The "Alternation effect" in Africa

Journal of Democracy

Volumn 15, Issue 4, 2004, Pages 147-158

  Bratton, Michael ^a  

^a MICHIGAN STATE UNIVERSITY   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

EID: 7544245577     PISSN: 10455736     EISSN: None     Source Type: Journal    
DOI: None     Document Type: Review

References (19)

1. "Key Findings on Public Opinion in Africa," Afrobarometer Briefing Paper 1 (April 2002), available at www.afrobarometer.org.
2. The Afrobarometer project is a joint enterprise of the Institute for Democracy in South Africa (Idasa), the Center for Democratic Development (CDD-Ghana), and Michigan State University (MSU). Hence I use a collective "we" in describing methods and reporting results.
3. Michael Bratton, Robert Mattes, and E. Gyimah-Boadi, Public Opinion, Democracy, and Market Reform in Africa (New York: Cambridge University Press, forthcoming 2005).
4. Averages are calculated on pooled data, weighted to standardize each national sample to N=1200. That is, each country carries equal weight in aggregate means, regardless of sample size or overall population.
5. That is, more people were unable to choose between the available response options and could not proffer any answer.
6. The increase in "don't know" responses also derives from persons in Ghana and Botswana who did not recognize the word "democracy" in English. Comparisons may be affected by this change of question wording between Rounds 1 and 2.
7. Response categories form a five-point scale from "disapprove strongly" through "disapprove," "neither," and "approve," to "approve strongly." In the text and figures, summary figures for rejection of authoritarian rule combine "disapprove" and "disapprove strongly."
8. Response categories form a four-point scale from "not at all satisfied" through "not very satisfied" and "fairly satisfied" to "very satisfied." In the text and figures, summary figures for satisfaction combine "fairly" and "very."
9. See endnote 5.
10. Samuel P. Huntington, The Third Wave: Democratisation in the Late Twentieth Century (Norman: University of Oklahoma Press, 1991), 266-67.
11. These outlying cases tend to cancel each other out: While Botswana undermines the argument that democratic attitudes decline in the absence of alternation, the case of Zimbabwe embodies it.
12. Always significant, the correlation coefficients (Pearson's r) ranged from -.081 for change in satisfaction with democracy to -.665 for change in the perceived extent of democracy.
13. In Lesotho and South Africa we counted parliamentary elections.
14. Down 5 percent, from 61 to 56 percent. See "Alarm Call for Latin America's Democrats," The Economist, 28 July 2000; and "Democracy Clings On in a Cold Economic Climate," The Economist, 26 August 2002.
15. Down 5 percent, from 61 to 56 percent. See "Alarm Call for Latin America's Democrats," The Economist, 28 July 2000; and "Democracy Clings On in a Cold Economic Climate," The Economist, 26 August 2002.
16. Down to 43 percent in 2003. See United Nations Development Programme report "Democracy in Latin America: Towards a Citizens' Democracy" (2004).
17. Up from 58 percent to 65 percent, then back to 57 percent. Richard Rose and Christian Haerpfer, "Trends in Democracies and Markets: New Democracies Barometer, 1991-98," Studies in Public Policy 308 (Glasgow: University of Strathclyde, 1998).
18. For a two-item construct of perceived supply of democracy (an average of satisfaction and extent), Pearson's r=.204, p<.001.
19. For the four-item construct of demand for democracy (an average of support democracy plus reject three alternatives), Pearson's r=-.043, p<.001.