Current Status and the Differences among the Self-concept and Social Support- 53

5.3.1 Current status and differences of factors relating to self-concept and social support

The purpose of this analysis was to understand the factors of self-concept. The scale contained six factors, including: Specialized subjects, stress tolerance, physical strength, friends, relationship with parents, and career planning. The mean and standard deviation of each factor are shown in Table 5.6. As seen, the mean score for Specialized subjects was 3.48, that for stress tolerance was 3.58, that for physical strength was 3.42, that for friends was 3.48, that for relationship with parents was 3.70, and that for career planning was 3.23.

Table 5.6 Summary of current status of self-concept

Factors

M SD

Specialized subjects 3.48 0.66

Stress tolerance 3.58 0.66

Physical strength 3.42 0.84

Friends 3.48 0.71

Relationship with parents 3.70 0.70

Career planning 3.23 0.75

After obtaining the mean scores of the various self-concept factors, this study used repeated measures of ANOVA to analyze the level of importance attached to those factors, in order to determine whether there were any significant differences among them. As shown in Table 5.7, the F value was statistically significant, suggesting that there was a significant difference among various factors of self-concept. This study further used the least significant difference (LSD) to conduct a post hoc comparison.

Table 5.7 Summary of ANOVA on self-concept factors

Source of Variation

SS df MS F

Subject SSs 173.56 4.66 37.28 92.76*

Independent Variable SSa 2628.00 6541.43 0.40

Error Variable SSsa 1779.15 1405.00 1.27

*p< 0.05

According to the LSD post hoc comparison, the level of importance of various self-concept factors is shown in Table 5.8. The students are most concerned about the relationship with their parents, followed by stress tolerance, specialized subjects, friends, and physical strength. The level of importance of career planning is the lowest among the factors.

Table 5.8 Summary of LSD post hoc comparison on self-concept factors

5.3.2 Current status and difference of social support factors

The purpose of this analysis was to understand the current status of factors of social support. This scale contained three factors, including families, teachers, and peers. The mean and standard deviation of the factors are shown in Table 5.9. As seen, the mean score for families was 3.87, that for teachers was 3.39, and that for peers was 3.92.

Table 5.9 Summary of current status of social support

Factors

M SD

Families 3.87 0.66

Teachers 3.39 0.62

Peers 3.92 0.62

After the mean scores of the social support factors were obtained, this study further used repeated measures of ANOVA to analyze the level of importance attached to them, in order to determine whether there are any significant differences among the factors. As shown in Table 5.10, the F value was statistically significant, suggesting that there was a significant difference among the various factors of social support. This study used LSD to conduct a post hoc comparison.

Table 5.10 Summary of ANOVA on social support factors

Source of Variation

SS df MS F

Subjects SSs 242.47 1.98 122.14 474.96*

Independent Variable SSa 717.25 2789.16 0.26

Error Variable SSsa 963.73 1405.00 0.69

*p< 0.05

According to the LSD post hoc comparison, the level of importance attached to the factors of social support is shown in Table 5.11. The students attached the highest level of importance to their peers, followed by their families. The lowest level of importance was attached to their teachers.

Table 5.11 Summary of LSD post hoc comparison on social support factors

Factors

M

5.4 Path Analysis on the Overall Model of Self-concept, Social Support, Academic Achievement, and Occupational Choice Intention

This study explored whether self-concept and social support would affect academic achievement and occupational choice intention. Structural equation modeling (SEM) was used to perform analyses and test whether the goodness-of-fit between the hypothetical model and the collected data is acceptable. This study intended to establish a correlation model to understand whether self-concept and social support would affect academic achievement and occupational choice intention. Various indices were used to test the model’sgoodness-of-fit and investigate the cause-and-effect relationship among various potential variables.

SEM was used to test the overall model on its the goodness-of-fit between the

theoretical hypotheses and the empirical data, and to further test the goodness-of-fit level.

The indices for the overall goodness-of-fit assessment, the absolute goodness-of-fit statistics, the incremental goodness-of-fit statistics and the parsimony goodness-of-fit statistics were mainly used as the bases for measurement and assessment.

Table 5.12 is a summary of the goodness-of-fit indices for the overall model. The absolute goodness-of-fit statistics for the overall theoretical model were: χ2 = 831.22 (p

= .00), GFI = 0.92, RMR = 0.03, and RMSEA = 0.08. The incremental goodness-of-fit statistics for the overall theoretical model were: AGFI = 0.89, NFI = 0.89, NNFI = 0.87, CFI = 0.90, RFI = 0.86, and IFI = 0.90. The parsimony goodness-of-fit statistics for the overall theoretical model were PNFI = 0.71 and PGFI = 0.72. The above indices showed that only thevaluesofχ2 ,AGFI,NFI,NNFI,and RFIfailed to meetthestandard,while the other seven indices all met the standard. The results met the requirement of majority rule as proposed by Huang (2007); namely, the model would be acceptable as long as at least half of the indices met the standard.

The results of the goodness-of-fit assessment showed that at least a half of the indices met the standard, suggesting that the goodness-of-fit between the overall hypothetical model (the model of self-concept, social support, academic achievement and occupational choice intention) and the empirical data was good. Therefore, there was no need to amend the model. Verification of the hypotheses for the SEM model, and the direct, indirect and total effects among the variables could therefore be investigated.

Table 5.12 Summary of the goodness-of-fit indices of the overall model



better (P≧αvalue) 831.22 (P=0.00)

Unacceptable

2 /

 Between 1~5 9.90

Unacceptable

GFI > 0.8 0.92

Acceptable

RMR At least < 0.1 0.03

Acceptable

RMSEA < 0.05: good; 0.05~0.08: excellent 0.08

Acceptable

AGFI > 0.9 0.89

Unacceptable

NFI > 0.9 0.89

Unacceptable

NNFI > 0.9 0.87

Unacceptable

CFI > 0.9 0.90

Acceptable

RFI > 0.9, 0.95 perfect goodness-of-fit 0.86

Unacceptable

IFI > 0.9 0.90

Acceptable

Test of parsimony goodness-of-fit

PNFI > 0.5 0.71

Acceptable

PGFI > 0.5 0.72

^Acceptable