-
1
-
-
54749086366
-
-
Cluster analysis is a set of statistical techniques that group objects into homogeneous categories on the basis of their similarities. This statistical tool discovers meaningful structures within the data set that would otherwise remain hidden. For further discussion of cluster analysis, see Maurice Lorr, Cluster Analysis for Social Scientists San Francisco: Jossey-Bass Publishers, 1983, pp. 1-6;
-
Cluster analysis is a set of statistical techniques that group objects into homogeneous categories on the basis of their similarities. This statistical tool discovers meaningful structures within the data set that would otherwise remain hidden. For further discussion of cluster analysis, see Maurice Lorr, Cluster Analysis for Social Scientists (San Francisco: Jossey-Bass Publishers, 1983), pp. 1-6;
-
-
-
-
3
-
-
84870974222
-
-
"Cluster Analysis," Statsoft, http://www.statsoft.com/textbook/ stcluan.html.
-
Cluster Analysis
-
-
-
4
-
-
27944501739
-
The Two South Koreas: A House Divided
-
Summer
-
Chaibong Hahm, "The Two South Koreas: A House Divided," The Washington Quarterly 28, no. 3 (Summer 2005): 57.
-
(2005)
The Washington Quarterly
, vol.28
, Issue.3
, pp. 57
-
-
Hahm, C.1
-
6
-
-
33947413068
-
How to Deal With South Korea
-
Winter
-
Sunhyuk Kim and Wonhyuk Lim, "How to Deal With South Korea," The Washington Quarterly 30, no. 2 (Winter 2007-08): 74-75;
-
(2008)
The Washington Quarterly
, vol.30
, Issue.2
, pp. 74-75
-
-
Kim, S.1
Lim, W.2
-
9
-
-
54749147922
-
-
Kim and Lim, How to Deal With South Korea, pp. 75-78;
-
Kim and Lim, "How to Deal With South Korea," pp. 75-78;
-
-
-
-
10
-
-
33750226826
-
Increasing Distrust of the USA in South Korea
-
October
-
Myongsob Kim, Suzanne L. Parker, and Jun Young Choi, "Increasing Distrust of the USA in South Korea," International Political Science Review 27, no. 4 (October 2006): 432-433.
-
(2006)
International Political Science Review
, vol.27
, Issue.4
, pp. 432-433
-
-
Kim, M.1
Parker, S.L.2
Young Choi, J.3
-
11
-
-
54749113703
-
-
Kim and Lim, How to Deal With South Korea, p. 76.
-
Kim and Lim, "How to Deal With South Korea," p. 76.
-
-
-
-
12
-
-
54749114837
-
-
Out of 1,001 respondents, only respondents who had complete data tor all of the questions were included in the analysis. As a result, 16 respondents whose data was incomplete were excluded from analysis, leaving 985. For the purposes of this study, the number of clusters was set at two. The resultant cluster 1 comprised 64 percent of the population, with cluster 2 the other 36 percent. The mean resultant age for cluster 1 was 39, with 27 percent born prior to 1960; for cluster 2, it was 47, with 51 percent born prior to 1960.
-
Out of 1,001 respondents, only respondents who had complete data tor all of the questions were included in the analysis. As a result, 16 respondents whose data was incomplete were excluded from analysis, leaving 985. For the purposes of this study, the number of clusters was set at two. The resultant cluster 1 comprised 64 percent of the population, with cluster 2 the other 36 percent. The mean resultant age for cluster 1 was 39, with 27 percent born prior to 1960; for cluster 2, it was 47, with 51 percent born prior to 1960.
-
-
-
-
13
-
-
54749106225
-
-
The sampling error is ±3.1 percent with a 95 percent confidence level
-
The sampling error is ±3.1 percent with a 95 percent confidence level.
-
-
-
-
14
-
-
54749119719
-
-
Because cluster analysis uses variables (the responses by survey participants) as the criteria for gouping, the variables must be comparable. The response scale for most of the survey questions were on a continuous scale, but some were on a nominal scale. Thus, this study used a two-step cluster analysis procedure integrating both types of survey questions. Among the survey questions with a continuous scale, most use an 11-point scale, but some use a 4- or 5-point scale. To eliminate problems associated with using different response scales, all variables were standardized internally by the SPSS computer analysis program. Through such standardization, each survey question was equally weighted. See Richard D. De Veaux and Paul F. Velleman, Intro Stats Boston: Pearson, 2004, pp. 82-83
-
Because cluster analysis uses variables (the responses by survey participants) as the criteria for gouping, the variables must be comparable. The response scale for most of the survey questions were on a continuous scale, but some were on a nominal scale. Thus, this study used a two-step cluster analysis procedure integrating both types of survey questions. Among the survey questions with a continuous scale, most use an 11-point scale, but some use a 4- or 5-point scale. To eliminate problems associated with using different response scales, all variables were standardized internally by the SPSS computer analysis program. Through such standardization, each survey question was equally weighted. See Richard D. De Veaux and Paul F. Velleman, Intro Stats (Boston: Pearson, 2004), pp. 82-83.
-
-
-
-
15
-
-
84963045424
-
Types of Political Attitude Structure: Results of a Cluster Analysis
-
See also, Fall
-
See also John A. Fleishman, "Types of Political Attitude Structure: Results of a Cluster Analysis," Public Opinion Quarterly 50, no. 3 (Fall 1986): 374.
-
(1986)
Public Opinion Quarterly
, vol.50
, Issue.3
, pp. 374
-
-
Fleishman, J.A.1
-
16
-
-
54749105858
-
-
All of the survey questions included in figure 1 used a continuous scale. For the full list of survey questions, see appendix at http://www.bw.edu/academics/pol/faculty/Chae/appendix.pdf. Most of the survey questions range from zero to 10, with a midpoint of 5. Survey questions with a 4- or 5-point scale were rescaled so that all the questions used in this graph have the same 11-point scale (0-10, For easy comparison, we linearly transform questions whose endpoint values are different, A and B respectively, to the same scale by using: y, mx, b, where m, 10/(B, A) and b, mA. This mapping preserves the shape of the histogram representing the respondent's answers while mapping the endpoints and midpoint to 10, 0, and 5, respectively
-
All of the survey questions included in figure 1 used a continuous scale. For the full list of survey questions, see appendix at http://www.bw.edu/academics/pol/faculty/Chae/appendix.pdf. Most of the survey questions range from zero to 10, with a midpoint of 5. Survey questions with a 4- or 5-point scale were rescaled so that all the questions used in this graph have the same 11-point scale (0-10). For easy comparison, we linearly transform questions whose endpoint values are different, A and B respectively, to the same scale by using: y = mx + b, where m = 10/(B - A) and b = -mA. This mapping preserves the shape of the histogram representing the respondent's answers while mapping the endpoints and midpoint to 10, 0, and 5, respectively.
-
-
-
-
17
-
-
54749092900
-
Election in South Korea Is Missing Its Suspense,
-
December 17
-
Norimitsu Onishi, "Election in South Korea Is Missing Its Suspense," New York Times, December 17, 2007, http://www.nytimes.com/2007/12/17/world/asia/ 17korea.html?ex=1355634000&en=2a3bd95f71f82679&ei= 5124&partner=permalink&exprod=permalink.
-
(2007)
New York Times
-
-
Onishi, N.1
-
18
-
-
54749114033
-
-
Lee Joo-hee, Lee Champions Realist, Pragmatic Approach to N.K., Korea Herald, December 21, 2007.
-
Lee Joo-hee, "Lee Champions Realist, Pragmatic Approach to N.K.," Korea Herald, December 21, 2007.
-
-
-
|