This chapter presents the statistical results of the study, and the data were analyzed using SPSS 15.0. Additionally, in the discussion section, the results would be compared with previous studies to see whether there were differences between them.
Demographic and Descriptive Statistics
As mentioned before, THEDS surveyed all the bachelor’s and master’s graduates, and there were a total of 100,606 questionnaires returned. However, not all the data were valid and required. First, the missing data were omitted, and 95,969 cases remained. Then, the sample required in the study was extracted as follows:
1. Not took in-service education program or in-service training
2. Majored in the engineering and technology, which included mathematics and computation
3. Worked full-time, and not run one’s own business Thus, a total of 2247 questionnaires were valid.
Demographic
The demographic of the sample can be found in Table 4.1. The following is the brief description.
Gender
The number of the female is much more than that of the male. The number of the male is 920 (40.9%), whereas the number of the female is 1327 (59.1%).
Marital status
Most of the respondents were single (N=2180; 97.0%), whereas only 66 respondents (2.9%) were married. And the respondents with other marital status were excluded while running regression.
Educational background
There were 1770 respondents with bachelor degree (78.8%), and 477 respondents with master degree (21.2%). Among those with bachelor degree, 26.9% of the respondents were university or college graduates, 25.7% of them were two-year technical college graduates, and 47.4% of them were four-year technical college graduates. In addition, respondents attended public schools were less than those attended private schools.
Employment status
Eighty percent of the respondents worked in the industry, 8.3% of them worked in the school, 3.2% of them obtained government employment, 1.1% of them got military employment, 2.0% of them worked in the non-profit organization, and 5.3% of them gained other types of employment. Of the 1798 respondents who worked in the industry, 2.1% of the respondents worked in the publicly owned corporations, and 97.9% of them worked in the private corporations. Further, among the respondents worked in the school, 55.9% of them worked in the public school, and 44.1% of them worked in the private school. Thus, the number of respondents worked in the public sector is 240 (10.7%), while that of respondents worked in the private sector is 1842 (82.0%).
Table 4.1.
The profile of the respondents (N=2247)
Demographic Number of Respondents Percentage (%) Gender
School that respondents worked in
Public school 104 4.6
Private school 82 3.6
Types of economic sectors by ownership
Public sector 240 10.7
Private sector 1842 82.0
Voluntary sector 46 2.0
Other sector 119 5.3
Descriptive statistics
As mentioned above, there were two items to evaluate employment-education mismatch (overeducation and mismatch). Firstly, overeducation can be measure by the additional years of education, which is the difference between the respondent’s schooling years and the required schooling years for the job. Secondly, mismatch also can be assessed by the relatedness between work and major.
If an individual’s additional year is lager (smaller) than zero, one is classified as overeducated (undereducated), whereas if one’s additional year is equal to zero, one is properly educated. Table 4.2 contains the distribution of respondents being overeducated.
Master’s graduates were more likely to be overeducated. However, the mean additional years of overeducation of bachelor’s graduates are significantly larger than that of master’s graduates (see Table 4.3).
On the other hand, Table 4.4 shows that 58.3% of participants obtained jobs which were not or somewhat related to their majors, while 41.7% of them gained jobs which were mostly or closely related to their majors. Most of them had jobs somewhat related to their majors.
Additionally, the percentage of bachelor’s graduates having jobs not and somewhat related to their majors is more than 60%, and is larger than that of master’s graduates. As seen in Table 4.5, the mean of the relatedness between employment and education is shown, and the mean for master’s graduates is significantly larger than that for bachelor’s graduates. Hence, master’s graduates usually find jobs more related to their majors compared that of bachelor’s graduates.
Overall, most bachelor’s graduates might find jobs somewhat related to their majors, while most master’s graduates might find jobs that require less education. However, the extent of overeducation for bachelor’s graduates is greater than that for master’s graduates.
Table 4.2.
Overeducated: Frequency of the sample (N=2247)
Properly educated Undereducated Overeducated n Percentage (%) n Percentage (%) n Percentage (%)
Total sample 1286 57.2 34 1.5 927 41.3
Males 552 60.0 17 1.8 351 38.2
Females 734 55.3 17 1.3 576 43.4
Single 1250 57.3 34 1.6 896 41.1
Married 35 53.0 0 0 31 47.0
Bachelor’s graduates 1056 59.7 33 1.9 681 38.5
Master’s graduates 230 48.2 1 .2 246 51.6
Public school 301 61.9 8 1.6 177 36.4
Private school 985 55.9 26 1.5 750 42.6
Public sector 155 64.6 4 1.7 81 33.8
Private sector 1058 57.4 25 1.4 759 41.2
Voluntary sector 23 50.0 2 4.3 21 45.7
Other sector 50 42.0 3 2.5 66 55.5
Table 4.3.
Analysis of educational attainment to the additional years of overeducation (N=927)
M SD t
Total sample 2.76 1.10
Bachelor’s graduates 2.97 1.15 10.511*
Master’s graduates 2.16 0.60
Note: The range of mean for bachelor’s graduates is from 2 to 7, whereas the range of mean for master’s graduates is from 2 to 9.
*p<.05.
Table 4.4.
Percentage of respondents working outside their majors (N=2247) Not related Somewhat
Analysis of educational attainment to match between work and majors (N=2247)
M SD t
Total sample 2.43 0.999
Bachelor’s graduates 2.31 0.996 -11.162*
Master’s graduates 2.87 0.883
Note: The range of mean for bachelor’s graduates is from 1 to 4, and so is the range of mean for master’s graduates.
*p<.05.
As seen in Table 4.6, most respondents could find jobs within one and a half month.
Further, bachelor’s graduates significantly have longer unemployment time than master’s graduates do. Thus, the hypothesis two is confirmed.
Table 4.6.
Analysis of educational attainment to jobsearch duration (N=2247)
M SD t
Total sample 1.46 2.255
Bachelor’s graduates 1.56 2.350 4.096*
Master’s graduates 1.08 1.819 Note: *p<.05.
The results demonstrated that most of the participants reported that these nine work values were important to them. As shown in Table 4.7, the rank of the mean scores of the nine work values is as follows: compensation, security, benefit, advancement, work challenge, work location, work responsibility, work independence, and contribution to the society.
Except for work responsibility, work independence, and contribution to the society, the mean scores of other seven work values are above three; in other words, these three work values were somewhat unimportant to them. And the mean concern for security and work challenge were significant different between bachelor’s graduates and master’s graduates. Master’s graduates perceived that security and work challenge are more important to them.
Table 4.7.
Analysis of educational attainment to work values (N=2247)
M SD t
Compensation
Total sample 3.41 0.583
Bachelor’s graduates 3.40 0.586 -1.923
Master’s graduates 3.46 0.570
Benefit
Total sample 3.39 0.578
Bachelor’s graduates 3.38 0.580 -1.235
Master’s graduates 3.42 0.569
(table continues)
Table 4.7. (continued)
M SD t
Security
Total sample 3.41 0.586
Bachelor’s graduates 3.40 0.595 -2.319*
Master’s graduates 3.47 0.551
Work location
Total sample 3.04 0.720
Bachelor’s graduates 3.03 0.713 -1.478
Master’s graduates 3.09 0.744
Advancement
Total sample 3.16 0.670
Bachelor’s graduates 3.15 0.665 -0.725
Master’s graduates 3.18 0.692
Work challenge
Total sample 3.04 0.651
Bachelor’s graduates 3.02 0.659 -2.319*
Master’s graduates 3.10 0.618
Work responsibility
Total sample 2.96 0.622
Bachelor’s graduates 2.97 0.632 0.106
Master’s graduates 2.96 0.586
Work independence
Total sample 2.94 0.680
Bachelor’s graduates 2.94 0.681 -0.159
Master’s graduates 2.94 0.676
Contribution to the society
Total sample 2.75 0.775
Bachelor’s graduates 2.74 0.769 -0.606
Master’s graduates 2.77 0.797
Note: The range of mean for bachelor’s graduates is from 1 to 4, and so is the range of mean for master’s graduates.
*p<.05.
Hierarchical Regression Analysis of Employment-education Mismatch on Factors of Demographic, Work Values, and Jobsearch Duration
In this section, the additional years of overeducation and the relatedness between work and major could be predicted by these independent variables including demographic variables (gender, marital status, educational attainment, school types, and sector types), intrinsic work values (security, work challenge, work responsibility, work independence, and contribution to the society), extrinsic work values (compensation, benefit, work location, and advancement), and jobsearch duration.
The result show that the demographic predictors accounts for 12.4% of the variance in the additional years of overeducation. After controlling the demographic variables, the intrinsic work values accounts for 0.8% of the variance in the additional years of overeducation, while the extrinsic work values only accounts for 0.2% of the variance in the additional years of overeducation. As shown in Table 4.8, master’s graduates, people who worked in the private sector, and people who concern about work challenge were more likely have shorter additional years of overeducation. However, when one concerned about work independence, one had longer additional years of overeducation. In addition, jobsearch duration was not significantly related to overeducation.
On the other hand, females, bachelor’s graduates, and people attended private school were more likely to be mismatched, or to find jobs not or somewhat related to their majors.
People who concern about security, work challenge, and compensation probably could find jobs mostly or closely related to their majors, while people who concern about work independence and advancement were likely to find jobs not or somewhat related to their majors (see Table 4.9).
As can be seen in Table 4.10, the summary of the possible work values predicting employment-education mismatch (mismatch and overeducation) was demonstrated.
Table 4.8.
Hierarchical regression analysis for variables predicting the additional years of overeducation (N=927)
Table 4.9.
Hierarchical regression analysis for variables predicting the relatedness between work and major (N=2246)
Table 4.10.
Summary ofthe possible work values predicting employment-education mismatch
Dependent variables Work value
Additional years of overeducation
Shorter additional years Work challenge Longer additional years Work independence Relatedness between work and major
High relatedness Security, work challenge, and compensation Low relatedness Work independence and advancement
Logistic Regression Analysis of Employment-education Mismatch on Factors of Demographic, Work Values, and Jobsearch Duration
In this sector, a logistic regression was used to study the factors that might influence the possibility of being overeducated and being mismatched. And these factors contained demographic variables (gender, marital status, educational attainment, school types, and sector types), intrinsic work values (security, work challenge, work responsibility, work independence, and contribution to the society), extrinsic work values (compensation, benefit, work location, and advancement), and jobsearch duration.
The result indicated that females, master’s graduates, or people attended the private school were more likely to be overeducated (in terms of the actual educational level being higher than the educational level required for the job). Yet, people who worked in the public sector or private sector were more likely to be properly educated. Further, people who concern for advancement were likely to be overeducated, while people who concern for work challenge were likely to be adequately educated. In addition, the relationship between jobsearch duration and overeducation was positive, but it was not significant (see Table 4.11).
Likewise, females and people who attended the private school were likely to be mismatched (between job and major), and people who worked in the public sector or private sector were more likely to be properly matched. Nevertheless, as shown in Table 4.12, master’s graduates were more likely to be matched, that is they were more likely to find jobs mostly or closely related to their major. And there is a high possibility to be mismatched for people who concerned for work independence and advancement; whereas there is a high possibility to be matched for those who concerned for work challenge and compensation. In addition, the relationship between jobsearch duration and overeducation was positive, but it was not significant.
Also, the summary of the possible work values predicting employment-education mismatch was demonstrated in Table 4.13.
Table 4.11.
Logistic regression analysis for variables predicting overeducation (N=2212)
Level 1 Level 2 Level 3 Level 4
Table 4.12.
Logistic regression analysis for variables predicting mismatch (N=2246)
Level 1 Level 2 Level 3 Level 4
Table 4.13.
Summary of the possible work values predicting employment-education mismatch
Dependent variables Work values
Overeducation
Overeducated Advancement
Not overeducated Work challenge
Mismatch
Mismatched Work independence and advancement
Not mismatched Work challenge and compensation
Hierarchical Regression Analysis of Jobsearch Duration on Factors of Demographic Variables, and Work Values
A hierarchical regression was utilized to examine the hypothesis six, work values are significantly associated with the length of jobsearch duration. The length of jobsearch duration also could be predicted through these variables including demographic variables (gender, marital status, educational attainment, school types, and sector types), intrinsic work values (security, work challenge, work responsibility, work independence, and contribution to the society), and extrinsic work values (compensation, benefit, work location, and advancement).
The results show that jobsearch duration is significantly related to educational attainment (β=-.068, p<.05), school type (β=.050, p<.05), and public sector (β=.143, p<.05).
After controlling the demographical variables, preference for work challenge is significantly related to jobsearch duration (β=-.095, p<.05).
Thus, bachelor’s graduates, people who attended the private school, and people who worked in the public sector had longer jobsearch duration. Yet, people who concerned for work challenge had shorter jobsearch duration. Extrinsic work values were not significantly related to the length of jobsearch duration (see Table 4.14).
At last, the statistic results of six research hypotheses were demonstrated in Table 4.15.
Table 4.14.
Hierarchical regression analysis for variables predicting jobsearch duration (N=2246)
Level 1 Level 2 Level 3
B β B β B β
Gender 0.128 .028 0.128 .021 0.100 .022
Marital status -0.257 -.019 -0.257 -.017 -0.218 -.016 Educational
attainment -0.378 -.068* -0.378 -.065* -0.356 -.065*
School type 0.274 .050* 0.274 .051* 0.274 .050*
Public sector 1.041 .143* 1.041 .139* 1.016 .139*
Private sector 0.157 .027 0.157 .030 0.175 .030 Other sector 0.722 .045 0.722 .051* 0.810 .051*
Security 0.128 -.026 -0.096 -.025
Work challenge -0.257 -.095* -0.341 -.098*
Work
responsibility -0.378 .037 0.142 .039
Work
independence 0.274 -.018 -0.064 -.019
Contribution to
the society 1.041 .030 0.088 .030
Compensation -0.086 -.022
Benefit 0.083 .021
Work location -0.059 -.019
Advancement 0.044 .013
R2 .024 .032 .033
Adjusted R2 .021 .027 .026
F 7.854* 6.121* 4.698*
Note. *p<.05.
Table 4.15.
Statistic results of research hypotheses
Hypothesis Result
Hypothesis1: Educational attainment is significantly related to education-employment mismatch.
Hypothesis 1.1: Master’s graduates are significantly overeducated than bachelor’s graduates.
Hypothesis 1.2: Bachelor’s graduates are significantly mismatched than master’s graduates.
Hypothesis 1: Accepted.
Hypothesis 1.1: Accepted.
Hypothesis 1.2: Accepted.
Hypothesis 2: The jobsearch duration of master’s graduates is significantly shorter than that of bachelor’s graduates.
Hypothesis 2: Accepted.
Hypothesis 3: The perceived work values of bachelor’s graduates are significantly different from that of master’s graduates.
Hypothesis 3: Accepted.
Hypothesis 4: Jobsearch duration is significantly and positively related to overeducation.
Hypothesis 4.1: Jobsearch duration is significantly and positively related to overeducation.
Hypothesis 4.2: Jobsearch duration is significantly and positively related to mismatch.
Hypothesis 4: Rejected.
Hypothesis 4.1: Rejected.
Hypothesis 4.2: Rejected.
Hypothesis 5: The certain perceived work values are significantly and positively associated with
education-employment mismatch.
Hypothesis 5.1: The certain perceived work values are significantly and positively associated with
overeducation.
Hypothesis 5.2: The certain perceived work values are significantly and positively associated with mismatch.
Hypothesis 5: Accepted.
Hypothesis 5.1: Accepted.
Hypothesis 5.2: Accepted.
Hypothesis 6: Work values are significantly associated with the length of jobsearch duration.
Hypothesis 6: Accepted.
Discussions
There were six hypotheses in the study, and they were examined accordingly as follow.
And these results were compared with the previous reviewed literature as well.
Hypothesis1: Educational Attainment Is Significantly Related to Education-employment Mismatch.
There were two sub-hypotheses: (a) master’s graduates are significantly overeducated than bachelor’s graduates; (b) bachelor’s graduates are significantly mismatched than master’s graduates. Based on the results, when education-employment mismatch was evaluated by mismatch (between work and major), bachelor’s graduates are significantly mismatched than master’s graduates. However, when education-employment mismatch was assessed by overeducation (the difference between required schooling years and the actual schooling years), master’s graduates are more significantly overeducated than bachelor’s graduates. Although master’s graduates are more overeducated, the extent of the overeducation is greater among bachelor’s graduates.
Likewise, people with master degree were overeducated than people with bachelor degree (Wang, 2000). In addition, the result of this study consisted with Robst’s (2007) study that people with master degree or doctoral degree are less likely to be mismatched than those with bachelor degree.
Hypothesis 2: The Jobsearch Duration of Master’s Graduates Is Significantly Shorter Than That of Bachelor’s Graduates.
As assumed, the result of the study demonstrated that the jobsearch duration of master’s graduates is significantly shorter than that of bachelor’s graduates by nearly a half month.
The result was correspondent with the findings of the previous studies (Chen, 2008; Tao & Li, 2006).
Hypothesis 3: The Perceived Work Values of Bachelor’s Graduates Are Significantly Different From That of Master’s Graduates.
Generally speaking, both the bachelor’s graduates and master’s graduates perceived that most work values were critical to them. Master’s graduates, nevertheless, considered that security and work challenge are of vital importance among these work values. The result also showed that high educated people more emphasized on the intrinsic work values (Chiu, 1993;
Sinisalo, 2004). In addition, Li et al. (2008) expressed that high educated people work for the pursuit of self-fulfillment. This expression might explain why master’s graduates more concerned work challenge.
Hypothesis 4: Jobsearch Duration Is Significantly and Positively Related to Education-employment Mismatch.
According to the results, jobsearch duration was positively related to education-employment mismatch (overeducation and mismatch), and the result was consistent with Wang’s study (2000). The longer one’s jobsearch duration, the more likely one is overeducated or mismatched. However, the relationships were not significant.
Hypothesis 5: The Certain Perceived Work Values Are Significantly and Positively Associated With Education-employment Mismatch.
There were two sub-hypotheses: (a) the certain perceived work values are significantly and positively associated with overeducation; (b) the certain perceived work values are significantly and positively associated with mismatch. Those who cared about advancement or work independence were more likely to obtain jobs which require less education. In addition, for this group of people the extent of overeducation was greater. Contrarily, those who cared about work challenge were more likely to obtain jobs which meet with their education level. In addition, for this group of people the extent of overeducation was smaller.
On the other hand, those who concerned for work independence and advancement were more likely to be mismatched, while those who concerned for work challenge, compensation or
security might be matched.
As described, Bender and Heywood (2006) suggested that people might be willing to accept jobs that require lower educational level because of the perception of advancement or compensation. Likewise, the result of this study showed that people might be overeducated due to the emphasis of advancement. Nevertheless, people might be properly educated or look for jobs which meet their educational level due to the emphasis of compensation. Yet, the relationship was not significant.
Moreover, benefit was positively associated with overeducation and mismatch, while work responsibility and work location were negatively associated with overeducation and mismatch, but these relationships were not significant.
Hypothesis 6: Work Values Are Significantly Associated With the Length of Jobsearch Duration.
Concerning intrinsic work values, when people placed importance on security and work challenge, they probably might have shorter jobsearch duration. Further, preference for work challenge was significantly and negatively related to the length of jobsearch duration. When it comes to extrinsic work values, people with preference for compensation and work location might have shorter jobsearch duration, whereas those with preference for benefit and advancement might have longer jobsearch duration. Yet, the relationships were not significant. And the results were inconsistent with the explanation of Chen’s study (2002) that people might spend time finding a job with good pay.