The Riddle Part 2, and the Trouble with Experiments

A public opinion riddle, part 2:

To explain the 1965 election public opinion shifts, Noelle-Neumann offered a new theory of social influence: the spiral of silence. The catalyst for the phenomenon was the favorable television coverage of the Christian Democrat party during the Queen of England’s state visit to Germany. In the summer of 1965, the Queen of England had embarked on an 18-city tour of the country. This symbolic state visit, which was covered widely on German television networks, undoubtedly served as good publicity for the Christian Democrats. Christian Democratic Chancellor Ludwig Erhard had accompanied the Queen to most sites, and TV broadcasts often showed him appearing with the Queen to greet cheering crowds (Kennamer 1990, Kaid 2007). Noelle-Neumann argued that, amidst this wave of publicity, supporters of the Social Democrats perceived a lack of social support for their party among the German public. They thus began to see themselves as belonging to a distinct “opinion minority” whose political positions were perceived as less legitimate than those of their dominant Christian Democratic opposition. This lack of perceived legitimacy compounded with a universal fear of social isolation to discourage supporters of the Social Democratic platform to publicly disclose their political opinions and voting intentions in the polls leading up to the election.        

In the 40 years since Noelle-Neumann first articulated her spiral of silence theory, political scientists and communications scholars alike have subjected it to countless empirical tests on a variety of controversial political and social issues (Kennamer 1990). The vast majority of these studies have employed survey methodology. Respondents first indicate their opinions on a controversial issue, as well as what they perceive to be the prevailing opinion in the community and among the public. The survey then typically presents a hypothetical discussion of the controversial and asks respondents whether they would be willing to express their own view during a spontaneous conversation on the topic (Glynn, Hayes, and Shanahan, 1997). 

My study is an experimental test of Noelle-Neumann’s theory. The experiment is a 2×2 factorial design, where I have manipulate both opinion agreement/disagreement and social endorsement. After collecting my data and doing an initial analysis of my results, however, I’ve come across a problem — apparently, my sample size of 209 respondents is too small — underpowered, in experimental design speak — to detect statistically significant results between treatment groups. I’m going to read up on how to deal with (salvage?) underpowered experiments, and report back in the next post.


Glynn, C. J., Hayes, A. F., & Shanahan, J. (1997). Perceived Support for One’s Opinions and Willingness to Speak Out: A Meta-Analysis of Survey Studies on the” Spiral of Silence”. Public opinion quarterly, 452-463.

Kaid, L. L., & Holtz-Bacha, C. (Eds.). (2007). Encyclopedia of political communication. SAGE publications.

Kennamer, J. D. (1990). Self-Serving Biases in Perceiving the Opinions of Others Implications for the Spiral of Silence. Communication Research, 17(3), 393-404.