Disclaimer. Don't rely on these old notes in lieu of reading the literature, but they can jog your memory. As a grad student long ago, my peers and I collaborated to write and exchange summaries of political science research. I posted them to a wiki-style website. "Wikisum" is now dead but archived here. I cannot vouch for these notes' accuracy, nor can I say who wrote them.
Norris. 2002. Democratic Phoenix. Cambridge: Cambridge University Press.
Norris uses cross-national data to test existing theories about turnout. She first predicts each country's aggregate turnout, then she uses survey data to predict individual turnout within each country. Though this study is informative, her method leaves much to be desired, as discussed below.
This list summarizes the variables from chapters 3 and 4. I discuss them in more detail below. Each bullet indicates a specific indicator, whether it is significant, and the direction of the relationship. The indicators are grouped into larger variable classes.
Measuring (aggregate) turnout: She uses number of voters over total population size--ignoring cross-national variations in who is eligible or registered. Critique: This is no light matter; McDonald and Popkin (2001) show that this decision dramatically changes the "trends" in American turnout, a finding that Norris acknowledges but does not adequately address (p 41).
"Core hypothesis": Modernization. Economic, social, and political development are closely intertwined; changes in one area affect the others. Turnout will rise as countries move from low to medium development, then stabilize due to a "ceiling effect."
Test: See Table 3.1 and Figures 3.3-4 (in book). Sure enough, postindustrial societies have the highest turnout, followed by medium- and low-development societies. Also, turnout is higher in older democracies than in new ones.
Other findings: Over the last 50 years, turnout in postindustrial societies has been quite stable, but it has steadily increased in less developed societies. It is not clear to me what relevance this has to the modernization hypotheses--other than to demonstrate that other variables must also be having an influence.
Critique: Norris has fabulous time-series data at her disposal. Yet she doesn't really test her "core hypothesis." She needs to show that when countries change from low to high development that turnout increases, not just that turnout is higher when development is. She claims in her conclusion that "countries with rapid human development have experienced substantial growth in electoral turnout," but she hasn't really tested this.
Motivational hypotheses: This chapter looks at "motivational" theories of turnout--basically, the standard rational calculus of voting. She calls C, B, and P "costs," "choices," and "decisiveness," and uses this language to introduce institutional, legal, and ease-of-voting concerns. These issues, together with the modernization variables from chapter 3, are summarized in the last table.
Critique: Three variables have unexpected findings: compulsory voting, proxy voting, and number of polling days. In each case, Norris suggests that these results might be an artifact of her cross-sectional approach. After all, if countries with low turnout introduce these reforms in an effort to boost turnout, then it might appear that these reforms actually hurt turnout. She's right. She ought to be doing time-series analysis, not cross-sectional analysis. These three unexpected results should cast doubt on all her results; even the expected findings might be false positives, artifacts of the cross-sectional design.
Norris now switches to examining individual-level turnout data from an international survey. Like Brady, Verba, and Schlozman (1994), she argues that people don't turn out because they can't (structural), they won't (cultural), or nobody asked (mobilization). Here's another list of indicators, formatted like the one above:
Research by the same authors
Research on similar subjects