The Year of the Flood By Thomas Petersen

The Year of the Flood: The "Strength-of-Personality Scale" and the 2002 German Federal Elections

In his 1944 book, The People’s Choice, Paul F. Lazarsfeld hypothesized that there are people in all social strata who have especially strong personalities, who make a greater impression on the people in their surroundings than the mass media do, and who use their personal powers of persuasion to influence the voting decisions of their friends and acquaintances. In the social sciences and the field of survey research, these people have become known as “opinion leaders.”

Right from the start, the description of opinion leaders was linked with the concept of the “two-step flow of communication,” which held that even if—according to the state of research at the time—the mass media did not seem to have any direct influence on public opinion, it was nevertheless conceivable that opinion leaders used the media more intensively as a source of information on political topics than the remaining population did and passed on to others the ideas and arguments they gleaned.

In the early 1980s, the Allensbach Institute developed a tool to identify these opinion leaders. The so-called “strength-of-personality scale” consisted of ten simple statements presented to survey respondents, who were asked to say which they would apply to themselves.

Analysis of this scale, which had initially been designed for use in readership research, revealed that respondents who applied an especially great number of these statements to themselves also displayed a striking number of traits that Lazarsfeld had ascribed to opinion leaders during the 1940 U.S. presidential campaign. Suddenly, the newly developed scale of personality strength offered researchers possibilities that went far beyond its original purpose. If those identified by the scale were, in fact, opinion leaders in Lazarsfeld’s sense, and if the two-step flow of communication were operative in election campaigns, then these persons ought to have a decisive impact on other people’s voting decisions.

Above all, opinion swings and reappraisals of the political parties and candidates during the course of a campaign ought to be observed first among opinion leaders and only later among the remaining population. Proving this would not only further validate the opinion leader theory; it would also make the strength-of-personality scale an important election forecasting tool, one that would enable researchers to detect trends in a campaign right when they were beginning to emerge—that is, before they took hold among the general population and thus became evident in findings of representative surveys.

The question, therefore, was whether those identified as opinion leaders by the strength-of-personality scale preceded the rest of the population when it came to their voting preference. An analysis of trends ascertained by the Allensbach Institute during the 1990 German federal election campaign confirmed this supposition.

The election to the German parliament was held on December 2, 1990. From May to September of that year, there was little change in party strength. Among opinion leaders, the governing Christian Democratic Party (CDU/CSU) was clearly leading the opposing Social Democrats (SPD), whereas the two parties were neck-and-neck among the remaining population.

Suddenly, however, in early October, the share of opinion leaders who said they would vote for the Christian Democrats if the election were held the following Sunday swelled from about 45 percent to over 50 percent. Presumably, the festivities relating to German reunification on October 3 contributed to this mood swing. No changes in voting intentions were observed among the remaining population, though, until three weeks later, when they also clearly shifted in favor of the Christian Democrats (Figure 1).

As Figure 2 shows, a similar pattern was observed during the 1994 election campaign.