Book Review Ole Borre & Jørgen Goul Andersen: Voting and Political Attitudes in Denmark. Aarhus University Press, Aarhus 1997. 352 pp.

For the first time, Denmark's electoral researchers have written an English language monograph on Danish electoral behavior: Voting and Political Attitudes in Denmark, by Ole Borre, Denmark's grand old man in electoral research, and the prolific welfare state researcher Jørgen Goul Andersen. Similar books on Finland. Norway. and Iceland already exist (Pertti Pesonen: An Election in Finland from 1968 and Sammi Borg & Risto Sänkiaho (eds.). The Finnish Voter from 1996: Ola Listhaug: Citizens, Parties and Norwegian Electoral Politics 1957-1985 from 1989; and Olafur Hardarson: Parties <&. Voters in Iceland from 1995). and all these works increase the pressure on Swedish electoral researchers to make an equally impressive effort and improve communication with the international audience.

Voting and Political Attitudes in Denmark deals with the electorate's votes and opinions in connection with Denmark's 1994 parliamentary election, but also with trends in Danish electoral behavior from 1971. i.e., from the ten elections. The book contains 12 chapters and a short introduction and is divided into two sections: Chapters 1-6 and 12 discuss Danish electors' voting and why the 1994 election went the way it did. Chapters 7-11 focus on the electorate's opinions on a number of issues, such as the scope of government, the welfare state, unemployment, the European Union, and trust in the political system. The author of each chapter is stated in the preface.

As many of its predecessors, the book lacks a common thread. Therefore, the order of the chapters seems a bit random, an indication of both strength and weakness: strength because each chapter merits attention by itself; weakness because the chapters would have benefited from a little more "mutual pollination." more of a summarizing multivariate analysis of why people vote as they do. and more of a common basis for the chapters on opinions, how they are formed, and how opinion differences can be explained.

As it stands, this is still a solid piece of research which contains an enormous amount of information, many intelligent observations, and it is highly readable. Voting and Political Altitudes in Denmark is. in short, new proof that research on political behavior in general and election and opinion research in particular are among the most prominent sub-disciplines in political science.

But, as we know, nobody is perfect. The authors exhibit specifically three types of weaknesses, which, unfortunately, are all too common in our field: too much emphasis on the absolute level of attitudes, missing causal models, and standardized regression coefficients.

1) First of all. Jørgen Goul Andersen relies too much on the absolute level of attitudes. Most people who do this kind of research know that the formulation of the interview questions determines the outcome of the opinion distributions: we usethis knowledge to create the necessary variance in the dependent variables. The simple

rule in accounting for and interpreting results is therefore to avoid journalistic
comments on the attitude level and instead focus on changes over time and
differences between groups.

Unfortunately, Jørgen Goul Andersen does exactly the opposite in chapter 8, which discusses the Danes' attitudes towards the welfare state. What he mainly does is argue for political theses by way of statements of the type, "most people really do want to maintain the welfare state" (p. 221) and "most people appear satisfied with present welfare arrangements, satisfied with the balance between welfare and taxation, but troubled by the expected long-term economic implications" (p. 223). There are already way too many welfare state researchers — to the left and the right on the political spectrum - who have trouble keeping cool. One wonders if Voting and Political Attitudes in Denmark would have benefited if Ole Borre had written the chapter on the welfare state and instead left something less politically controversial to Jørgen Goul Andersen.

2) Some of the multivariate analyses in the book would have been more complete had they been carried out within the framework of a causal model with a theoretically fixed order between the independent variables. The problem is most obvious in sections 4.7 and 4.8 where several variables, without mutual causal order and with a probable high covariation, are used to explain votes for the government and for the bourgeois parties. The results of the analyses are unclear since the presented direct effects most likely give an incorrect picture of the causal impact of the variables, which must be found in the not shown total effects.

The same superficial method is unfortunately also used in one of the most interesting analyses of explanations of political trust. The concluding multivariate analysis tells us that "policy distance is the main source of political distrust," but that also school education and spending attitudes have clear effects (p. 318). How this concluding sentence would do in a stricter causal model with maybe one or two more variables is an interesting question for future research.

3) Fortunately, several of the many regression analyses build on the interpretable unstandardized regression coefficients, but at the same time too many analyses build on standardized beta-coefficients. In the beginning of the 1980s, it seemed as if the standardized regression coefficients would be wiped out for good. Internationally, fatal shots against the beta-coefficients were fired by for example Christopher Achen in Interpreting and Using Regression (1982) and by Mark Franklin in The Decline of Class Voting in Britain (1985). In the Nordic countries, the successful battle was fought by the Norwegians Petter Laake and Trond Petersen in a long debate with Knud Knudsen, Tom Colbjørnsen and Gudmund Hernes in Tidsskrift for Samfunnsforskning from 1981^1983 - later supported by the Swedes Björn Halleröd's and Peter Stern's pedagogical article, "Att jämföra äpplen och päron med bananer" ("To compare apples and pears with bananas") in Sociologisk forskning in 1991.

However, it seems that the standardized regressional coefficient virus has only been dormant, because recently it has reappeared here and there. Unfortunately, Voting and Political Attitudes in Denmark belongs to the (newly) infected. The worst example is the chapter on class voting in which certain analyses exclusively contain dichotomous variables, and where, consequently, there was never any reason whatsoever to look askance - in an unmotivated state of panic from not being able to compare apples and bananas - at the hateful beta-coefficients.

My opinion - which may be rather magisterial - is that those who still want to use
standardized regression coefficients should start out by re-reading about the above

mentioned "fatal shots" and then explain whether any new arguments have emerged
recently that could possibly explain this resurrection.

Voting and Political Attitudes in Denmark is, as already stated, so interesting and
informative that I cannot do it justice in this limited space. Let me just mention some
highlights and some weaknesses:

Chapter 1 is an excellent introduction to Danish politics and voting behavior, supplemented with interesting analyses of, for example, the restitution of the traditional parties after the 1973 earthquake election, the dilemma of the center parties, the voters' swing to the right, and the, after all. limited impact of election campaigns.

Chapter 2 presents an interesting and useful division of voters into four types. which I, for the sake of simplicity, choose to call tolerant left, non-tolerant left, tolerant right, and non-tolerant right. Left-right is measured in a traditional way with a number of questions about the state's relations to business, equality, and the size of the public sector, while tolerance is measured by attitudes towards foreign aid. immigration, punishment for violent crime, and environmental efforts. The new toleration dimension, which some readers may interpret in terms of political correctness, is a recurrent theme in several other chapters and contributes to an increased understanding of phenomena such as the new left, the right-leaning trends in public opinion, the success of the Progressive Party, and the growing distrust in politicians.

Chapter 3 contains, among many other things, a Danish showdown with the supertrendy, but empirically poorly founded, directional theory of issue voting that was introduced by George Rabinowitz. Stuart Elaine Macdonald and Ola Listhaug. Ole Borre here shows that the incapacity ot the directional theory to compete with Anthony Downs' traditional proximity theory holds also in the Danish case.

Several chapters - especially 5 and 8 - fight an. in my opinion, unnecessary battle against the so-called self-interest perspective and claims are made that it is primarily the voters' values that determine their opinions and behavior. However, a sufficient theoretical foundation in the form of clear definitions of "what is what" is never presented, nor are we dealing with a systematic test of the eternal and maybe impossible question of self-interest versus values and public interest. We are only offered scattered results and subjective interpretations that generally land on the "good" side, i.e., Danish voters exhibit very little self-interest. These lax reasonings and passages should have been left out. because they may reduce the credibility of other, more well-founded conclusions.

Chapter 9 is an innovative and thorough chapter about the voters' views on various anti-unemployment strategies. The voters turn out to have opinions on the various strategies they are presented with, and these opinions do not fit well with how the economists and the political parties want to solve the problems. This is most obvious in the contrast between the voters' positive attitudes to "taking leases in turn in order to share jobs which is rejected by most experts as well as by most political parties" (p. 278). If anything is missing in the chapter, it is a better recognition of the fact that answers to this kind of complex, specialized economical interview questions should be taken with a grain of salt.

Finally, chapter 11 is one of the richest in the book, where we again see the aforementioned analyses of how education, toleration and policy distance can explain trust in parties and politicians. Denmark, or possibly Danish electoral research, seems to have been pioneers in this area: "The association between distrust

and new right position has not been reported for other countries in a form that is as
unambiguous as in the Danish electorate" (p. 316).

At the normative level, Ole Borre sticks his neck out here and there may be some truth to what he says. In any case, it is a new thought to me that distrust in parties and politicians is not necessarily something that must be remedied in a democracy: "Political distrust in this sense is probably a price to be paid for a development of society which is spearheaded by an intellectually and technologically advanced elite with an international orientation" (p. 319).

It was a pleasure reading the book.

Department of Political Science, Göteborg University