This paper examines the record of western democracies to measure the impact of differing electoral formulae on the rate of voter turnout. The record of 509 national elections in 20 countries provides the basis for a regression analysis that clearly identifies higher turnout rates in PR systems that cannot be explained by a wide variety of control variables or traditional arguments about PR. The data also reveal a marked increase in electoral turnout over the last century.

Previous research on voter turnout has concluded that institutional factors such as compulsory voting, registration laws and the degree of disproportionality of the electoral system have a determining impact on the number of people who cast their ballots. This article goes beyond such an institutional explanation. It is hypothesized that a diverse political landscape which offers a great menu of political choices to electors is more likely to have a higher voter turnout than a political system in which the differences between the parties are small. The greater the range of political expression available the more people are stimulated to vote. This study found that countries with a high party polarization have significantly higher turnout rates than countries with a rather narrow political spectrum. In addition, countries in which a ‘postmaterialist party’ is present in the national parliament also display significantly higher voter turnout rates than countries which lack such a party. In order to explain the determinants of voter turnout, not only institutional variables but also intrinsically political variables such as the number of parties, the type of parties and party polarization have to be introduced. To solve the puzzle of varying voter turnout rates, ‘politics’ has to be brought back in.

The way we talk about politics has an impact on the way we perceive them. Thus, language not only describes a world, but gives sense and meaning to it. Metaphors in particular imply ways of seeing ‘something’, as well as instructions on how to handle ‘something’. These insights are used to develop and operationalize the concept of ‘mental pictures’. Mental pictures of the state as they appear in the platforms of the West German Social Democratic and Christian Democratic parties (SPD and CDU) were examined by means of qualitative and quantitative content analysis. The documentary texts date mainly from the 1970s. Apart from different judgements of the state perceived as an apparatus, differences between the parties can be found where ‘human’ attributes such as strenght, leadership or the ability to inspire confidence are evoked. In spite of its liberalistic distance from the state, which it stresses in other contexts, the CDU prefers a state conceived of as a leadership figure. In contrast, the SPD is averse to such a mental picture. A technocratic image is the (necessary?) price the SPD pays for this aversion.

The paper tests the ‘folk’ hypothesis that the vote at elections in different countries is affected by international opinion swings which are sometimes to the right and sometimes to the left. An hypothesis that, if it had been true, would have allowed us to predict the election outcome in one country by considering the outcome in recent elections in other countries. In order to test this hypothesis all (governmental) elections between 1948 and 1985 in 17 main democracies are considered. A set of series S(i, t), where i is a country and t time, showing the left/right-swing at the election is constructed. We next delete the i-index and reach the S(t)-swing series. The hypothesis is then operationalized as saying that S(t) contains positive autocorrelation. The S(t)-series is shown to behave as perfect white noise, so the hypothesis is rejected.

A wide variety of publications on the Italian party system can be reduced to essentially two opposite approaches. The first approach, explains the dysfunctioning of Italy's party system by analysing the practices of political elites. These are said to manipulate society by using the state as an instrument. Authors who take this approach generally neglect the impact of ideological discord and inter-party dynamics. This approach is often related to sharp criticism of the governing parties, and a rather positive assessment of the Communist Party. By contrast, Giovanni Sartori and others who take the opposite approach underscore the importance of ideological heterogeneity and destabilizing inter-party dynamics. Political elites are thought to react to an unstable environment. This can be used in justification of the governing parties, despite the weak performance of Italian government. Sartori's analysis of the Italian party system has been sharply criticized. Pointing out the fundamental difference between these two approaches, this article sheds some light on the roots and limitations of this criticism.

Dekker and Ester's (1987) position regarding working class authoritarianism, using Wright's (1979, 1985) neo-marxist class-conceptualization, is challenged. It is shown in replication that Wright's class concepts are only weakly and non-linearly related to ‘class-proxies’ such as income, wealth, educational level and subjective social class. Thus there is reason for doubt regarding the validity of Wright's class-concept, which might be partly responsible for the weak associations found with authoritarianism and authoritarian attitudes. Nevertheless, we found that Dekker and Ester's (1987) hypotheses 1 and 2 cannot be rejected: the working class is somewhat more authoritarian than all other classes combined. We maintain that the well-known relatively strong relationship between educational level and authoritarianism remains of major theoretical and social relevance.

Our reply to the forgoing article by Middendorp and Meloen rejects their critique for three reasons:

1. 1. their requirement of ‘empirical validity’ has no relevance with respect to our argumentation to use E.O. Wright's class measurements.

2. 2. they fail to see the relational and qualitative character of the class measurements we used, and

3. 3. they do not show any empirical evidence which contradicts our findings.

We do agree with Middendorp and Meloen on the importance of education for explaining authoritarianism. This is exactly the reason not to mix up education with class: education is just too important to misuse it as a class indicator.