Content uploaded by Fritz Sager
Author content
All content in this area was uploaded by Fritz Sager on Oct 24, 2017
Content may be subject to copyright.
1
Habermas' Models of Decisionism, Technocracy, and
Pragmatism in Times of Governance
The Relationship of Public Administration, Politics, and
Science in the Alcohol Prevention Policies of the
Swiss Member States
in Public Administration Vol. 85(2) (2007): 429-447
final version before typesetting
Contact:
Prof. Dr. Fritz Sager
University of Bern
KPM Center for Public Management
Schanzeneckstrasse 1
P.O.Box
CH-3001 Bern
Switzerland
phone +41 31 631 32 85
fritz.sager@kpm.unibe.ch
2
Habermas' Models of Decisionism, Technocracy, and
Pragmatism in Times of Governance
The Relationship of Public Administration, Politics, and
Science in the Alcohol Prevention Policies of the
Swiss Member States
Abstract
The present article assesses the effects different constellations of science, administration,
and the political sphere display on the policy process in the field of alcohol prevention policy
in the Swiss member states. First, it is argued that Habermas' models of Decisionism, Tech-
nocracy, and Pragmatism can be interpreted as distinct modes of governance that take
place in today's policy-making and implementation. Second, as for the effects these different
constellations of science, administration, and the politics, the findings from a written sur-
vey of the Swiss cantons imply that a broad and adequate policy design can be found in
cantons with an alcohol prevention policy dominated by administrative actors. Output per-
formance proves to be comparatively higher in cantons with a high influence either of sci-
entific actors or of political actors. In an overall comparison a combined model of adminis-
trative dominance with strong affiliation to the scientific community prevails against the
other models considered.
Keywords: Jürgen Habermas; relation of public administration, science, and the political
sphere; technocracy; alcohol prevention policy; democratic legitimacy of public action
3
Habermas' Models of Decisionism, Technocracy, and
Pragmatism in Times of Governance
The Relationship of Public Administration, Politics, and
Science in the Alcohol Prevention Policies of the
Swiss Member States
Fritz Sager
INTRODUCTION
In 1963, the sociologist Jürgen Habermas came up with three models of how the two social
subsystems of science and politics work together. In his article "The Scientization of Politics
and Public Opinion" (orig. "Verwissenschaftlichte Politik und öffentliche Meinung"), Haber-
mas (1976: 120-145) developed the idea that a political system's response to the growing
complexity of modern society can be threefold: either scientifically little informed politics
dominate a professional administration, or a highly specialized and scientifically well in-
formed administration dominates the politics, or, finally, both work together in a pragmatic
and non-hierarchic way in order to find the best solutions for the social problems at stake.
The present article has a double aim: First, it is argued that Habermas' models can be inter-
preted as distinct modes of governance that are actually found in today's policy-making and
implementation. Second, as an exploratory effort, the article aims to assess the effects these
different constellations of science, administration, and the political sphere actually display in
the policy-process. For this goal, two measures for the quality of policy-making and imple-
mentation are chosen, namely the policy design and the output performance, and applied in
the field of alcohol prevention policy which is argued to be an adequate venue to test the
questions raised, for three reasons. First, the political field of alcohol prevention policy is
4
regulated with singular inconsistency, since the cantons, i.e. the middle federal rung in the
Swiss state structure, are equipped with very extensive programming and implementation
competencies, leading to a pronounced heterogeneity and correspondingly interpretable sta-
tistic variance in cantonal prevention efforts (Sager 2003). Second, alcohol prevention as a
public health policy is characterized by the need to meet the standard of distributive equity
since it has a very heterogeneous target group ranging from the general public to the actual
alcohol addicts, so that cantonal differences in performance efficiency are not only of theoret-
ical, but also of immediate practical significance (Cattacin/Lucas 1999; Meier 1994). Third
and finally, studies in the area of addiction policy in particular have proven the positive influ-
ence of scientific-professional administrative actors on the innovative ability of public action.
Thus, Sabet and Klingner (1993: 263) find "that a high degree of professionalism among
those administrators responsible for policy adoption will increase the likelihood that public
agencies will adopt innovative administrative policies, and those policies probably will be
more radical than incremental."
The article proceeds as follows: In the next section, a refinement of Habermas' models in the
light of governance theory is presented. This being done, both the policy design and the pro-
vision of the respective services are introduced as adequate measures for the assessment of
the theoretical models in question and eight respective hypotheses are formulated. The re-
search design in which the Swiss member states are shown to be an appropriate test for the
theoretical models, the operationalization of the variables, and the methodology are present-
ed in the section that follows. Subsequently, the results will be reported and the hypotheses
will be tested. The conclusion will consist of a final analysis summarizing both the theoretical
arguments and the empirical findings.
A REFINEMENT OF THE HABERMAS TYPOLOGY IN VIEW OF GOVERNANCE
THEORY
Jürgen Habermas (1976: 120-127) articulates three possibilities for the two sub-systems of
science and politics to deal with each other: decisionistically, technocratically or pragmatical-
5
ly. In the decisionist model, politics dominate science. Politicians use the results of research
selectively and ad hoc in pursuit of their own interests, while scientific discoveries and their
political consequences hardly gain a hearing in public and in the political decision-making
process. It is this model that can be considered compliant with what is labeled classic gov-
ernment structures as opposed to new modes of governance which we will discuss more in-
depth below.
The counterpart to the decisionist model is the technocratic model. Particularly at times
when new problems arrive on the political agenda, parties and associations initially turn to
science for information in order to get their bearings. At this early stage, the political actors
have not yet determined their positions, which gives experts with the relevant knowledge the
opportunity to influence key political actors’ opinions and increases the consideration given
to consolidated scientific findings in political practice. With its results, science can anticipate
political decisions, and in extreme cases even replace them. Freiburghaus and Zimmermann
(1985) conclude the existence of typical “neuralgic points” of high contingency at which sci-
entific findings become politically effective during the individual phases of the political deci-
sion-making process and via certain channels.
In the third variant, the pragmatic model, science and politics mutually complement each
other and no hierarchical relationship exists. It is this latter model that at a first glance would
be considered closest to what in recent years has been labeled 'Governance'.
"In the most common current usage of the term, 'Governance' is seen as implying a
move away from the previous government approach (a top-down legislative approach
which attempts to regulate the behaviour of people and institutions in quite detailed
and compartmentalised ways) to governance (which attempts to set the parameters of
the system within which people and institutions behave so that self-regulation
achieves the desired outcomes), or put more simply, the replacement of traditional
'powers over' with contextual 'powers to' (Pierre and Peters 2000). In such a govern-
ance system, permeable and flexible system boundaries will facilitate communication
and will support the achievement of higher level goals. These assumptions underline
6
the switch from government to governance in debates about the modernisation of pol-
icy systems implying a switch from constraining to enabling types of policy or regula-
tion" (Lyall and Tait 2005: 3f.).
However, the term 'Governance' is used in many different ways. If, on the one hand, the non-
hierarchic levelness is emphasized as the main characteristic of governance structures, it is in
fact Habermas' pragmatic model that comes closest. On the other hand, if we stress the un-
derstanding of governance structures as networks of various types of public and private ac-
tors addressing the same problem, the concept applies very well to the technocratic model as
well in that it focuses on the various types of actors with differing degrees of democratic legit-
imacy that take part in the public policy-making. As a matter of fact, many of the critiques
which the concept of Governance is subject to are very similar to Habermas' democratic con-
cerns regarding the technocratic model. Wälti, Kübler and Papadopoulos (2004: 84) distin-
guish two respective strands of concern about new modes of governance:
"Deliberative criticism argues that governance extracts policy issues from the public
space and confines them into spheres of technocratic decision-making in which the
participants are no longer accountable, neither to the public nor to their representa-
tives (in particular Hunold; Duran; Gaudin 1999; Scharpf 1999; Bovens; Benz 1998).
Therefore, decisions tend to represent partial agreements among (nonelected) mem-
bers of specific policy networks. Participatory criticism stresses that governance may
limit citizen participation and, in effect, impinge on mechanisms of community-
building and solidarity (in particular Duran and Thoenig; Rose)."
Habermas makes no actual distinction between the technical expertise centered within the
public administration and scientific expertise from the outside. In the 1960s, when the article
was written, there was no special need for this sort of distinction as the public administration
was considered the main resort for technical expertise relevant to politics (Hall 1968; Scott
1965). However, in light of governance theory, this distinction is important as professional-
ism no longer is considered the main trait of public administration. On the contrary, in times
of insinuated state failure (Niskanen 1971), the need of professionalism is considered one im-
7
portant reason for outsourcing public tasks (Ostrom, Bish, Ostrom 1988). In his analysis of
the NIMBY ('not in my backyard')-phenomenon, Fischer (1990) makes an according distinc-
tion between the state's right to define public objectives and policy experts' right to define
public problems, both of which are negated by NIMBY-groups. Understanding public admin-
istration as a self-reliant stakeholder in the policy-process, we propose, thus, a respective di-
vision of Habermas' technocratic model into a state-centered technocratic model in which
public administration pursues goals that are mainly problem driven which, however, does not
mean that they are automatically also scientifically informed (Lodge/Hood 2003), and a sci-
entific-technocratic model where it actually is the policy experts that display a major role in
the policy-making.
This distinction implies that bureaucratic interests do not necessarily concur with profes-
sional interests. This view does not conflict with the assumption that both bureaucrats and
scientific professionals aim at the best possible solution for a certain problem. However, the
assumption that "professionals with recognized expertise and competence in a particular
domain and an authoritative claim to policy relevant knowledge within that domain or issue-
area" (Haas 1992: 3) would draw similar interpretations and make similar policy conclusions
confronted with the same data is not adhered for public servants with this distinction. The
more expertise is outsourced and the more bureaucrats are to become managers (Aucoin
1990), the more the perception of what is needed to achieve the best solution for a problem
changes. The bureaucratic perception unlike the merely scientific one may, thus, emphasize
conditions beyond scientific evidence such as former solutions (Kay 2005), political windows
of opportunity (Olsen 2001) or budget concerns (Niskanen 1971). Still, both the administra-
tive and the scientific perceptions have to be labelled technocratic in that their aim to achieve
the best solution is not power driven but problem driven.
The corresponding refinement of Habermas’ models for the reciprocal relationship of science
and politics, with the additional distinction between administrative actors and professional
experts, leads to the emergence of the four basis constellations as presented in Table 1.
8
Table 1: Four constellations of the relation of politics, administration and science
Empirically, the question arises which of the four basis constellations predominates in the
cantons and what the political consequences of the individual models are. Bressers and
O'Toole (1998) state that the two concepts of policy networks, as governance structures, and
policy instruments have, for a long time, been considered in isolation from each other in the
relevant literature. The authors put forward the hypothesis that considering the attributes of
different governance structures contributes to the explanation of policy instrument choice. In
the following section, therefore, we will introduce two respective dependent variables in or-
der to assess the four models and derive a set of respective hypotheses.
DEPENDENT VARIABLES AND HYPOTHESES
As Stoker (1998) argues that governance is ultimately concerned with creating the conditions
for ordered rule and collective action, its outputs are supposedly no different from those of
government. If we, thus, want to assess the effects of this "difference in processes", which Ly-
all and Tait (2005: 4) see as the actual significant feature of governance, we have to focus on
the output of these differently structured processes. In order to do so, we select two central
variables from the policy formulation and implementation process as dependent variables:
policy design and degree of goal attainment.
1.) Policy Design. The central feature of a cantonal alcohol prevention policy is the state's
provision of services discernible by the policy addressees, i.e. the policy design’s choice of in-
struments (Vedung 1998). Empirically, performance evaluations within and outside health
policy show that broad policy designs, above all, have a high chance of success (Howlett
2005, Sager 2005a). This applies particularly to alcohol prevention policy: “No strategy con-
stitutes a panacea. Alcohol problems have multiple causes, develop in many different situa-
tions and affect the most dissimilar people. The requisite strategies will thus inevitably be
based on combinations rather than on a single stroke of genius” (Edwards et al. 1997: 184;
also Babor et al. 2003). The essence of the matter, according to the authors, is to determine
9
to what extent which combinations may be successful. Performance studies show that combi-
nations of variously obliging instruments are the most successful in terms of behavioral
change and acceptance among political addressees (Howlett 2005, van der Doelen 1998). As
our first dependent variable, we select the mean combination in policy design, while defining
an as-broad-as-possible policy design as desirable.
2. Degree of Goal Attainment in Terms of Production and Provision of Public Services. As
our second dependent variable, we select the degree of goal attainment in the form of out-
puts, i.e. the provision of public services, the production of incentives and the enforcement of
regulations. Outputs as implemented policy design are characteristic for their creation of a
direct relationship between political addressees and the responsible state implementation
authority. That is, output is not to be confused with the actual outcome or the impact of a giv-
en policy (Patton 1997: 193f.). To give an example, in our case of alcohol policy an output
would be the provision of a certain therapy offer but not its therapeutic success. It was not
possible to collect data on the actual effects of these outputs, i.e. the decrease in number of
addicts. In cases where not all political addressees can be tended to or prosecuted punctually
and simultaneously by outputs, Knoepfel and Bussmann (1997: 72f.) speak of temporally or
locally occurring implementation deficits. Implementation deficits are thus measured accord-
ing to the formulated, operative objectives of policy design. Low implementation deficits de-
fine the optimum in goal attainment.
The two dependent variables interact in that a sound policy design by itself is no use if not
implemented and a successful output production will not solve a social problem if the imple-
mented policy design is not adequate. Thus, both are needed in order to achieve an appropri-
ate and effective alcohol prevention policy.
The two dependent variables are now being linked to the four models derived from Habermas
(1976). The two variables, first of all, serve for the systematic comparison of the four models.
However, there also are theoretical reasons for linking the two dependent variables to the
models. The respective eight hypotheses regarding the relationship of science, administration
10
and politics in order to assess our four theoretical models are presented and explained in the
following.
(H1a) Where the decisionist model predominates, the policy design of a canton’s alcohol pre-
vention is rather broad.
(H1b) Where the decisionist model predominates, the output goals of a canton’s alcohol pre-
vention are likely to be attained.
The reasoning behind the first two hypotheses is based on the public choice argument that
the actors in the dominant political sphere seek to comply with voters' preferences due to the
need to be reelected after the end of a legislative term (Downs 1957). The evaluation of a poli-
tician's performance by the electorate can be expected to be drawn upon his or her success in
solving a social problem at stake. Considering that a broad policy design is more likely to
achieve a desired impact, politicians are expected to make respective decisions regarding in-
strument choice in alcohol policy and to provide sufficient resources for the implementation
of such policy designs.
(H2a) Where the pragmatic model predominates, the policy design of a canton’s alcohol pre-
vention is rather broad.
(H2b) Where the pragmatic model predominates, the output goals of a canton’s alcohol pre-
vention are likely to be attained.
The argument for the second two hypotheses derives from the proposition put forward in
governance literature that the deployment of so-called ‘command-and-control’ regulatory
policy instruments is the quintessence of government (Pierre 2000: 242) while "‘new’ (or
what are sometimes labelled ‘softer’) instruments" are characteristic for governance struc-
tures that "allow social actors more freedom to coordinate amongst themselves in pursuit of
societal goals" (Jordan, Wurzel, and Zito 2005: 478-479). As the pragmatic model consists of
the nonhierarchical relationship between societal actors in the form of science and govern-
11
ment in the form of political executives, we expect a mix of both types of policy instruments,
i.e. a broad policy design, as well as the provision of sufficient resources for implementation
to result from such configurations.
(H3a) Where the administrative-technocratic model predominates, the policy design of a
canton’s alcohol prevention is rather broad.
(H3b) Where the administrative-technocratic model predominates, the output goals of a can-
ton’s alcohol prevention are likely to be attained.
The argument behind the third two hypotheses is that public administration, as the dominant
actor, will seek the technically best policy design, i.e. the solution most adequate to the prob-
lem at stake. There are two lines of reasoning why a dominant public administration is able
to find resources sufficient for the realization of such solutions. The first one is based on to
the public choice perception of public servants as budget maximizers. Due to the information
asymmetry problem (Niskanen 1971), i.e. the "loss of the symmetry of helplessness" (Haber-
mas 1976), public agents will be able to convince politicians to provide the necessary funds
for implementation. Second, based on the view of public servants as managers, O'Toole and
Meier (1999, also Nicholson-Crotty and O'Toole 2004) show that managerial activities can be
substantive contributors to organizational success. Thus, the stronger the public administra-
tion's position is, the more there is room for managerial activities and the better are the
chances for successful implementation performance.
(H4a) Where the scientific-technocratic model predominates, the policy design of a canton’s
alcohol prevention is rather broad.
(H4b) Where the scientific-technocratic model predominates, the output goals of a canton’s
alcohol prevention are likely to be attained.
The argument behind the last two hypotheses is in line with the one behind hypotheses H3a
and H3b as far as the breadth of policy design is concerned. The professional interest pur-
12
sued by scientific and specialist actors in corresponding non governmental organizations is
expected to lead to an evidence-based policy design. The argument for the assumption of a
better goal attainment in terms of output performance is twofold: First, we already have de-
fined the scientific-technocratic model as a distinctive form of governance. Stoker (1998: 17)
states that the "essence of governance is its focus on governing mechanisms which do not rest
on recourse to the authority and sanctions of government." If, thus, the experts' dominance
does not only regard policy formulation but also implementation, we expect the specialists
outside the public administration to design their policy according to their possibilities for
implementation. Second, this scope of action will be augmented as its funds are not exclu-
sively provided by the state, but draws from other sources as well, such as donations and the
respective organizations' funds.
RESEARCH DESIGN, OPERATIONALIZATION, AND DATA
Our research approach is a quantitative comparison of cantonal decision-making and imple-
mentation structures in the field of alcohol prevention. Due to its 26 highly autonomous, but
also very different cantons and half-cantons in terms of their contextual features and internal
administrative organization, Switzerland represents an ideal field of investigation for system-
atic-comparative enquiries into the conditions of emergence and means of functioning of dif-
ferent forms of authority and administrative organization. A further advantage of such a
cross-sectional analysis is the fact that the superordinate context for all the objects of enquiry
(cantons) remains constant.
As Germann (1998: 18) states, the Swiss state is very similar to the US-American liberal sys-
tem. Unlike its neighboring countries, where we find various traditions of strong states (Knill
1999; Raphael 2000), Switzerland does not have a tradition of a strong and autonomous bu-
reaucracy. On the contrary, local and cantonal government was based on self-administration
structures for a very long time. Another limitation to bureaucratic autonomy consists of di-
rect democracy which is especially well established at the cantonal level. These idiosyncrasies
of its political system make Switzerland an outstanding field of investigation for the analysis
13
of new modes of governance. For the afore-mentioned reasons, this is especially the case in
the field of alcohol prevention policy (Cattacin/Lucas 1999).
Empirically, the paper draws from a written survey of the Swiss Cantons (Sager and Vatter
2000). For the data collection, a standardized questionnaire was used to ask the appropriate
decision-making and implementation bodies in the Swiss cantons for information regarding
the contents and organization of alcohol prevention policy. The survey took place in summer
1999. The survey questions used were developed especially for the study. The questionnaire
was mostly filled in by the executive of the authority respectively responsible for alcohol pre-
vention. 21 of 26 cantons responded. The rate of return amounted to 81%. It was not possible
to examine the extent to which the self-declaration in the questionnaires reflects actual can-
tonal policy on the basis of the actual survey. In order to ensure validity of both the data and
especially the finding gained from them, additional in-depth case studies have been conduct-
ed in five cantons based on expert interviews and document analyses.
For the statistical analysis, the predominantly dichotomous variables were added in order to
transform them into ordinal scaled index variables. Even though it would have been desirable
to employ multivariate statistical analysis to get to the problem, we opted for a bivariate
analysis of the data. There are three reasons for this methodological decision. First, the ade-
quate methodology for ordinal dependent variables is the ordinal log-linear analysis. An or-
dinal logit requires a minimum of 50 cases to be calculated (Backhaus et al. 2000: 107). Ur-
ban (1993: 13) names a minimum of 100 observations for sufficiently meaningful estima-
tions. With our N=21, therefore, respective estimations would not lead to interpretable re-
sults. Second, a study of the determinants of the adoption of restrictive policy instruments in
the cantons comes to the conclusion that neither socioeconomic factors such as urbanisation,
religion, or even the actual alcohol consumption nor political factors such as party system,
strengths of parties in the executive, or the degree of direct democracy explain the coercive-
ness of policy design (Sager 2005b). These findings correspond to a large degree with the re-
sults of Meier's (1994: 216-225) analysis of U.S. alcohol policies. In contrast, policy specific
organisational factors regarding the responsible administration and implementation struc-
14
tures turn out to be good predictors of the choice of restrictive policy instruments in the can-
tons (Sager 2005b). We, therefore, consider it legitimate to concentrate on the bivariate cor-
relations between our four models and the dependent variables, not controlling for socioeco-
nomic and political variables. Third, the present article is an exploratory investigation and
does not contest its tentative character.
Correspondingly, we use contingency table analysis as our method for checking the hypothet-
ical associations. The correlation coefficient is Kendall tau-b. Kendall tau-b assumes a value
between 1 for a perfect positive correlation and –1 for a perfect negative correlation, with 0
representing no connection at all. In the following, we describe a correlation as important
whenever Kendall tau-b is greater than or equal to ¦.3¦ (Gehring and Weins 1998: 130). Since
the availability of data for 21 cantons corresponds to a quasi-complete inventory count, we
also include statistically non-significant values which only count towards the investigative
entirety. Because of the ordinal quality of the data and the rather small number of cases, we
dispense with multivariate model calculations. The individual results were qualitatively vali-
dated and supplemented by expert interviews and case studies.
As for the measurement of the four theoretical models, indices were formed from various
questionnaire items. Index construction in this process follows a textual motivation and con-
sistency. Thus, the variable “decisionism” consists of the items “Executive determines course
without consulting the administration”, “Parliament has decisive influence” and “Science
plays no role for politics”; the pragmatic model is measured via addition of the items “Execu-
tive determines course in close consultation with the administration”, “Parliament is influen-
tial in terms of budget, but not of contents” and “Science flows into politics with a delay”; and
the scientific-technocratic model with the items “We define our tasks ourselves”, “Parliament
has no influence” and “Science flows into politics rapidly”. For the administrative-
technocratic model, administrative self-perception is enlisted with the item “Administration
is, above all, a performance aid for politics and does not dispose of much scope for action” vs.
“The administration often dictates the line of attack to politics and disposes of a large scope
for action”. Accordingly, unlike the three other models, the administrative-technocratic mod-
15
el is dichotomously coded. This is due to the better possibility to capture the model in a single
survey question. However, this difference in item scaling does not affect the bivariate statisti-
cal analysis.
Table 2 reports the cantonal scores for the four models.
Table 2: Cantonal scores for the four models
Table 3 provides information on the inter-independence of the four models tested. It tran-
spires that no systematic associations exist between the four index variables formed.
Table 3: Correlations between the four index variables
The results in Table 3 provide an indication of the index variables’ validity. Thus the negative
correlation between the scientific-technocratic model and both the decisionist and the prag-
matic model, as well as that between pragmatism and administrative technocracy, speak for
the validity of the index formed. The weak positive correlation between administrative tech-
nocracy and decisionism does not fully reflect the two models’ theoretical distinctness, some-
thing which can be explained by the lack of an explicit item regarding the role of administra-
tion in the decisionism index. The positive correlations between scientific and administrative
technocracy, as well as between decisionism and pragmatism, are theoretically plausible and
again speak for the validity of the index. Based on these results, the independent interpreta-
tion of their correlations with the dependent variables can be considered permissible.
The first dependent variable, the policy design, then, was queried with the question “How
important are the following instruments in your canton’s current alcohol policy?”, in which
respondents were asked to rate a selection of 17 measures from very important to completely
unimportant on a scale of one to four. The breadth of policy design, was subsequently meas-
ured by means of an index in which estimations of the significance of the individually queried
16
policy instruments employed in a canton were added up and re-coded on a scale of 0-3, with
0 designating a very limited, 3 a very broad policy design.
The second dependent variable, goal attainment, finally, was queried in the questionnaire
through a self-assessment in which the respondents were asked to agree with the applicable
statement: “The canton is capable of attaining its self-set goals in alcohol prevention. True/
untrue”. Replies were subsequently appropriately dichotomously coded.
Upon being asked to what extent new scientific findings flow into political practice, 12 can-
tons (57.1%) state that this occurs only with large delays, while in another question, nine can-
tons (42.9%) state that scientific discoveries are, as a rule, of no significance whatsoever for
politicians. A rapid and intensive exchange between science and politics takes place in a mere
three cantons (14.3%). These initial results document the generally low ranking which re-
search holds for the political decision-makers in the cantons.
In considering the variance of the dependent variables, one is struck by the pronounced de-
gree of heterogeneity. Regarding policy design, both secondary prevention (advisory centers,
risk group programs and individual intervention) and treatment (advisory centers, out and
in-patient treatment) are described as important or very important by 20 cantons (95.2%).
Education and further education are mentioned second most often (18 cantons, 85.7%), pri-
mary prevention (general risk prophylaxis) and tertiary prevention (aftercare and survival
help) third most often (17 cantons each, 81.0%). 14 cantons (66.7%) describe the protection
of minors as a very or rather important component of their alcohol prevention policy. Beyond
this, the majority of the cantons evaluates availability control, i.e. regulation of opening
hours, physical availability in stores or catering businesses’ “necessity clause” (17 cantons,
81.0%), price regulation and specialized research (15 cantons each, 71.4%) as “rather or com-
pletely unimportant”. Thus, criminally binding instruments display a lesser importance than
instruments, which are less binding for the political addressee, such as information and in-
frastructure measures.
With a total of 11 mentions, cantonal alcohol prevention policies were described as successful
in around half of all the cases (52.4%). Simultaneously, the information additionally request-
17
ed contains a validation of the first dependent variable, in that in nine cases (42.9%) cantonal
alcohol policy is described as a success, which is attributed to an as-broad-as-possible combi-
nation of measures. Cantonal action programs are primarily and specifically aimed at the al-
cohol problem in only three cases (14.3%). Six cantons (28.6%) integrate their alcohol pre-
vention policy into general addiction prevention, three cantons (14.3%) into general health
promotion. Seven cantons (33.3%) state that they carry out both alcohol-specific and more
general measures.
HYPOTHESIS TEST AND DISCUSSION
Table 4 reports the correlations resulting from the analysis of the cantonal results.
Table 4: Hypothesis Test
A statistically significant correlation exists for hypothesis H4b, while hypotheses H3a and
H1b are merely not rejected for the cantons examined as tau-b is not statistically significant.
From this, we can derive the following statements:
• In the cantons examined, the more strongly the administrative-technocratic model pre-
dominates, the broader is the policy design of a canton’s alcohol prevention, thus the
broader is the range of different instruments it supplies.
• The more strongly the scientific-technocratic model predominates, the more likely a can-
ton’s alcohol prevention policy is to attain its goals.
• In the cantons examined, the more strongly the decisionist model predominates, the
more likely a canton’s alcohol prevention policy is to attain its goals.
The observations are interesting in terms of a comparison of the four models. A broad policy
design can be found in cantons where the administration dominates. On the other hand, po-
litical dominance, pragmatic consensus solutions and a strong leaning on scientific discover-
18
ies do not significantly influence the breadth of policy design. However, like political domi-
nance in the cantons observed, the latter constellation leads to a high output performance.
Neither the breadth of policy design nor goal attainment are, taken on their own, indicators
of an appropriate alcohol prevention. Only the enforcement of an adequate choice of means
implies an effective prevention policy. In keeping with this criterion, the combination of the
two technocratic arrangements scores highest marks, although this entails various democrat-
ic-theory problems regarding the legitimacy of political action. According to Habermas
(1973), if scientific-technical progress is elevated to being a system’s primary productive
force, the public becomes increasingly depoliticized and democratic formation of an informed
opinion is replaced by technocracy. Specifically this contradiction between formal democracy
and factual technocracy generates a chronic need for legitimation, which takes us back to the
respective discussion in the Governance-literature where Rhodes (1996) remarks that gov-
ernance poses a considerable "challenge for democratic accountability", and King (1996)
stresses that lines of accountability between voters and private organizations have become
somewhat obscure. "Critics also point out that, as governance is most concerned with in-
creasing the effectiveness and the efficiency of policy outputs, it may do so at the expense of
democratic legitimization on the input side" (Wälti et al. 2004: 84). So, why does technocracy
prove more successful than political dominance?
Based on expert interviews, the circumstance can be explained by means of two specific fea-
tures of prevention policy in general and of alcohol prevention in particular. For one thing,
the concept of prevention has a central flaw from the point of view of politics: it leads to no
short-term, concrete tangible results. Successfully altered addictive behavior expresses itself
in long-term tendencies, but is not suitable as manifest proof of the effects of political work.
Politics, however, depends on short-term successes and is hence reserved in its reaction. Sec-
ond, alcohol prevention is further disadvantaged in terms of its communicability compared to
illegal drug prevention, something which an interview quotation vividly illustrates: “The
sympathy bonus and, thus, acceptance of the use of tax payers’ money for addiction preven-
tion is incomparably higher in the case of a 20-year-old drug-addicted woman on the streets
19
than in the case of a 50-year-old, outcast and isolated alcoholic. That’s why illegal drugs are
politically much more popular than alcohol. The most alcohol prevention can hope for is to
bask in the illegal drugs’ reflected prominence.”
In other words: within the bounds of prevention policy, which is politically fairly unpopular,
alcohol prevention leads a shadowy existence. Hence, for obvious reasons, prudent alcohol
policies tend to be pursued in areas where decisions are not necessarily bound to political
considerations. At the same time, the results explain why not only scientific technocracies,
but politically dominated arrangements, too, are successful in terms of goal attainment, i.e.
why both hypotheses H1b and H4b are corroborated: scientists and specialists in the first,
political actors in the second case dispose of enough scope of action to perform their
measures – even if in the latter case it seems reasonable to suppose that political, rather than
specialist criteria determine the choice of these measures. These explanations exemplify the
significance of the dimensions of the contents of cantonal alcohol policies. Accordingly, the
question poses itself in which way the four models affect not only the breadth, but also the
contents of policy design, i.e. the choice of measures which are actually put into force. Table 5
presents the connections observed in this respect:
Table 5: Correlations between relation of science/ administration/ politics and alcohol pre-
vention measures
The bias of the results for the examined cantons confirms the above assumption that the de-
cisionist model tends to prioritize a concrete, tangible policy, while rather technocratic ar-
rangements are less considerate of short-term, more communicable political contents. Thus,
only the use of primary preventative measures (with the obvious exception of education) cor-
relates with the two technocratic models, while in the decisionist and pragmatic models, the
influence of politics manifests itself in a higher weighting of early diagnosis (secondary pre-
vention) and treatment. Last but not least, the use of the necessity clause and the enforce-
20
ment of the protection of minors correlate with the pragmatic model, something which can be
attributed to these measures’ high societal acceptance and obvious moral value.
Finally, in view of the four models’ repercussions, the question begs itself as to their basic in-
stitutional and organizational requirements. For this reason, and to conclude, we examine
the nature of the correlation between the four models and forms of institutional authority
arrangement, i.e. where the models are most likely to occur. This enquiry reveals the correla-
tions reported in Table 6.
Table 6: Correlations between relation of science / administration/ politics and authority
structures
As Table 6 shows, the observed connections are negative, which means that the decisionist
model is especially dominant in cantons where alcohol prevention policy is decentrally orga-
nized, while the administrative-technocratic model tends to dominate in cantons where alco-
hol prevention is not delegated to state service providers. These results are amenable to logi-
cal interpretation, since it is to be assumed that a fragmented administration is more easily
dominated by politics. Similarly obviously, the more technocratic an arrangement, the more
specialist capacities it possesses; and the more strongly this knowledge at the same time
dominates policy formulation, the less are tasks delegated.
CONCLUSION
Which relational models predominate between administration, politics and science, and what
role do they play regarding political practice? In order to answer these core questions, we
have derived four governance arrangements of science, politics and administration based on
the work of Jürgen Habermas: the decisionist model, in which politics dominates administra-
tion; the pragmatic model, which stands for a non-hierarchical relationship; the administra-
tive technocratic model, in which administrative actors dominate; and the scientific techno-
21
cratic model, in which experts dominate administration and administration dominates poli-
tics.
A comparison of cantonal alcohol prevention policies reveals that a broad policy design is ap-
plied in the cantons in which administrative actors dominate. On the other hand, political
dominance, pragmatic consensus solutions and a strong leaning on scientific findings do not
significantly influence the breadth of policy design. However, the latter constellation and po-
litical dominance lead to a high output performance. No systematic relationship exists be-
tween the pragmatic model and the political results. If we observe both dependent variables
together – which seems sensible in an assessment of cantonal alcohol prevention – the com-
bination of the two technocratic arrangements comes off best in the comparison undertaken.
The results suggest that professionally sensible, evidence-based policies tend to be practiced
wherever there is less need for decisions to be bound to short-term political considerations.
After all, alcohol prevention policy, as a field, is difficult to communicate and is politically less
utilizable than, for instance, restrictive measures in the field of illegal addiction. By the same
token, politically or professionally motivated political goals tend to be achieved whenever the
relevant authorities dispose not only of the power of definition in policy formulation, but also
of scope for action in terms of enforcement.
These findings contribute in two ways to the development of governance theory: First, while
Habermas' (1992) work on discourse theory holds a prominent place in the recent study of
politics (Holzinger 2001, Spörndli 2004, Steiner et al. 2004) also contributing to the research
on governance structures (Hunold 2001, Wälti et al. 2004), the present study introduces an-
other strand of argument drawn upon Habermas for the analysis of governance. While Wälti
et al. (2004) use Habermas' arguments to discuss governance-induced problems of democra-
cy, the work presented here applies Habermas in order to categorize specific modes of gov-
ernance and to assess the different models in comparison. The findings not only show that
Habermas' typology still has great analytical value decades after being put to paper, but also
that the present assessment of the models differs from the original evaluation. While Haber-
mas saw great democratic problems with technocracy, the present study finds virtues that
22
speak for a re-evaluation of bureaucracy, professional expertise and specialisation in times
state scepticism. These findings, as a second contribution to the theoretical debate on gov-
ernance, support Olsen's (2006: 2) cue "that for those interested in how contemporary public
administration is organized, functions, and changes, it is worthwhile to reconsider and redis-
cover bureaucracy as an administrative form, an analytical concept and a set of ideas and ob-
servation about public administration and formally organized institutions." The results pre-
sented here strongly corroborate this claim to shift the focus of governance analysis back to
questions of bureaucratic organization and success.
However, despite this contribution of the results to the debate on governance, it must be
pointed out that due to the ordinal quality of the data ascertained, the present investigation
provides only initial indications and cannot claim conclusive validity. This qualification needs
to be made not only because we have leaned exclusively on bivariate analyses, but also be-
cause of the small N of 21 cantons. Although the results point in the direction expected based
on theoretical considerations, further investigations are needed into the relationship between
politics, administration and science, based on broader sets of data which additionally permit
multivariate estimates. Only once intervening influences are systematically controlled will it
be possible to resolve these complex issues adequately. In this sense, this contribution is only
one step on a path requiring further research.
23
REFERENCES
Aucoin, P. 1990. 'Administrative Reform in Public Management: Paradigms, Principles, Par-
adoxes, and Pendulums', Governance, 3, 115-137.
Babor, T. F., R. Caetano, S. Casswell, G. Edwards, N. Giesbrecht, K. Graham, J. Grube, P.
Gruenewald, L. Hill, H. Holder, R. Homel, E. Österberg, J. Rehm, R. Room, and I. Rossow.
2003. Alcohol: No Ordinary Commodity— Research and Public Policy. Oxford and Lon-
don: Oxford University Press.
Backhaus, K., B. Erichson, W. Plinke, and R. Weiber. 2000. Multivariate Analysemethoden.
Eine anwendungsorientierte Einführung. Berlin: Springer.
Benz, A. 1998. 'Postparlamentarische Demokratie? Demokratische Legitimation im koopera-
tiven Staat', in M. T. Greven (ed). Demokratie—eine Kultur des Westens? Opladen: Leske
& Budrich, pp. 201–222.
Bovens, M.A. P. 1998. The Quest for Responsibility. Accountability and Citizenship in Com-
plex Organisations. Cambridge, New York: Cambridge University Press.
Bressers, H. T. A., and L. O'Toole Jr. 1998. 'The Selection of Policy Instruments: a Network-
based Perspective', Journal of Public Policy, 18(3), 213-239.
Cattacin, S., and B. Lucas. 1999. 'Autorégulation, intervention étatique, mise en réseau: les
transformations de l'État social en Europe (les cas du VIH/sida, de l'abus d'alcool et des
drogues illégales)', Revue française de science politique, 49, 379-98.
Downs, A. 1957. An Economic Theory of Democracy. New York: Harper and Brothers.
Duran, P. 1999. Penser l’action publique. Paris: Librairie générale du droit et de la jurispru-
dence.
Duran, P., and J.-C. Thoenig. 1996. 'L’Etat et la gestion publique territoriale', Revue fran-
çaise de science politique, 46, 580–623.
Edwards G. et al. 1997. Alkoholkonsum und Gemeinwohl: Strategien zur Reduzierung des
schädlichen Gebrauchs in der Bevölkerung. Stuttgart: Enke.
Fischer, F. 1990. Technocracy and the Politics of Expertise, Newbury Park CA.: Sage.
24
Freiburghaus D. and W. Zimmermann. 1985. Wie wird Forschung politisch relevant? Bern:
Haupt.
Gaudin, J.-P. 1999. Gouverner par contrat. Paris: Presses de science po.
Gehring U. and C. Weins. 1998. Grundkurs Statistik für Politologen. Opladen: Westdeut-
scher Verlag.
Germann, R. E. 1998. Öffentliche Verwaltung in der Schweiz, Bd. 1, Der Staatsapparat und
die Regierung. Bern/Stuttgart/Wien: Haupt.
Haas, P. M. 1992. 'Introduction: Epistemic Communities and International Policy Coordina-
tion', International Organization, 46, 1-35.
Habermas J. 1973. Legitimationsprobleme im Spätkapitalismus. Frankfurt am Main: Suhr-
kamp.
Habermas J. 1976. 'Verwissenschaftlichte Politik und öffentliche Meinung', in J. Habermas,
Technik und Wissenschaft als ‘Ideologie’. Frankfurt am Main: Suhrkamp, pp. 120-145.
Habermas, J. 1992. Faktizität und Geltung. Diskurstheorie des Rechts und des demokrati-
schen Rechtsstaats. Frankfurt a.M.: Suhrkamp.
Hall, R. H. 1968. 'Professionalization and Bureaucratization', American Sociological Review,
33, 92-104.
Holzinger, K. 2001. 'Verhandeln statt Argumentieren oder Verhandeln durch Argumentie-
ren? Eine empirische Analyse auf Basis der Sprechakttheorie', Politische Vierteljahres-
schrift, 42, 414-46.
Howlett, M. 2005. 'What Is a Policy Instrument? Policy Tools, Policy Mixes, and Policy Im-
plementation Styles', in P. Eliadis, M. M. Hills, and M. Howlett (eds), Designing Govern-
ment. From Instruments to Governance. Montreal&Kingston, London, Ithaca: McGill-
Queen's University Press, pp. 31-50.
Hunold, C. 2001. 'Corporatism, Pluralism, and Democracy: Toward a Deliberative Theory of
Bureaucratic Accountability', Governance, 14, 151–167.
25
Jordan, A., R. K. W. Wurzel, and A. Zito. 2002. 'The Rise of "New" Policy Instruments in
Comparative Perspective: Has Governance Eclipsed Government?', Political Studies, 53,
477–96
Kay, A. 2005. 'A Critique of the Use of Path Dependency in Policy Studies', Public Admin-
istration, 83(3), 553-571.
King, D. 1996. 'Conclusion', in D. King and G. Stoker (eds). Rethinking Local Democracy.
Government Beyond the Centre. Houndmills: Macmillan, pp. 214–23.
Knill, C. 1999. 'Explaining Cross-National Variance in Administrative Reform: Autonomous
versus Instrumental Bureaucracies', Journal of Public Policy 19(2): 113-39.
Knoepfel P. and W. Bussmann. 1997. ‘Die öffentliche Politik als Evaluationsobjekt’, in W.
Bussmann, U. Klöti and P. Knoepfel (eds), Einführung in die Politikevaluation. Ba-
sel/Frankfurt a. M.: Helbing und Lichtenhahn, pp. 58-77.
Lodge, M., and C. Hood 2003. 'Competency and Bureaucracy: Diffusion, Application and Ap-
propriate Response?', West European Politics, 26(3), 131-52.
Lyall, C., and J. Tait. 2005. 'Shifting Policy Debates and the Implications for Governance', in
C. Lyall and J. Tait (eds.), New Modes of Governance. Developing an Integrated Policy
Approach to Science, Technology, Risk, and the Environment. Aldershot: Ashgate, pp. 3-
17.
Meier, K. J. 1994. The Politics of Sin. Drugs, Alcohol, and Public Policy. Armonk/London:
M.E. Sharpe.
Nicholson-Crotty, S., and L. J. O'Toole Jr. 2004. 'Public Management and Organizational
Performance: The Case of Law Enforcement Agencies', Journal of Public Administration
Research and Theory, 14(1), 1-18.
Niskanen, W. A. 1971. Bureaucracy and Representative Government. Chicago/New York:
Aldine and Atherton.
26
O’Toole, L. J. Jr., and K. J. Meier. 1999. 'Modeling the impact of public management: Impli-
cations of structural context', Journal of Public Administration Research and Theory,
9(4), 505–26.
Olsen, J. P. 2001. 'Garbage cans, new institutionalism, and the study of politics', American
Political Science Review, 95, 191-8.
Olsen, J. P. 2006. 'Maybe It Is Time to Rediscover Bureaucracy', Journal of Public Admin-
istration Research and Theory, 16(1), 1-24.
Ostrom, V., R.L. Bish, and E. Ostrom. 1988. Local Government in the United States. San
Francisco: Institute for Contemporary Studies.
Patton, M. Q. 1997. Utilization-Focused Evaluation: The New Century Text. Thousand Oaks
CA, London, New Delhi: Sage.
Pierre, J. 2000. ‘Conclusions: Governance Beyond State Strength’, in J. Pierre (ed), Debating
Governance. Oxford: Oxford University Press, pp. 241–6.
Pierre, J., and B.G. Peters. 2000. Governance, Politics, and the State. New York: St. Martin’s
Press.
Raphael, L. 2000. Recht und Ordnung. Herrschaft durch Verwaltung im 19. Jahrhundert.
Frankfurt a. M.: Fischer.
Rhodes, R. A. W. 1996. 'The New Governance: Governing without Government', Political
Studies, 44, 652–67.
Rose, L. 1995. Consumers, Taxpayers, Citizens: An Inquiry into Democratic Citizenship in
the Context of Local Government. Oslo: University of Oslo, Department of Political Sci-
ence.
Sabet M. and D. Klingner. 1993. ‘Exploring the Impact of Professionalism on Administrative
Innovation’, Journal of Public Administration Research and Theory, 3(2), 252-266.
Sager, F. 2003. 'Kompensationsmöglichkeiten föderaler Vollzugsdefizite. Das Beispiel der
kantonalen Alkoholpräventionspolitiken', Swiss Political Science Review, 9(1), 309-333.
27
Sager, F. 2005a. 'Die Wirksamkeit verkehrspolitischer Massnahmen: Eine Meta-Analyse', in
T. Bieger, C. Laesser and R. Maggi (eds). Schweizerische Verkehrswirtschaft - Jahrbuch
2004/2005. St. Gallen: SVWG/IDT-HSG, pp. 213-227.
Sager, F. 2005b. Assessing the 'Exhaustion of Tools'-Phenomenon. Cantonal Alcohol Preven-
tion Policy Designs in Comparison. Paper presented at the 2005 ECPR Joint Sessions,
Granada, April 14-19 2005.
Sager, F., and A. Vatter. 2000. Strukturen und Strategien der Kantone in der Alkoholprä-
ventionspolitik. Eine komparative Analyse unter besonderer Berücksichtigung der Ver-
wendung des Alkoholzehntels. Bern: Studie zuhanden der Schweizerischen Stiftung für
Alkoholforschung.
Scharpf, F. 1999. Governing in Europe: Effective and Democratic? Oxford, New York: Ox-
ford University Press.
Scott, R. 1965. 'Reactions to Supervision in a Heterogeneous Professional Organization', Ad-
ministrative Science Quarterly, 10, 65-81.
Spörndli, M. 2004. Diskurs und Entscheidung. Eine empirische Analyse kommunikativen
Handelns im deutschen Vermittlungsausschuss. Wiesbaden: VS Verlag für Sozialwissen-
schaften.
Steiner, J., A. Bächtiger, M. Spörndli and M. R. Steenbergern. 2004. Deliberative Politics in
Action. Analysing Parliamentary Discourse. Cambridge: Cambridge University Press.
Stoker, G. 1998. 'Governance as theory: five propositions', International Social Science Jour-
nal 50: 17-28.
Urban, D. 1993. Logit-Analyse. Statistische Verfahren zur Analyse von Modellen mit quali-
tativen Response-Variablen. Stuttgart/Jena/New York: Gustav Fischer Verlag.
Van der Doelen, F. C. J. 1998. 'The "Give-and Take" Packaging of Policy Instruments: Opti-
mizing Legitimacy and Effectiveness', in M.-L. Bemelmans-Videc, R. C. Rist, and E. Ve-
dung (eds). Carrots, Sticks, and Sermons: Policy Instruments and Their Evaluation, New
Brunswick, N.J. and London: Transaction Publishers, pp. 129-148.
28
Vedung, E. 1998. 'Policy Instruments: Typologies and Theories', in M.-L. Bemelmans-Videc,
R. C. Rist, and E. Vedung (eds). Carrots, Sticks, and Sermons: Policy Instruments and
Their Evaluation. New Brunswick, N.J. and London: Transaction Publishers, pp. 21-58.
Wälti, S., D. Kübler, and Y. Papadopoulos. 2004. 'How Democratic Is “Governance”? Lessons
from Swiss Drug Policy', Governance, 17(1), 83–113.
29
Tables
Table 1: Four constellations of the relation of politics, administration and science
Decisionist model Pragmatic model Administrative-
technocratic model
Scientific-technocratic
model
Politics (executive and
parliament) dominates
The relationship be-
tween politics, admin-
istration and science is
non-hierarchical
Administration domi-
nates
Science and specialists
dominate
Table 2: Cantonal scores for the four models
Canton Decisionist
model
Pragmatic
model
Administrative-
technocratic
model
Scientific-
technocratic
model
AG 1 2 2 2
AI 1 2 2 1
BE 1 2 2 1
BL 0 2 2 0
BS 2 2 2 0
GE 0 0 2 2
GR 0 3 2 1
JU 1 2 2 1
LU 0 0 2 1
NE 0 2 -- 2
OW 0 2 1 1
SG 0 1 2 1
SH 1 0 2 2
SO 2 -- 2 0
SZ 0 1 1 1
TG 1 0 2 2
TI 1 1 2 1
UR -- -- 1 1
VD -- 2 1 1
VS 1 2 1 0
ZG 0 1 2 0
30
Table 3: Correlations between the four index variables
Scientific tech-
nocracy
Administrative
technocracy
Pragmatism
tau-b
N
tau-b
N
tau-b
N
Decisionism -.13 19 .21 18 .11 18
Pragmatism -.31 19 -.20 18 -- --
Administrative techno-
cracy
.12 20 -- -- -- --
bold: important connection (tau-b > .3)
Table 4: Hypothesis test
Hypotheses (variable) H1 (Decisionism) H2
(Pragmatism)
H3 (Administra-
tive technocracy)
H4 (Scientific
technocracy)
tau-b
N
tau-b
N
tau-b
N
tau-b
N
a) Policy design .19 19 -.13 19 .31 20 .19 21
b) Goal attainment .30 19 -.23 19 .12 20 .39* 21
* Sig. >=0.1; bold: important connection (tau-b > .3)
31
Table 5: Correlations between relation of science/ administration/ politics and alcohol pre-
vention measures
Measures
Decisionist mod-
el
Pragmatic model
Administrative-
technocratic
model
Scientific-
technocratic
model
tau-b
N
tau-b
N
tau-b
N
tau-b
N
Primary prevention
• Information
.04
19
-.27
19
.14
20
-.30
21
• Education
-.06
19
-.40**
19
.29
20
.04
21
• Health promotion
.07
19
-.22
19
.25
20
-.31
21
Price control
.08
19
.03
19
-.02
20
.31
21
Availability control
• Opening hours
.03
19
.04
19
.07
20
.08
21
• phys. Availability
.08
19
.06
19
.20
20
-.01
21
• Necessity clause
.11
19
.31*
19
.17
20
-.04
21
Secondary prevention
• Advice centres
.38**
19
-.53***
19
.30
20
.00
21
• Risk group programme
.43**
19
-.16
19
.28
20
-.07
21
• Individ. interventions
.56***
19
.40**
19
.07
20
-.11
21
Protection of minors
.01
19
-.34**
19
.05
20
.26
21
Treatment
• Advice centres
.29
19
.13
19
-.24*
20
.00
21
• Out-patient treatment
.31
19
-.02
19
-.04
20
.00
21
• In-patient treatment
-.07
19
.12
19
.12
20
.22
21
Tertiary prevention
.39*
19
.17
19
-.01
20
.07
21
Specialist research
.46***
19
.11
19
.29
20
.03
21
Education and further
education
.18
19
.19
19
.26
20
.42**
21
* Sig. >=0.1; ** Sig. >=0.05; *** Sig. >=0.01; bold: important connection (tau-b > .3)
Table 6: Correlations between relation of science / administration/ politics and authority
structures
Decisionist mod-
el Pragmatic model Administrative-
technocratic
model
Scientific-
technocratic
model
tau-b
N
tau-b
N
tau-b
N
tau-b
N
Centralization
-.51***
19
.02
19
-.06
20
-.13
21
Delegation
-.05
19
-.13
19
-.40
20
-.19
21
*** Sig. >=0.01; bold: important connection (tau-b > .3)
