Research Analysis Lib Data

In the subsequent regression analysis therefore the

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: ariables are basically independent of each other. In only one instance does a zero-order correlation exceed .50. However, the correlation of –.69 between the south and midwest regions raises questions concerning multicollinearity. In the subsequent regression analysis, therefore, the Midwest is the “omitted” or “criterion variable.” Thus, we conclude that the multiple regression analysis utilizes a set of variables that are relatively independent of each other in their effects on the dependent variable, that is, the rate of currently divorced persons. Descriptively, the counties studied show a mean rate of 75.5 currently divorced persons per 1,000 population and a mean religious homogeneity score of .29 (as measured by the Herfindahl Index). The Index score means that the odds are 29 percent that two persons selected at random from the sampled counties will be affiliated with an organized religious denomination of the same type. Table 1 Descriptive Statistics Variables Divorced rate Religious concentration % Population change % 15–34-years % White % Native American % Asian/Pac. Islander % Hispanic % Female % Unemployed % Employed manufacturing % Urban Median family income Northeast region South region West region Midwest region ¯ X S.D. 75.45 .29 3.57 28.65 86.95 1.66 .76 4.12 50.90 6.91 18.57 34.83 28,094.03 .08 .41 .14 .37 18.59 .15 15.15 4.82 16.58 7.39 3.40 9.66 1.95 4.92 10.81 29.28 7,013.84 .27 .49 .34 .48 THE IMPACT OF RELIGIOUS HOMOGENEITY 347 Further, the descriptive results indicate that within the selected counties there was a 3.6 percent population increase between 1980 and 1990 and that 28.7 percent of the population was between 15 and 34 years of age. Racial/ethnic percentages show that the population of these counties in 1990 was 87 percent white, 1.7 percent Native American, 0.8 percent Asian or Pacific Islander, 4.1 percent Hispanic and, by subtraction, 10.5 percent African American and other races. Females comprised 51 percent of the total population for these counties. The economic stability measures indicate that 6.9 percent were unemployed, 18.6 percent were working in manufacturing occupations, 34.8 percent resided in urban areas, and that the average median income for families was $28,094. Geographically, 8 percent of the counties were located in the Northeast, 41 percent in the South, 14 percent in the West, and 37 percent in the Midwest. Results of Regression Analysis The essential issue addressed in this research is the impact of religious concentration on the rate of currently divorced, using county of residence as the main contextual variable. Examination of Block 1, Table 2, shows that the zeroorder correlation between the divorce rate and religious concentration, r = –.144 ( p < .05), is weak, but statistically significant. Only about 2 percent of the variance in the divorce rate is explained by religious concentration. Still, irrespective of Table 2 Block Regression Analysis of the Divorced Rate on Religious Concentration and Selected Variables ( N = 621) Independent variables Block 1 Religious concentration Block 2 Religious concentration % Population change % 15–34-years % White % Native American % Asian/Pac. Islander Slope (b) S.E. (b) Beta ( β ) –.018** .005 –.144 Constant = 80.89; S.E. = 1.68 (t = 48.30; p < .001) R2 = .021 (1/619df; F = 13.10; p < .001) –.017** .457** –0.342* .029 .050 .560** .004 .050 .170 .048 .098 .217 –.136 .372 –.089 .026 .020 .102 348 LARRY C. MULLINS ET AL. Table 2 (continued ) Slope (b) % Hispanic % Female % Unemployed % Employ. manuf. % Urban Median family income –.224** .075 –.116 –2.114** .409 –.222 .572** .142 .151 .301** .065 .175 .252** .030 .397 –.004** .000 –.160 Constant = 187.88; S.E. = 24.20 (t = 7.76; p < .001) R2 = .338 (12/608df; F = 25.79; p < .001) R2 change Block 1 to Block 2 = .317 F of change = 26.41 (11/608df; p < .001) Block 3 Religious concentration % Population change % 15–34-years % White % Native American % Asian/Pac. Islander % Hispanic % Female % Unemployed % Employ. manuf. % Urban Median family income West Northeast South *p < .05, **p < .01. S.E. (b) Beta ( β ) Independent variables –.015** .004 –.119 –.361** .047 .294 –.166 .161 –.043 .119* .048 .106 .033 .094 .013 .486* .203 .089 –.366** .072 –.190 –1.523** .388 –.160 .551** .133 .146 .360** .061 .109 .243** .027 .382 –.001** .000 –.147 20.124** 2.094 .372 1.212 2.406 .018 7.862** 1.561 .208 Constant = 137.60; S.E. = 23.08 (t = 5.96; p < .000) R2 = .433 (15/605df; F = 30.74; p < .001) R2 change Block 2 to Block 3 = .095 F of change = 33.83 (3/605df; p < .000) THE IMPACT OF RELIGIOUS HOMOGENEITY 349 any particular denominational categorization, the greater the homogeneity of the religious environment, the less the likelihood of divorce. The question remains,...
View Full Document

This note was uploaded on 01/27/2014 for the course SOCI 3040 taught by Professor Lauramckinney during the Fall '13 term at Tulane.

Ask a homework question - tutors are online