Chapter 4 Research result & data analysis
4.1 Research subject & Data analysis
This part would be bringing the description of the survey information as well as analytic methods utilized for breaking down collected data. Thus, collected data would be analyzed with several different analysis, namely, reliability, validity, descriptive analysis, Pearson correlation test, paired sample-t test for examining the first three hypotheses. Yet, the fourth method of regression analysis would be implementing to test on the fourth hypothesis. All analyses would be running through IBM SPSS statistics 22 for processing tool produced by IBM Company.
4.1.1 Research subject
This research aims at current and resigned employees from 85℃ coffee shops around Taiwan districts from which to randomly select subjects to advance the questionnaire. On the aspect of data analysis, I firstly use SPSS statistical software published by IBM Company in order to test both the reliability and validity of the questionnaire which is essential for the designing to be efficient and trust-worthy in terms of research accuracy and professionality.
4.1.2 Sampling method & estimated sample size
The approach to test of this questionnaire is to execute cluster random sampling method in order to extract test sampling subjects. By following the indicated unit of number of people, I randomly selected 35 subjects and successfully retrieved in a total of 31 effective questionnaires, which the retrieving rate is estimated 88%. The result is quite satisfying because the pre-estimated number of questionnaire in advance is
32
anticipated to approximately 25 effective survey data in accordance with the chosen 35 test subjects.
4.1.3 Reliability
This section would be breaking down the data analysis on 3 scale of questionnaire in order to test the reliability as well as validity for professionality and accuracy on question design. The research uses Cronbach’s α coefficient to measure the consistency between questions in questionnaire. According to the suggestion by Cuieford (1965), when α value becomes bigger than 0.7, it is regarded as high reliability, when α value sets between 0.7-0.35, the reliability is acceptable, when α value becomes lower than 0.35, this factor should be excluded (Cuieford, 1965). Therefore, it is a necessity for a well-designed questionnaire to be possessed high reliability by examining through the test and deleting low reliable questions in order to increase the consistency of questionnaire.
4.1.3.1 Reliability on work value scale
According to the pre-test of the first 5 data collected from the questionnaire, the Cronbach’s α value appears to be 0.912. All 10 questions are proved to be effective which all of them would be contain in the work value scale. The organized chart is shown as following:
33
Case processing summary (Proportion: All variables)
Table 6. Reliability of work value scale
N %
Cases
valid 5 100.0
Excluded 0 0
Total 5 100.0
Statistics of reliability
Table 7. Statistics of reliability on work value scale
Cronbach’s Alpha item
.912 10
4.1.3.2 Reliability on job engagement scale
According to the pre-test of the first 5 data collected from questionnaire, the Cronbach’ α value is 0.818, which also represents the reliability in job engagement scale is high and effective. Yet, all 10 questions in the scale would be included in job engagement part. The organized chart is shown as following:
Case processing summary (Proportion: All variables)
Table 8. Reliability of job engagement scale
N %
Cases
valid 5 100.0
Excluded 0 .0
Total 5 100.0
Statistics of reliability
Table 9. Statistics of reliability on job engagement scale
Cronbach’s Alpha item
.818 10
34
4.1.3.3 Reliability on turnover intention scale
According to the pre-test of the first 5 results data collected from questionnaire, the Cronbach’s α value is 0.677, which also indicates the high reliability of questions in turnover intention scale. The organized chart is shown as following:
Case processing summary (Proportion: All variables)
Table 10. Reliability of turnover intention scale
N %
Cases
valid 5 100.0
Excluded 0 0
Total 5 100.0
Statistics of reliability
Table 11. Statistics of reliability on turnover intention scale
Cronbach’s Alpha item
.677 4
In conclusion, the entire 24 questions designed within the questionnaire have high reliability, which Cronbach’s Alpha values are 0.912, 0.818 and 0.677, which are all higher than recommend Cronbach’s Alpha value 0.6 stated by Fornell and Larcker (1981) through the pre-test analysis of SPSS statistics software from the first 5 test subjects (Fornell & Larcker, 1981). Even with the entire 31 effective questionnaire results, the reliability of the 3 main scales are respectively estimated into 0.837, 0.876 and 0.803 which the reliability are higher than the results from pre-test.
4.1.4 Validity
In order to test the validity of the designed questions, Pearson product-moment correlation coefficient test is utilized at the beginning for finding out the correlation between the 3 constructs. For the construct validity, 3 new variables of the main
35
constructs are inserted, which are work value, job engagement and turnover intention.
Yet, input the variables into Pearson test through SPSS software and finally the correlation are calculated as following:
Table 12. Pearson correlation test
Work value Job engagement Turnover intention Pearson
Work value 0.000 0.286
Job
From the previously mentioned reliability of the 3 main constructs, they are known as 0.837, 0.876 and 0.803 Cronbach’s Alpha values, which all of them are bigger than 0.7 to have convergent validity. Moreover, all of the reliability value (α) are bigger than the correlation coefficient (r) which represent that the relationship between the 3 constructs have discriminant validity.
Overall, since the variables are estimated to have convergent validity as well
36
discriminant validity, construct validity is also acquired, which provides the evident to the validity of questions in questionnaire.
4.1.5 Descriptive Analysis
This research utilizes frequency distribution, percentage, average and standard deviation so as to execute descriptive analysis. Aiming at breaking down and better knowing the original basic information of the questionnaire which the basic features and actual condition of retrieving questionnaire relating to results of test subjects would be concisely presented into descriptive approaches of showing the data situation.
4.1.5.1 Personal attributes
The distribution relating to population statistics variables are listed as following with description:
1. Gender: According to the results of frequency distribution chart, it is shown that the retrieving samples of female test subjects are slightly higher than that of male test subjects, which the former variable has 54.8% (17 samples) while the latter has taken up 45.2% (14 samples) among the 31 samples.
2. Age: According to the result, the distribution of age of the research goes as the ages of test subjects lower than 17 appear to be 0%, ages between 18-25 account for up to the majority of 77.4% (24 samples), which indicates that employees between age 18-25 have the higher inclination of engaging into occupation in 85℃coffee shops compare to that of other ages. Test subjects between age 26-38 account for around 19.4% (6 samples), and only 3.2% (1 sample) of the test subjects are within the range between age 39-45. The overall distribution has the indication that people who were born in age 18-38 engage in 85℃coffee shops, that is , the generation of millennials.
3. Marriage: According to the distribution chart and the statement claimed in the
37
age section, it could be indicated that millennials are the concentrated human resource in the service industry like 85 ℃ coffee chain stores. Therefore, percentage of the unmarried test subjects accounts for up to 87.1% (27 samples), which seems coherent with the phenomenon of millennials engaging into 85℃ coffee shops. Besides, around 12% (4 samples) of employees are married according to the data results.
4. Educational level: The majority of test subjects are centralized in college level of education, which accounts for about 64.5% (20 samples). This could also indicate that millennials who engage in 85℃coffee shops are mostly college students. The second largest proportion in this section is the people received education level at graduate school, which takes up approximately 25.8% (8 samples). Thirdly, people who received the lowest education level remains at 9.7% (3 samples) which is at the level of high school. As a result, the main educational level of employees working at 85℃lies in the level of college which could further indicate that more and more millennials within the educational range of college have the tendency to spare effort both into work and study spontaneously.
5. Service seniority: The survey result shows that the majority length of employees contributing into 85℃coffee shops lies in seniority of 2 years or above which takes up to 48.8% (15 samples). Moreover, seniority length under half year is considered as secondary most distributed, 22.6% (7 samples).
Length between half year to 1 year accounts for 16.1% (5 samples) and lastly, seniority length ranged from 1 year to 2 years takes up around 12.9% (4 samples).
6. Position: In terms of position, most of the employees are working in 85℃
coffee shops as part-time job which are perhaps college students. As the result
38
shown, 67.7% (21 samples) of employees are in the position of working part-time and viewing from this outcome, it could be indicated that employees whose educational level at college hold the mindset or value of pursing engaging in part-time occupation while studying. For the position of full-time, there is also an indication that merely 25.8% (8 samples) of the survey employees are perhaps regarding the occupation of 85℃coffee shops as their
“real” job compare to the majority of employees under the position of part-time. Lastly, there comes to 6.5% (2 samples) associating with the position like chief store manager or assistant store supervisor and so forth. Overall, it could be predicted that the occupation of 85℃is seen as more like a part-time job for millennials domestically in the modern society which is also a common phenomenon regarding to service industries like other similar coffee or bubble tea shops.
Table 13. Descriptive analytic chart of personal attributes
Research construct Category Frequency Percentage (%)
Gender
39
Graduate school or above Total
4.1.5.2 Workvalue construct
In accordance with the work value scale mentioned in scale design section, the research utilizes the scale model originated by Wu (吳鐵雄，1996), who categorized work value construct into 8 types listed as following shown in the figure 18. Therefore, this part aims to calculate the minimum value, max value, average value and standard deviation through descriptive analysis according to the provided constructs attached to each questions in questionnaire. The organized chart is given down below:
40
Table 14. Descriptive statistics 1 Work value constructs
(followed by questions)
N Min. Max .
Average Standard deviation
1. Self-growth 31 2.00 5.00 4.1290 0.80589
2. Self-actualization 31 2.00 5.00 4.0968 0.78972
3. Self-dignity 31 3.00 5.00 4.0968 0.65089
4. Dignity 31 2.00 5.00 3.7742 0.84497
5. Social interaction 31 2.00 5.00 3.7097 0.78288
6. Freedom from anxiety 31 1.00 5.00 3.3548 0.95038
7. Traffic 31 2.00 5.00 4.2581 0.77321
8. Social relationship 31 3.00 5.00 4.3871 0.76059
9. Economic stability 31 2.00 5.00 4.2258 0.84497
10. Leisure health 31 3.00 5.00 4.6774 0.54081
Valid N (listwise) 31
By observing and focusing on the average and standard deviation in the chart, it is shown that the highest score with 4.6774 in terms of average segment is question 10 relating to leisure health and also got the lowest standard deviation score with 0.54081 which could be indicated that most of the employees (test subject) engaging into 85℃
coffee shops strongly agree the comprehensive labor and health insurance offered by the company or organization with high satisfaction and yet, the result is distributed in a highly centralized degree. The second highest average score goes to question 8 relating social relationship within the company which has the score of 4.3871 and the standard deviation score of 0.76059. Therefore, the construct could be indicated that most employees agree with the harmonious atmosphere and social relationship between colleagues are satisfied or they enjoy indulging in such particular working environment
41
which workers respect and support each other in terms of working tasks. Moreover, question 7 relation to traffic are also considered as importance for employees working in 85℃coffee shops whose has an average score of 4.2581 and standard deviation score of 0.77321. Employees also have a certain agreement on the comfortable working environment offered by 85 ℃ coffee shops as well as the location convenience associating with traffic issue; Economic stability has the average score of 4.2258 and standard deviation score of 0.84497 which both scores disclose employees place an average-above agreement on being satisfied with their currently paid wages or salaries.
However, the distribution appears to be quite dispersed according to the score provided which thus indicates some employees are satisfied with the financial stability given by 85℃ coffee brand organization which could be related to employees with position of full-time. On the other hand, it could also be indicated that some part-time employees are not quite satisfied with salaries received engaging in 85℃coffee shops because part-time shift are assigned with instability according to part-part-time employees’ schedule or plan; Self-growth has an average score of 4.1290 and standard deviation score of 0.80589. The scores could be indicated that employees has a certain degree of gaining brand new skills or knowledge while contributing into tasks under the brand’s shops which the working experience allows them to excel their personal improvement; Both question 2 and 3 have the same average score of 4.0968 but different on standard deviation scores, which the former has the score of 0.78972 and the latter scoring 0.65089. Both constructs have the under-average score which could be indicated that employees do not quite agree with the matter of working in 85℃coffee shops could achieve self-actualization as well as self-dignity; Question 4 relating to dignity has the average score of 3.7742 and standard deviation score of 0.84497 which not quite many employees agree with being rewarded or complimented by colleagues after
42
accomplishing excellent tasks in 85℃organization. However, a few employees might have received the experience associating a sense of dignity relating to similar matters within the organization according to the slightly high score on standard deviation;
Question 5 relating to social interaction with people within the organization has scored at a low 3.7097 and the standard deviation score of 0.78288. This data result indicates that employees engaging in 85℃organization seldom agree with working in such service industry does not required the responsibility to contribute to public or society;
Last but not least, question 6 has the lowest average score of 3.3548 and the standard deviation score of 095038. The data outcome could be indicated that very few employees have an agreement on indulging in occupation within 85℃coffee shops could effective avoid anxiety or pressure.
4.1.5.3 Job engagement construct
The chart listed below shows the organized descriptive analysis in accordance with job engagement scale introduced by Kanungo (1982). Since the scale does not required any specific constructs, both questions and the scale as a construct are respectively calculated. Therefore, the overall average value is calculated at the score of 3.9580 and the standard deviation score of 0.63444. The result indicates that employees focusing on millennials engaging in 85℃fail to perform a highly involving job engagement and thus, the distribution condition is shown with minimal difference which the standard deviation score is nowhere near value 1. The descriptive analysis chart relating to job engagement scale is shown as following:
43
Table 15. Descriptive statistics 2 Job engagement questions N Min. Max
.
Average Standard deviation
Question 1 31 3.00 5.00 4.3871 0.71542
Question 2 31 2.00 5.00 4.1613 0.93441
Question 3 31 1.00 5.00 3.4516 1.12068
Question 4 31 1.00 5.00 3.7097 1.03902
Question 5 31 1.00 5.00 3.6452 1.05035
Question 6 31 3.00 5.00 4.1290 0.71842
Question 7 31 3.00 5.00 4.2581 0.72882
Question 8 31 2.00 5.00 4.3226 0.83215
Question 9 31 2.00 5.00 3.6452 1.01812
Question 10 31 2.00 5.00 3.8710 0.95715
Total 31 2.50 5.00 3.9580 0.63444
Valid N (listwise) 31
4.1.5.4 Turnover intention construct
According to the adopted turnover intention scale design by Huang (黃開義 1984), the turnover intention questions are all descriptive data analyzed into the following results shown down below. All four questions have standard deviation inclined to value 1 which mean the distribution appears to be quite centralized. Yet, the average standard deviation has the score of 0.92159 which also literally represents concentrated distribution according to this question scale. From viewing question number 4 as representation of the scale, who has the highest average value of 3.8065 indicates the high intention of leaving or resigning working in 85℃after confirming to gain other job opportunities and the distribution extension is also quite concentrated. The descriptive
44
analysis of turnover intention chart is listed as following:
Table 16. Descriptive statistics 3
Turnover intention questions N Min. Max. Average Standard deviation
Question 1 31 1.00 4.00 2.8065 1.04624
Question 2 31 1.00 4.00 2.5161 1.02862
Question 3 31 1.00 5.00 3.3226 1.27507
Question 4 31 1.00 5.00 3.8065 1.27591
Total 31 1.00 4.50 3.1129 0.92159
Valid N (listwise) 31
4.1.6 Pearson correlation coefficient test
This section would be using the Pearson correlation coefficient test in order to test the correlation between constructs such as personal attributes, work value, job engagement and turnover intention from the perspective of linear correlation and what is more, to evident the hypotheses proposed in chapter 2 which especially aiming at the correlation between turnover intention (dependent variable) and personal attributes of age as well as educational level (independent variables). Moreover, hypothesis relating to whether the relationship between work value held by millennials and turnover intention have significance is also taken into the test for further verification.
4.1.6.1 Pearson test: Hypothesis 1
According to the Pearson test result relating the relationship between work value held by millennials and turnover intention, the correlation does not have enough evidence to support the proposed hypothesis which the correlation appears to be not significant with a correlated Pearson’s R-value of -0.198 and the two-tailed P-value of 0.286. Although the two variables have modest degree of correlation, the reason leading
45
to the unanticipated data outcome could be the problem of designing the questionnaire.
However, the questions are cautiously developed and confirmed in terms of reliability as well as validity through pre-test. Viewing from different aspect, the cause of the data result could be attributed to the samples collected in the questionnaire, which is rather low at a number of 31 samples in total to test out proposed hypothesis.
Table 17. Pearson correlation 1
Work value Turnover intention Pearson correlation Work value
Turnover intention -0.198
Sig. (two-tailed) Work value 0.286
Turnover intention 0.286
N Work value 31 31
Turnover intention 31 31
N: 31; *p<0.05 ; **p<0.01 ; ***p<0.001 4.1.6.2 Pearson test: Hypothesis 2
According to the shown result, independent variable of gender has highly significance to dependent variable of turnover intention. The Pearson’s R-value is calculated at -0.423 with the two-tailed significance p value of 0.018, which the number is smaller than the standard P-value of 0.05. Therefore, the correlation associating with variables of gender and turnover intention evidently supports the second hypothesis proposed in chapter three which the test result indicates that female employees has lower turnover intention compared to male employee in terms of engaging in 85℃
coffee shops mentioned in literature review.
46
Correlation
Table 18. Pearson correlation 2
Gender Turnover intention
Pearson correlation Turnover intention -0.423*
Sig. (two-tailed) Gender 0.018
Turnover intention 0.018
N Gender 31 31
Turnover intention 31 31
N: 31; *p<0.05 ; **p<0.01 ; ***p<0.001
4.1.6.3 Pearson test: Hypothesis 3
According to the result of Pearson analysis, variables of educational level and turnover intention relating to engaging in 85℃have significant correlation. This result indicates that the higher educational level employees of 85℃received, the higher turnover intention it becomes in terms of engaging in 85℃organization. According to the test subjects, employees who are receiving or received at low educational level of junior high or senior high school have high possibility of easily resigning the occupation in 85℃coffee shops because the millennials at this stage possess strong intention of seeking their ideality or dreams eager to pursue which also correlates with their unique characteristics of achievement oriented. The chart is listed as following:
47
Correlation
Table 19. Pearson correlation 3
Educational level Turnover intention
Pearson correlation
Turnover intention 0.585**
Sig. (two-tailed) Educational level 0.01
Turnover intention 0.01
N Educational level 31 31
Turnover intention 31 31
N=31; *p<0.05; **p<0.01; ***p<0.001
4.1.7 Paired sample t-test
This section would be utilizing paired sample t-test in order to test out the correlation by observing whether the calculated P-values of two variables are significant or not. Thus, the T-value would also be observed for significance which the result could be indicated that whether the samples are correlated or inter-influenced.
4.1.7.1 Paired sample: Hypothesis 1
In the analysis, there are 31 employees involved with their work values and turnover intention. Variable of work value has the mean of 40.7097 and standard deviation value of 4.97456 held by employee of millennial generation. Besides, variable of turnover intention has the mean value of 12.4516 and standard deviation value of
In the analysis, there are 31 employees involved with their work values and turnover intention. Variable of work value has the mean of 40.7097 and standard deviation value of 4.97456 held by employee of millennial generation. Besides, variable of turnover intention has the mean value of 12.4516 and standard deviation value of