one of the regional officers of health shortly afterwards he had tried to manage an epidemic of dysentery in a sparsely populated area. The third account is the medical scientist Armauer Hansen’s own report on how he in his laboratory discovered the micro-organism that caused leprosy.

Today, both medical science and most social studies on medical knowledge find it irrational to question the scientific descriptions of the biological nature of disease. Likewise, statistics is made the preferred method of describing the incidence of disease in society. In this paper, I will present an alternative perspective on these three medical accounts: Contrary to traditional descriptions of scientific progress as processes of discovery of natural entities, I will focus on the production of the biological nature of epidemic disease as the construction of heterogeneous relations between medical theories of disease and empirical studies of disease both in the society, in the clinics and in the laboratory, statistical knowledge on the course of specific epidemics, and also as the construction of new relations between nature and the modern society of disease.

1B: Recognition - Redistribution                                           

What kind of Sámi figures? A discussion on status and future prospects for contemporary Sámi statistics

Torunn Pettersen

The overall knowledge about the indigenous Sámi people in the northernmost Europe is widely recognized as weak and rather characterized by stereotypes. One reason for this is the quite limited kinds and amount of data available for research and information. For instance, numeric data collected to produce official statistics on Sámi demography and Sámi social conditions have "always" been insufficient, often totally lacking.

The West-European states, included the three Nordic states that divide the traditional Sámi living area (Sápmi), are not among those states who have long, though disputed, traditions for official recording of ethnicity data. Today there are neither clear criteria nor institutional procedures for dealing with contemporary official Sámi statistics. A particular challenge is that such statistics to some extent must be based on ethnicity data about individuals' Sami affiliation (ancestry and/or identity), while relevant stakeholders (institutions and individuals) have different opinions on both whether and eventually how such data should be officially recorded.

Political changes during the last decades, resulting in increased Sámi rights to influence and self-determination, have re-actualized various stakeholders' needs for access to Sámi statistics for documentation, society planning, political decisions and evaluations. This challenges established norms on not recording Sámi/ethnical data for official statistics in the respective states.

How do we define for whom the bell curves? Measuring ethnicity in Canada

Ellen Balka

The Canadian Health Act states that healthcare must be delivered to Canadians in an accessible, equitable and universal fashion. Canada is an ethically diverse society, and in practical terms, determining whether or not the principles of the Canada Health Act are being upheld in its implementation requires an analysis of health service delivery in relation to ethnicity. While for policy analysts assessing whether or not Canadian health services are delivered in a fair and equitable matter is a manner of equity, for countless health services researchers, evaluating health outcomes and social determinants of health in relation to ethnicity is a matter of delivering effective care.

Health researchers and health care practitioners interested in diversity face many challenges. Often health researchers are interested in ethnicity because they have observed differences in how members of varied ethnic groups respond to health interventions, or because they have observed that some populations appear to be more susceptible to certain diseases than other populations. While it may be possible to document the increased prevalence of a disease among a particular ethnic group, even such seemingly simple assertions raise many questions. For example, how should a particular ethnic group be defined? How might researchers capture the complex interplay of cultural factors (such as belief systems), region of origin (e.g., South Asia), cultural practices (e.g., common diet), and social determinants of health (e.g., socio-economic status, length of time since immigration into Canada, extent to which one is integrated into the dominant culture, etc.), and other variables such as the relative status of various “ethnic” populations in relation to the dominant culture(s) in a region (e.g., a French as first language Haitian may have greater social status in Quebec where French is the dominant first language, than the same person would in British Columbia)?

Poverty measures as boundary objects between science and state:

A business approach to users and producers of poverty statistics.

Julia Mensink

As poverty affects societies in detrimental ways, poverty statistics have been of major interest to economic science and the state. As such, poverty measures have acted as facts that travel between economic science and the state. This paper analyses poverty statistics in these domains, which differ considerably. During the twentieth century, economists have employed increasingly sophisticated quantitative and statistical approaches, while politicians have preferred more intuitive and policy oriented measures. Moreover, the aims and usage of one and the same measure have varied greatly between the user and producer domains. This paper investigates these understudied but essential differences between users and producers of poverty measures.

The innovative methodology of the paper introduces conceptual analogies from business research concepts to analyse poverty statistics and their impact. Economic measures have not just been there, they have been produced. Actors and processes, such as ‘investors’, ‘raw materials’, ‘manufacturing processes’, ‘marketing’, and ‘stakeholders’, have all determined the end product: the quantitative poverty statistics. These statistics then have had an impact on governance and eventually on lives of citizens. Using case studies this paper reveals how the production and usage of poverty measures interconnects the academic and policy domains. With a focus on the practice of the production of the statistics and the practice of governance, the paper analyses how poverty statistics act as boundary objects between science and state.

Ideas in Action: 'Human Development' and 'Capability' as Boundary Objects

Asuncion Lera St. Clair

Ideas are a driving force in the current political debate as to what appropriate development policy is and how to conceptualize poverty. Ideas in development are concepts which powerfully influence development policy, yet they are often elaborated by the UN System itself. Following insights from sociology of science and science and technology studies (STS), the paper proposes that ideas and theories on poverty and development are neither scientific truths nor knowledge separated from the realm of politics.  UN System institutions are hybrid sites of co-production; they produce knowledge and politics, global ideas and global social orders. The paper argues that ideas in the multilateral system straddle a shifting divide between science in the making and global politics in the making; ideas are subjected to what social studies of science refers to as  "boundary work;" a constant redefinition of what is knowledge and what is non-knowledge.  The paper elaborates on the history and use of the notions ‘human development’ and ‘capability’ and depicts them as ‘intellectual boundary objects’, and the Human Development Report Office as a ‘boundary organization.’ These intellectual boundary objects and the boundary organization are attempts to stabilize the tensions that arise between the realms of knowledge and politics.

2A: Administration and Performance indicators                  

Mediating and Translating Materialities of Governance - Performance Indicators as ‘Investments in New Forms’ of Governance?

Sonja Jerak-Zuiderent, Roland Bal, Christopher Pollitt, Stephen Harrison & George Dowswell

The introduction of performance indicators (PIs) in the Dutch health care system in 2003 has mainly triggered discussions between protagonists and antagonists of ‘new public management’. In contrast, this research focuses on the empirical analysis of what these indicators actually “do” in health care organisations and systems in general and in hospitals in particular. The central questions posed in this research are (1) how does this ‘investment in (new) forms’  (Thévenot 1984) - by developing and introducing PIs -interfere with old forms of complex (public) ecologies of knowledges (Star 1995) and practices and (2) if and in what ways PIs, initially meant for the public accountability of hospitals, are taken up by the hospitals and ‘publics’ for improving the quality of health care.

In this paper I will present the first results of the empirical and qualitative research into the consequences of PIs in the Dutch healthcare system. The findings are based on (1) interviews with key actors who have been involved in developing the set of now 29 PIs for/with the Dutch Health Inspectorate and introducing them in the Dutch health care system and (2) on participatory observations in three hospitals.

Health System Indicators as Boundary Objects: Statistics and the State in Canadian Healthcare

Ellen Balka

In Canada, a basic level of healthcare is provided to all Canadians and landed immigrants. The right to healthcare is enshrined in the Canada Health Act, which includes as two of its five principles that care should be accessible to all Canadians, and that it should be universally available.

In an era of increased accountability, fuelled by the convergence of computers and telecommunication equipment that has made it at least theoretically possible to collect vast amounts of statistical data, the Canadian state has increasingly invested in infrastructure designed to support increased collection of data in the health sector. Several things have fuelled recent interest in health data in Canada. These have included an emphasis on evidence based medicine (EBM), and fiscal pressures which are fuelling a need for data upon which to base decisions about the allocation of health resources. In an era of increased accountability, health data can be used to assess the extent to which the Canadian state is meeting its commitments to universal and accessible healthcare, as outlined in the Canada Health Act.

Technologies of governing and processes of self-governing - The use of statistics in the quality assessment of Norwegian schools

Sigrunn Tvedten & Svein Hammer

Skoleporten.no is a website administrated by The Norwegian Directorate of Education and Training, and launched in 2004 as part of the development of a national quality assessment system for Norwegian primary and secondary schools. The main content of the website are statistical data from schools, for example the results from the national tests. The website also offers resources to contribute to interpretation, assessment and development within the school. Teachers, principals, and local school administrators and politicians are encouraged to use the information as resources for local work concerning evaluation and development.

Our paper will analyse this website through the perspective of governmentality, that is, a focus on 1) the relations between concrete technologies of governing (the website itself, the data reported, the procedures for production and use of the numbers, etc) and related, but more abstract governing rationalities (processes where instrumental attempts to formalise, standardise and quantify, liberal attempts to govern through freedom, etc are questioned, conceptualised and discussed) – and 2) the relations between vertical systems of government (national and municipal laws, structures and procedures constructed to govern the school) and horizontal practices of subjectivation and self-governing (the way different actors find their positions, form their identities, make their choices and assess their results).

The empirical discussions along these two dimensions, will probably indicate some important adjustments in how the Norwegian school system functions; a discursive change where ambitious attempts on holistic, planned governing, have been at least partly replaced by a liberal focus on decentralised strategies, where the humans as creating entrepreneur, calculating economists and choosing customers are cultivated – or, in other words, an increasing tendency to operationalise the (seemingly) free choices of humans. Our main focus here will be on indicators and numbers on the one hand, and different processes of self-governing on the other.

Quantitative basis for decision-making – boundary objects in the arena of Swedish road and railroad planning?

Karin Thoresson

The paper elaborates on quantitative basis for decision-making in Swedish infrastructural planning using a boundary object perspective. The main issue is to investigate whether this quantitative material plays the role of a boundary object between different social worlds involved in the process of planning new roads and railroads.

This study is part of an ongoing PhD project at the Swedish National Road and Transport Research Institute. The study is to make a contribution of knowledge concerning the use of quantitative basis for decision-making in infrastructural planning within the transport area. In the planning of new infrastructure economic calculations, models and forecasts over (social) costs and benefits are used as basis for investment decisions. The project started due to lack of insights regarding the role of calculations, models and forecasts in road and railroad planning practices and also due to earlier criticism towards these quantitative basis for decision-making. The main critical remarks have focused on a shortage in practical usability due to complexity and lack of transparency.

Actors involved in the planning process are primarily civil servants at the Swedish National Road and Railroad Administrations national and regional units. When planning new roads and railroads other actors such as the Ministry of Enterprise, Energy and Communications, the Swedish Government, politicians at the municipal level, and other local actors with different interests are also involved in the process. All these different actors can be regarded as part of the infrastructural planning arena.

The foundation of this paper consists of previous research, reports and inquiries within the area of interest. I examine the meaning and significance of quantitative basis for decision-making in negotiations, discussions and translations between actors and their possible function as boundary objects between different social worlds in the infrastructural planning arena.

Statistics and their relation to formalizing processes in social systems

Jon Hovland

Statistics, our understanding of them and their meaning, are in this papers point of view the sum of the central debates we enter in discussion of statistics. I argue we may narrow these down to four. A: Standardization / Categorisation, B: Quantification, C: Model Analysis and D: Structural formalizing. The first three are intuitive and well known, and I will briefly sum up the pro’s and con’s of these debates. D is more fluid, and in a way represents the way in which all three other debates both contribute to formalizing structures in social systems, and vice versa.

2B: Crime and statistics                                                          

Governance or government? Two faces of criminal justice statistics in nineteenth century England and Wales

Chris A. Williams

Interpretations of historical criminal statistics in the UK fall into several categories. They are variously seen as: fictional constructs (Taylor, 1998); starting points for the debate on criminality (Sindall, 1990); measures of the activity of the criminal justice system (Williams, 2000), or even as measures of crime (Gatrell, 1980). The bureaucratic processes that produced the statistics also allow us to examine the nature of nineteenth century governance.

In his work on the private sector, Beniger (1986) identified the 'crisis of control' as a motor for a change in administrative and managerial techniques, and it is possible to see a similar process operating in public sector crime control institutions (Williams, 2007). One way that this can be assessed is by studying the duration of the feedback loops between the collection of information and its processing as a part of the policy-making process, or its publication. A general characteristic of modernisation has been the acceleration of these feedback loops. This paper examines the various techniques used by the British government to process and publicise information concerning habitual criminals, in an attempt to see whether or not there is a secular trend towards faster feedback in this area.

It will also attempt to characterise the way that the British state reacted to and used expertise. Traditionally (Moran, 2003), historians have explained the government of the UK in terms of the action of a small clique of generalists, whose main task was to manage the appearance of government. A recent critique of this position has been advanced by Agar (2002), who argues that by the early twentieth century Whitehall was concerned to make use of information to best increase the effectiveness of government.

Agar, J. (2002) The Government Machine: A Revolutionary History of the Computer, London: MIT Press.

Beniger, J.R. (1986) The control revolution: technology and the economic origins of the information society, Cambridge: Harvard University Press.

Moran, M. (2003) The British Regulatory State: High Modernism and Hyper-Innovation, Oxford: Oxford University Press.

Gatrell, V.A.C., (1980) The decline of theft and violence in Victorian and Edwardian England, in Gatrell, V.A.C., Lenman, B., and Parker, G., (Eds.), Crime and the law: the social history of crime in early modern Europe, London, Europa.

Sindall, R., (1990) Street violence in the nineteenth century: media panic or real danger? Leicester: Leicester University Press.

Taylor, H., (1998) ‘Rationing crime: the political economy of criminal statistics since the 1850s’, in Economic History Review, Vol. 51, No. 3, 569-590.

Williams, C.A., (2000) ‘Counting crimes or counting people: some implications of mid-nineteenth century British police returns’ in Crime, Histoire & Sociétés/Crime, History & Societies. Vol. 4 no.2, 77-93.

Williams, C.A., (2007) 'Trains, planes, then automobiles: the advent of the police control room in the UK, 1907-1975' submitted to Technology and Culture, November 2006.

Declining Crime Rates – Good News or Bad News?

Heidi Mork Lomell

Since the 1960’s, most countries have experienced at times dramatic increases in the recorded crime rate. Although some, such as academics and the statistics bureaus, have been skeptical about both the validity and the reliability of the figures, pointing at alternative interpretations, new counting rules, redefinitions, etc., both politicians and the police have taken these figures at face value, and exploited them politically to ask for more police, harsher punishments etc. The rising official crime rates were used by politicians and police unions to produce a general perception that crime was ’getting out of control’.

The last decade, however, there have been two important changes in this pattern: First, criminal justice agencies have become increasingly subject to inspection, audit, evaluation, and other means of testing their ’performance’ against defined criteria. Second, we have seen a drop in the crime rate, not only in Norway, but in most part of the western world. Now, when the statistics show a decrease, some of the groups that took the increase at face value, such as the police union, question the statistics. Politicians aspiring to power also question the figures, while politicians in office, together with the heads of the police force, embrace the figures.

The crime drop debate illustrates how there is always both a methodological and a political discourse surrounding official statistics. It also shows how these two discourses at times overlap, that methodological concerns can be used politically. By looking at who accepts and who is skeptical about the figures one can get a better understanding of  the way in which the production and deployment of (crime) statistics is an important part of criminal-justice politics.

How to convince with (non-) statistics.

Johanne Yttri Dahl

DNA is based on probabilities, but is not always presented as such in courtrooms. Norwegian expert witnesses sometimes paint pictures to make the statistics more understandable to jurors. This paper will explore how DNA-evidence is presented in courts.

In defence of Numbers

Ann R Sætnan

2C: Co-production of Statistics and general perspectives    

Sociology in the making; The co-production of statistical sociology and society in context of a withdrawing modern episteme.

Christoffer Kullenberg

This paper will explore the establishment of the largest social scientific research institute in Sweden, the SOM institute, which was founded in 1986 in Göteborg. It highlights the intersections of epistemology related to the social sciences, with special focus on the use of statistics as a way of constructing positivities in order for the “social” and the “human” and “identities” to emerge as an empirical ground for knowledge, with the co-production argument proposed by Sheila Jasanoff (2004). There will thus be a joint theoretical foundation of Michel Foucault’s (1970) works on the episteme of the human and social sciences, and the contemporary co-productionist idiom. Together they will function as a way of understanding how society and the social sciences are interrelated through the use of statistical measurements, and what the consequences are for the domains of the political, subjectivity and knowledge.

The empirical material consists of an in-depth study of the report series publications of the SOM-institute. Special attention will be drawn to the use of standardized scales and measurement techniques, which were imported from other contemporary surveys, such as the World Value Survey and the Eurobarometer. This will both indicate the global aspects of social statistics, as well as the co-production of a modern episteme with society at large. Moreover, this paper will also draw attention to the study object of the social sciences, namely to the “statistical subjects” (respondents), which in order to be enrolled by the SOM-institute need to reply to the survey forms in concrete everyday life situations. As there is a general declining trend in response rates for the SOM institute, this phenomenon will be discussed in terms of how respondent agency may have epistemological consequences, thus re-shaping the use of both social theory and statistical methodology.

The statistical object in the digital age: Scientific classifications under disturbance

Christine Hanke

Drawing on my previous work on the constitution of the categories ‹race› and ‹sex› and their interconnection in anthropology around 1900, this paper focuses on recent anthropological research which uses statistics for the identification of body features, such as biometrics and sport anthropometry. I will especially draw on the continuities until present and ask for the actual normalising impact of these statistical practices for the ‹population› as well as for the individual.

The main focus of my paper will be the blurring of categories. As I have described in my research on the anthropological constitution of ‹race› and ‹sex›, statistically constituted categories are not only boundary objects in the sense of Star/Griesemer (1988), but they have peculiar effects for classification work in a more literal sense: By focussing on the boundaries of the categories it appears that ‹race› and ‹sex› are construed but at the same time their boundaries are blurred, with the statistical objects becoming part of a ‹deconstructive flow›. This can be understood – as I would propose – as an effect of the epistemic structure of the metric-statistical and the (precursor of the) digital. Taking visual statistics into account, this becomes explicit: In data visualisations boundaries between signals and noise, between data and disturbances remain undecidable, even if statistical or mathematical operations try to eliminate the noise or to configure the signals before visualising the data. This observation is closely linked to the ‹dubitative› status of the image (Lunenfeld 2000) in the ‹digital age› of computer processing.

Uncertainty: how do experts and policy makers deal with it?

Annick de Vries

Uncertainty is an issue of growing concern. Uncertainty assessments and risk calculations are becoming increasingly important and are, as a result, provided more and more by diverse expert organizations. The outcomes assist policy makers in decision making.

I look at both an economic and an environmental expert organization that deal with uncertainties in different ways; the presentation of uncertainty varies essentially. One organization turned out to produce extensive uncertainty information whereas the other seemed to prefer a selective and ‘educating’ way of presenting uncertainty. This implies that the former leaves less room for policy makers to interpret the uncertainty information themselves than the latter. How do experts decide their presentation form of uncertainties? And what is perceived as more useful for policy makers?

One of the most important findings is that experts negotiate about uncertainties in meetings and that uncertainties are structured according to three general heuristics. I describe the way in which this process takes place. I use the do-ability concept and articulation work with respect to the way an expert organization deals with uncertainties. I first state that the experts transform uncertainties into do-able uncertainties. Second, I argue that do-ability of uncertainties is achieved by a process of articulation. Finally, I claim that articulation takes place along three junctions, which are based on three general heuristics found in the empirics.

From the policy-makers’ perspective, it turned out that the ready-to-use pieces of uncertainty information were used more frequently. I explain this with the help of the political context and of the various (strategic) purposes of policy makers. Further, I argue that the main reason of this difference is the extent of alignment between the expert organization on the one hand and the main ministries on the other hand. I define four factors that enhance this alignment.

Comparability at the expense of sensitivity to national features? Standardization of social statistics in the EU.

Marja Alastalo

The understanding of the social characteristics of nation states is mostly based upon social statis­tics. In the globalizing world today the comparable information provided by social statistics has become increasingly important if the importance is measured by the growing numbers of social statistics produced by supranational organizations such as Eurostat, the OECD and the UN. Social statistics include statistics on population and its living conditions and such phenomena as poverty, inequality and immigration. The significance of social statistics lies in the fact that the information they provide impacts heavily on our concep­tions of social reality at national and international levels.

Standardization of statistical work is a key method in the production of comparable statistical data. This presentation will discuss the con­sequences of standardization to the contents and methods of social statistics in the EU. Is the comparability gained in this process at the expense of sensitivity for national features? The presentation will draw on Finnish data collected in the project on Harmonization of Social Statistics in the EU: Finland and Sweden as Contrast Cases.

Notes on a Rex

Nadji Aïssa Khefif

Pierre Bourdieu did point out a striking parallel between the powers generally granted to the statistical machine, that is the assemblage formed by statistical tools and their operators, and the powers that the linguist Émile Benveniste described as the attribute of an Indo-European social institution once known as the rex. The rex had the power of drawing the lines of a social world that would be perceived as ‘just’, in the both sense of ‘fair’ and ‘right’. It was a religious, rather than a political institution. Prolonging the line drawn by this intuition, this paper will attempt to demonstrate that it may well be the function of a rex that the statistical machine fulfills within and for modern society.