Continued from page 2. Table 2 reports the means and standard deviations of variables for the groups of Americans and Chinese and for men and women. A multivariate analysis of the variance (MANOVA) was conducted. Four value variables were used as the dependent variables to examine the effect of culture (country affiliation) on individualism-collectivism. There was a significant difference in values between Americans and Chinese; F (4, 249) = 32.86, p [is less than] .0001. Univariate analyses indicated that Americans had significantly higher scores on separation from the ingroup [F(1,252) = 30.07, 12 [is less than] .0001] but significantly lower scores on family integrity [F(1,252) = 66.05, P [is less than] .0001] and self-reliance [F(1,252) = 55.09, 12 [is less than] .0001] than Chinese. The difference in the scores on interdependence was insignificant; F(1,252) = .81,12 [is greater than] .05.

Table 2 Means and Standard Deviations of Perceived Health, Self-Efficacy, Values. and Life Satisfaction

USA
Variable Male Female Total
m SD m SD m SD

Health 9.4 1.8 8.7 1.9 9.1 1.8
Efficacy 135.8 22.9 126.9 27.5 132.6 24.9
Reliance 34.3 11.1 29.6 11.6 32.6 11.5
SFI 11.2 2.2 11.3 2.0 11.2 2.1
FI 6.1 2.9 5.3 2.9 5.8 2.9
ITDPNDT 12.3 2.5 12.0 2.2 12.2 2.4
LSS 99.4 18.5 93.4 20.2 97.2 19.3

China
Variable Male Female Total
m SD m SD m SD

Health 5.5 1.5 6.6 1.6 5.9 1.5
Efficacy 110.1 25.2 114.2 26.2 111.7 25.6
Reliance 43.4 8.3 39.8 8.7 42.0 8.6
SFI 8.7 3.5 10.1 3.0 9.3 3.4
FI 8.6 2.8 8.9 2.8 8.7 2.8
ITDPNDT 12.3 2.4 12.9 2.1 12.5 2.3
LSS 71.0 17.7 75.6 21.2 72.8 19.2

Note. SFI = separation from ingroups; FI = family integrity; ITDPNDT = interdependence;

LSS = life satisfaction.

Culture, Gender and Life Satisfaction

The effects of culture and gender and the interaction effect between these two variables were examined using a univariate analysis of the variance. Values and income were used as covariate variables. The effects of the covariate variables were assessed first. The effect of income was significant; F(1,234) = 5.88, 12 [is greater than] .05. The effect of values, however, was small and insignificant (the F values ranged from .05 to .81,12 [is greater than] .05). Then, the effects of culture and gender were determined after statistically controlling for group difference in values and income. Culture had a strong effect on life satisfaction; F(1,242) = 7.76, 12 [is less than] .001. Americans had much higher life satisfaction than Chinese. However, the effect of gender was not significant; F(1,242) = .17, 12 [is greater than] .05. Males and females with SCI had a similar level of life satisfaction. The interaction effect between culture and gender was significant; F(1,242) = 6.98, 12 [is less than] .01. Chinese males had lower life satisfaction scores (m = 71.0) than Chinese females (m = 75.6) and American males (m = 99.5) and females (m = 94.5).

Correlates of Life Satisfaction in the Two Groups

Multiple linear regression analyses were performed for each of the groups separately. Life satisfaction was used as the dependent variable. Values, gender, and self-efficacy were used as the independent variables. Because the literature indicated that age, age at the time of injury, marital status, educational level, perceived health status, and income were the major contributors to life satisfaction in people with SCI (Krause & Dawis, 1992; Schulz & Decker, 1985); these variables were also included as independent variables. Prior to the regression analyses, gender and marital status were recoded using a dummy coding method. Females and individuals who were divorced, separated, widowed, or single were assigned a value of `0'; males and married individuals were assigned a value of `1.'

Because the entry order of the independent variables could affect results, a stepwise regression procedure was chosen to examine relationships between the independent variables and life satisfaction. It is important to note that, in a stepwise regression procedure, variables enter the equation according to their potency in contributing to the prediction. No hierarchy for inclusion of variables is preestablished; therefore, the results of a stepwise regression procedure reflect the predicative utility of the variables (Pedhazur, 1982).

As can be seen in Table 3, for Americans, perceived health status was the strongest predictor and entered the equation first, followed by self-efficacy, self-reliance, and marital status. The combination of the four predictors accounted for 50% of the variance in life satisfaction, F (4, 121) = 28.79, 12 [is less than] .001. Gender, separation from ingroups, family integrity, interdependence, and other demographic variables did not significantly contribute to life satisfaction when other variables in the model were controlled.

Table 3

Summary of Stepwise Regression Analysis for Gender, Values, Self-Efficacy, and Demographic Variables Predicting Life Satisfaction in Americans with SCI

Variable B SE B [Beta]

Step 1
Health 5.79 .75 .58(***)

Step 2
Health 4.39 .76 .44(***)
Efficacy .26 .06 .34(***)

Step 3
Health 4.39 .73 .44(***)
Efficacy .28 .06 .37(***)
Reliance -.37 .11 -.22(**)

Step 4
Health 4.47 .72 .44(***)
Efficacy .27 .05 .36(***)
Reliance -.38 .11 -.23(**)
Marital status 3.72 1.78 .14(*)

Note. [R.sup.2] = .33 for step 1 (p<.001); [Delta] [R.sup.2] = .43 for Step 2 (p<.001); [Delta] [R.sup.2] = .47 for Step 3 (p<.001); [Delta] [R.sup.2] = .50 for Step 4 (p<.001);

(*) p<.05; (**) p<.01; (***) p<.001.

For Chinese, self-efficacy was the strongest predictor and entered the equation first, followed by perceived health status. The combination of the two predictors accounted for 39% of the variance in life satisfaction, F (2, 112) = 33.83, p [is less than] .0001. Gender, the four value variables, and other demographic variables did not significantly contribute to life satisfaction when the influences of other independent variables in the model were controlled (Table 4).

Table 4

Summary of Stepwise Regression Analysis for Gender, Values, Self-Efficacy, and Demographic Variables Predicting Life Satisfaction in Chinese with SCI

Variable B SE B [Beta]

Step 1
Efficacy .35 .06 .49(***)

Step 2
Efficacy .28 .05 .40(***)
Health 4.69 .93 .39(***)

Note. [R.sup.2] = .24 for Step 1 (p<.001); [Delta] [R.sup.2] = .39 for Step 2 (p<.001);

(***) p<.001.

A one way analysis of the variance was conducted to examine the influence of employment status on life satisfaction. For the American group, the means of the LSS of individuals who were unemployed, employed part-time, or employed full-time were 95.94, 96.83 and 99.30, respectively. For the Chinese group, the means of the LSS were: 70.86 (unemployed), 73.05 (part-time), and 74.50 (full-time), respectively. Although participants who were employed were more satisfied with their lives than those who were unemployed, the differences were not significant at the .05 level. In addition, a one way analysis of the variance was conducted to examine differences in life satisfaction between participants with different levels of injuries. No significant difference was found for either Americans or Chinese.

Discussions

This study focused on value differences between Americans and Chinese people with SCI. the influences of culture and gender on life satisfaction, and the correlates of life satisfaction. It should be noted that all instruments used in this study were developed in the U.S. Although these instruments had been pretested with the Chinese people who had SCI, the constructs measured may have different meanings in the two cultures. Hence, the results of this study must be interpreted with caution.

Nevertheless, the results of the study indicated that the values of Americans and Chinese with SCI differed significantly and reflected the differences between individualistic cultures and collectivistic cultures. Americans tended to value keeping distance from their ingroups. On the other hand, the Chinese valued a close family relationship.