Emanating political participation: Untangling the spatial structure behind participation

British Journal of Political Science

Volumn 38, Issue 2, 2008, Pages 273-289

  Cho, Wendy K Tam ^a   Rudolph, Thomas J ^a  

^a UNIVERSITY OF ILLINOIS AT URBANA CHAMPAIGN   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

EID: 48849116820     PISSN: 00071234     EISSN: 14692112     Source Type: Journal    
DOI: 10.1017/S0007123408000148     Document Type: Article

References (107)

1. Sidney Verba, Kay Lehman Schlozman and Henry E. Brady, Voice and Equality: Civic Voluntarism in American Politics (Cambridge, Mass.: Harvard University Press, 1995), p. 1.
2. Steven J. Rosenstone and John Mark Hansen, Mobilization, Participation, and Democracy in America (New York: Macmillan, 1993);
3. Verba, Schlozman and Brady, Voice and Equality.
4. Robert Huckfeldt, Politics in Context: Assimilation and Conflict in Urban Neighborhoods (New York: Agathon, 1986);
5. Robert Huckfeldt and John Sprague, Citizens, Politics, and Social Communication: Information and Influence in an Election Campaign (New York: Cambridge University Press, 1995);
6. James G. Gimpel, Separate Destinations: Migration, Immigration, and the Politics of Places (Ann Arbor: University of Michigan Press, 1999);
7. Scott D. McClurg, 'Social Networks and Political Participation: The Role of Social Interaction in Explaining Political Participation', Political Research Quarterly, 56 (2003), 449-64;
8. Diana Mutz, 'The Consequences of Cross-Cutting Networks for Political Participation', American Journal of Political Science, 46 (2002), 838-55.
9. Robert Huckfeldt, Eric Plutzer and John Sprague, 'The Alternative Contexts of Political Behavior: Churches, Neighborhoods, and Individuals', Journal of Politics, 55 (1993), 365-81;
10. Carol Kohfeld and John Sprague, 'Racial Context and Voting Behavior in One-party Urban Political Systems', Political Geography, 14 (1995), 543-69;
11. Gimpel, Separate Destinations.
12. Mutz, 'The Consequences of Cross-Cutting Networks for Political Participation'.
13. Douglas S. Massey and Nancy Denton, American Apartheid: Segregation and the Making of the Underclass (Cambridge, Mass.: Harvard University Press, 1993);
14. Douglas S. Massey and Eric Fong, 'Segregation and Neighborhood Quality: Blacks, Asians and Hispanics in the San Francisco Metropolitan Area', Social Forces, 69 (1990), 15-32.
15. Rosenstone and Hansen, Mobilization, Participation, and Democracy in America.
16. Robert Huckfeldt, Politics in Context: Assimilation and Conflict in Urban Neighborhoods (New York: Agathon, 1986).
17. Rosenstone and Hansen, Mobilization, Participation, and Democracy in America.
18. Rosenstone and Hansen, Mobilization, Participation, and Democracy in America.
19. Huckfeldt and Sprague, Citizens, Politics, and Social Communication.
20. Huckfeldt and Sprague, Citizens, Politics, and Social Communication;
21. John D. Sprague, 'Is There a Micro Theory Consistent with Contextual Analysis?' in Elinor Ostrom, ed., Strategies of Political Inquiry (London: Sage, 1982), pp. 99-121.
22. Huckfeldt and Sprague, Citizens, Politics, and Social Communication.
23. Robert Huckfeldt, 'Political Participation and the Neighborhood Social Context', American Journal of Political Science, 23 (1979), 579-92;
24. Huckfeldt, Politics in Context;
25. Huckfeldt and Sprague, Citizens, Politics, and Social Communication;
26. J. Eric Oliver, Democracy in Suburbia (Princeton, N.J.: Princeton University Press, 2001).
27. Christopher B. Kenny, 'Political Participation and Effects from the Social Environment', American Journal of Political Science, 36 (1992), 259-67;
28. Jan E. Leighley, 'Social Interaction and Contextual Influences on Political Participation', American Politics Quarterly, 18 (1990), 459-75;
29. McClurg, 'Social Networks and Political Participation';
30. Mutz, 'The Consequences of Cross-Cutting Networks for Political Participation'.
31. Ronald La Due Lake and Robert Huckfeldt, 'Social Capital, Social Networks, and Political Participation', Political Psychology, 19 (1998), 567-84;
32. Leighley, 'Social Interaction and Contextual Influences on Political Participation';
33. McClurg, 'Social Networks and Political Participation';
34. Mutz, 'The Consequences of Cross-Cutting Networks for Political Participation'.
35. McClurg, 'Social Networks and Political Participation'.
36. Mutz, 'The Consequences of Cross-Cutting Networks for Political Participation'.
37. Huckfeldt and Sprague, Citizens, Politics, and Social Communication, p. 10.
38. Brady Baybeck and Scott D. McClurg, 'What Do They Know and How Do They Know It? An Examination of Citizen Awareness of Context', American Politics Research, 33 (2005), 492-520.
39. Bernard R. Berelson, Paul F. Lazarsfeld and William N. McPhee, Voting: A Study of Opinion Formation in a Presidential Campaign (Chicago: University of Chicago Press, 1954).
40. David E. Campbell, 'Community Heterogeneity and Participation' (paper presented at the annual meeting of the Midwest Political Science Association, 2004).
41. Huckfeldt and Sprague, Citizens, Politics, and Social Communication, p. 16.
42. Baybeck and McClurg, 'What Do They Know and How Do They Know It?'
43. Robert D. Putnam, Bowling Alone: The Collapse and Revival of American Community (New York: Simon and Schuster, 2000).
44. Mark S. Granovetter, 'The Strength of Weak Ties', American Journal of Sociology, 78 (1973), 1360-80.
45. For an exception, see Oliver, Democracy in Suburbia.
46. The SCBS data were collected by telephone interview in July-November 2000.
47. Wendy M. Rahn and Thomas J. Rudolph, 'A Tale of Political Trust in American Cities', Public Opinion Quarterly, 69 (2005), 530-60.
48. A complete listing of these cities is reported in the Appendix. While each city contains a random sample of voting age adults, it should be noted that the cities themselves do not constitute a representative sample of all American cities. Fortunately, however, the SCBS data also contain a nationally representative baseline survey. A comparison of the marginals from our 32-city sample with those of the national sample indicates that our sample is representative for the variables under analysis.
49. The mean of the variable is 1.14, and the standard deviation is 1.12. In our analysis, we rescale this variable from 0 to 1. Absent from our measure of political participation is voter turnout. Our decision to exclude turnout was based primarily on measurement grounds. The instruments measuring petition, rally, march and project all asked whether respondents had engaged in such activities during the last twelve months. Since the 2000 SCBS was a pre-election survey, the turnout instrument had a very different time horizon, asking respondents whether they had voted four years prior in 1996. Additionally, some work suggests that the determinants of turnout are not always identical to those of other forms of participation. See Rosenstone and Hansen, Mobilization, Participation, and Democracy in America;
50. Verba, Schlozman and Brady, Voice and Equality.
51. Putnam, Bowling Alone;
52. Rosenstone and Hansen, Mobilization, Participation, and Democracy in America.
53. Putnam, Bowling Alone.
54. Rosenstone and Hansen, Mobilization, Participation, and Democracy in America;
55. Verba, Schlozman and Brady, Voice and Equality.
56. Alberto Alesina and Eliana La Ferrara, 'Participation in Heterogeneous Communities', Quarterly Journal of Economics, 115 (2000), 847-904;
57. Huckfeldt, Politics in Context;
58. Oliver, Democracy in Suburbia;
59. Rosenstone and Hansen, Mobilization, Participation, and Democracy in America.
60. Perhaps part of the reason is that although econometric texts commonly discuss issues related to autocorrelation on the time dimension, the spatial dimension has been much more neglected. Accordingly, spatial methods have not as quickly been adopted as part of the 'standard' toolbox. An exception is J. Johnston, Econometric Models (New York: McGraw-Hill, 1984).
61. However, issues related to spatial autocorrelation are absent from many commonly cited basic texts (see, e.g., G. Judge, R. C. Hill, W. E. Griffiths, H. Lutkepohl and T. C. Lee, Introduction to the Theory and Practice of Econometrics (New York: Wiley, 1982);
62. William H. Greene, Econometric Analysis, 2nd edn (New York: Macmillan, 1993);
63. and D. J. Poirier, Intermediate Statistics and Econometrics: A Comparative Approach (Cambridge, Mass.: MIT Press, 1995))
64. and advanced econometric texts (see, e.g., T. B. Fomby, R. C. Hill and S. R. Johnson, Advanced Econometric Methods (New York: Springer-Verlag, 1984);
65. T. Amemiya, Advanced Econometrics (Cambridge, Mass.: Harvard University Press, 1985);
66. G. Judge, W. E. Griffiths, R. C. Hill, H. Lutkepohl and T. C. Lee, The Theory and Practice of Econometrics, 2nd edn, (New York: Wiley, 1985);
67. and R. Davidson and J. G. MacKinnon, Estimation and Inference in Econometrics (New York: Oxford University Press, 1993)).
68. j are observations on a random variable at locations i and j in space.
69. Luc Anselin, Spatial Econometrics: Methods and Models (Dordrecht: Kluwer Academic Publishers, 1988).
70. L. Anselin, A. K. Bera, R. Florax and M. J. Yoon, 'Simple Diagnostic Tests for Spatial Dependence', Regional Science and Urban Economics, 26 (1996), 77-104.
71. Anselin, Spatial Econometrics;
72. Luc Anselin, 'What is Special about Spatial Data?' in Daniel A. Griffith, ed., Spatial Statistics: Past, Present, and Future (Ann Arbor, Mich.: Institute of Mathematical Geography, 1990), pp. 63-77.
73. Luc Anselin and Daniel Griffith, 'Do Spatial Effects Really Matter in Regression Analysis?' Papers, Regional Science Association, 65 (1988), 11-34.
74. Luc Anselin and Anil K. Bera, 'Spatial Dependence in Linear Regression Models with an Introduction to Spatial Econometrics,' in Aman Ullah and David E. A. Giles, Handbook of Applied Economic Statistics (New York: Marcel Dekker, 1998), pp. 237-85. Note that the specific mechanism that produces the spatial patterns is unknown and not determinable via spatial analyses. What we can uncover are patterns consistent with the specific mechanisms that produce the participation patterns that we observe. This is not unlike traditional regression analyses that are also unable to establish casual links/mechanisms.
75. The general decision rule for specification in a spatial model begins with an examination of the non-robust forms of the Lagrange Multiplier tests for the spatial error and spatial lag. Both of these may be significant. In this case, one then examines the robust forms of these Lagrange Multiplier tests and bases the specification choice (either lag or error) on the robust tests. For a discussion of the robust diagnostics, see A. K. Bera and M. J. Yoon, 'Specification Testing with Misspecified Alternatives', Econometric Theory, 9 (1993), 649-58;
76. Anselin et al., 'Simple Diagnostic Tests for Spatial Dependence'.
77. On the non-robust forms, see also P. Burridge, 'On the Cliff-Ord Test for Spatial Autocorrelation', Journal of the Royal Statistical Society, Series B, 42 (1980), 107-8.
78. Bera and McKenzie discuss the invariance of the non-robust diagnostics to different alternatives (see A. K. Bera and C. R. McKenzie, 'Alternative Forms and Properties of the Score Test', Journal of Applied Statistics, 13 (1986), 13-25).
79. Our weights matrix was created using a distance-based definition for neighbours. We are able to locate individuals in 'space' because our data identify the census tract in which the individual resides. We defined an individual's neighbour as anyone living within a two-mile radius of that individual. This calculation was made from the centroid of a census tract to the centroid of other census tracts. Importantly, this allows us to analyse the effects of geographic distance both within and across cities.
80. J. K. Ord, 'Estimation Methods for Models of Spatial Interaction', Journal of the American Statistical Association, 70 (1975), 120-6;
81. Anselin, Spatial Econometrics.
82. In our analysis, social context is clearly defined as residential location. To be sure, individuals have varied social experiences that are not limited by place of residence. Other social focal points may be at work or school. Obviously, our measure of social context does not capture these interactions, and we do not pretend that it does. Rather, our findings should be understood in the context of social interactions guided by residential location.
83. It is important to note that a spatial lag is not directly analogous to a time series lag. It seems intuitive and appealing to equate spatial autocorrelation with time series autocorrelation, but the spatial econometric literature is clear that this analogy is misleading and wrong (see Anselin, Spatial Econometrics, for an extensive discussion on this precise point). Instead, the differences are significant. Spatial autocorrelation is more complex because the nature of the dependence is multidirectional and multidimensional. In time-series data, events that occurred earlier in time affect those that occur later in time (uni-dimensional). The reverse is not true. In spatial analyses, neighbours affect each other and so the autocorrelation is two-dimensional. In addition, spatial autocorrelation can be viewed as multi-directional since each observation simultaneously affects multiple neighbours. The multi-directional and multi-dimensional nature of spatial data complicates the nature of dependence considerably and so spatial techniques are not and could not be a straightforward extension of time-series methods. This will become obvious in the main text when we further elaborate on the interpretation of the spatial lag and the spatial lag model.
84. Mutz, 'The Consequences of Cross-Cutting Networks for Political Participation';
85. McClurg, 'Social Networks and Political Participation';
86. Leighley, 'Social Interaction and Contextual Influences on Political Participation.'
87. For a discussion of this test statistic, see Anselin and Bera, 'Spatial Dependence in Linear Regression Models with an Introduction to Spatial Econometrics'.
88. 1.
89. Note that this is not the one-directional Lagrange Multiplier test that is designed to test a single specification assuming correct specification for the remainder of the model. That test would result in unwanted 'power' due to the presence of local lag dependence. Instead, in our specification, we have already noted the presence of a significant spatial lag effect. Accordingly, valid statistical inference needs to take this lag dependence into account when testing for error dependence. The specification of this Lagrange Multiplier statistic tests for error misspecification in a model with a lag term present, based on the residuals of a maximum likelihood estimation of the spatial lag model. For details of this test, see Anselin and Bera, 'Spatial Dependence in Linear Regression Models with an Introduction to Spatial Econometrics'. The insignificant test statistic provides evidence that we have sufficiently accounted for the spatial autocorrelation with the spatial lag term. In other words, the error term (a measure of the effects of variables omitted from the model) contains no remaining spatial autocorrelation and so we have evidence that the spatial autocorrelation in the data is not the result of unmeasured variables, but is sufficiently captured by the spatial lag term.
90. The lack of significant remaining spatial autocorrelation is satisfactory by traditional standards to allow us to conclude that any spatial structure has been modelled by the inclusion of the proper covariates and the spatial lag. We do note, however, that while our specification is clearly satisfactory in this regard, it may not be the only satisfactory specification. See Lance A. Waller and Carol A. Gotway, Applied Spatial Statistics for Public Health Data (New York: Wiley, 2004), chap. 9. Certainly this is akin to modelling in any regard, spatial or otherwise.
91. Huckfeldt, 'Political Participation and the Neighborhood Social Context';
92. Huckfeldt, Politics in Context;
93. Huckfeldt and Sprague, Citizens, Politics, and Social Communication;
94. Oliver, Democracy in Suburbia.
95. See, e.g., Jonathan Kelley and Ian McAllister, 'Social Context and Electoral Behavior in Britain', American Journal of Political Science, 29 (1985), 564-86.
96. Putnam, Bowling Alone.
97. Mutz, 'The Consequences of Cross-Cutting Networks for Political Participation'.
98. Leighley, 'Social Interaction and Contextual Influences on Political Participation'.
99. Moran's I was originally proposed as a simple test for correlation between nearest neighbours, a generalization of one of his earlier tests. See P. A. P. Moran, 'The Interpretation of Statistical Maps', Journal of the Royal Statistical Society, Series B, 10 (1948), 243-51;
100. P. A. P. Moran, 'The Interpretation of Statistical Maps', Biometrika, 37 (1950), 17-23;
101. ij is a standardization factor equal to the sum of the spatial weights.
102. See A. Cliff and J. K. Ord, 'Testing for Spatial Autocorrelation among Regression Residuals', Geographic Analysis, 4 (1972), 267-84;
103. A. Cliff and J. K. Ord, Spatial Autocorrelation (London: Pion, 1973). If the weights are row-standardized, Moran's I simplifies to I = e′ We/e′ e.
104. See, e.g., Leighley, 'Social Interaction and Contextual Influences on Political Participation';
105. Mutz, 'The Consequences of Cross-Cutting Networks for Political Participation'.
106. See, e.g., Alberto Alesina and Eliana La Ferrara, 'Participation in Heterogeneous Communities';
107. Dora L. Costa and Matthew E. Kahn, 'Civic Engagement and Community Heterogeneity: An Economist's Perspective', Perspectives on Politics, 1 (2003), 103-11.