In Chapter three, the content was composed of research framework, research hypothesis, research sample, research method, validity, reliability, instrumentation, data collection and data analysis.
Research Framework and Hypothesis
The conceptual model was listed in this study according to literature review. Therefore, the research framework of this research was depicted in figure3.1:
Figure 3.1 Research framework
There are independent variable (job stress), dependent variable (affective commitment) and moderator (social support) in this research framework (See figure 3.1). For independent variable, it is divided into role ambiguity; role conflict and role overload (Kahn & Byosiere, 1992). For dependent variable, according to Allen and Meyer (1991) study, organizational commitment was divided into three component model of commitment, including affective, continuance, and normative commitment. But, this study focuses on affective commitment.
On the other hand, the study investigates the relationship between the job stress and affective Social Support
H2
Affective Commitment Job Stress
(1)Role Conflict (2)Role Ambiguity
(3)Role Overload H1
Control Variables 1. Age 2.Gender 3 Educational Level 4 Seniority of Working
Experience
commitment. On the other hand, through the moderator of social support, the study also explores the relationship among the job stress, social support and affective commitment. The relationship among three was described in figure 3.1 and research hypotheses were listed below:
Table 3.1
Research Hypothesis Research Hypothesis
Hypothesis 1 Job stress is significant negatively related to affective commitment Hypothesis 1a – Role ambiguity is negatively related to affective commitment
Hypothesis 1b- Role conflict is negatively related to affective commitment Hypothesis 1c - Role overload is negatively related to affective
commitment
Hypothesis 2 There is a moderate effect of the social support on the relationship between job stress and affective commitment.
Hypothesis 2a –There is a moderate effect of the social support on the relationship between role ambiguity and affective commitment.
Hypothesis 2b –There is a moderate effect of the social support on the relationship between role conflict and affective commitment.
Hypothesis 2c –There is a moderate effect of the social support on the relationship between role overload and affective commitment.
Research Sample
According to the “Taiwan NPO Information Platform”, there are 6013 NPOs registered in this platform. The researcher collected the data of participants from varies of NPOs including 17 social welfare service foundations, four environmental protection foundation and one cultural education foundation.
For sampling method, the study used a non-experimental quantitative method (John &
Christensen, 2004). In order to examine the hypotheses, the respondents need to meet the following requirement:
The respondent has at least six months working experiences in nonprofit organizations in Taiwan
Based on above requirement,for pilot study, the study collected 30 copies of questionnaires from employees who have experienced working in the Taiwan’s nonprofit organizations. For the research sample, the study targeted to collect data from 156 employees who have at least six months working experiences in nonprofit organizations in Taiwan. The respondents would assure of confidentiality and informed that the important information would be used for research only.
Research Approach
The quantitative approach was used in the study. Except background information, all items in the questionnaire were adopted by previous studies (Meyer & Allen, 1991; Pineda, 2011;
Zhao & Rachid, 2010; Zimet et al. 1988 ). This questionnaire was divided into three scales: job stress, social support and affective commitment. Moreover, before conducting the formal study, the researchers have implemented the pilot study to make sure the quality of the instrument.
Control Variables
The study selected control variables from background information, including age, gender, and educational level and seniority of work experience. The purpose of using control variables is to better understand NPOs’ employees of relationship among job stress, social support and affective commitment,
Demographic Information
All items in the first part were divided into organizational and personal demographic information. The content of background information included the type and location of
organizations. organizational size, age , gender, job position, educational level, seniority of work experience, job content that employees are responsible for in the organizations, and what kind of training had received before in the organizations. Control variables were included within the background information.
Job Stress
The job stress (role ambiguity, role overload and role conflict) items in questionnaire were adopted from study conducted by Zhao and Rachid’s (2010) study. The part Ⅱ of the questionnaire was consisted of 16 items.The scale was rated from 1 (strongly disagree) to 5 (strong agree) (See table 3.2) .Reliability coefficient of job stress measure was also shown in table 3.2 (See table 3.2).
Table 3.2
Items and Reliability Coefficient of Job Stress Measure Construct Item
I feel certain about how much authority I have.
I know that I have divided my time properly.
I know that I have my responsibilities are.
I know exactly what is expected of me.
I know what the critical factor is in getting
There is a need to reduce some parts of my role.
I feel overburdened in my role.
I have been given too much responsibility.
My workload is too heavy.
The amount of work I have to do interfere with quality I want to maintain.
5 0.75
(Continued)
Table 3.2 (continued)
I have to do things that should be done differently.
I work under incompatible polities and guidelines.
I receive an assignment without the resources to complete it.
I have to buck (bend) a rule or policy in order to carry out an assignment.
I receive incompatible requests from two or more people.
5 0.78
Note: Adapted from The mediating role of work-leisure conflict of job stress and retention of it professionals by Zhao, L & Rashid, H (2010), Academy of Information and Management Sciences Journal, 13 (2), 31.
Social Support
All the social support 15 items in questionnaires were adopted from Zimet et al. (1988) and Pineda (2011). The part Ⅲ of the questionnaire was consisted of 12 items (See table 3.3).
Moreover, the study found that Pineda (2011) study additionally increase one dimensions (organizational support) in her questionnaire. Therefore, the other 3 items also included into the part III of questionnaire. Both Reliability coefficient of social support measure was also shown in table 3.3, table 3.4 and table 3.5. The scale was rated from 1 (strongly disagree) to 5 (strongly agree).
Table 3.3
Items and Reliability Coefficient of Social Support Measure Construct Item
I have a special person who is a real source of comfort to me.
I get the emotional help and support I need from my family.
My family really tries to help me.
I can talk about my problems with my
I can talk about my problems with friends.
I have friends with whom I can share my joys and sorrows.
I can count on my friends when things go wrong.
My friends really try to help me.
4 0.85
Table 3.4
Items and Reliability Coefficient of Social Support Measure
Construct Question Item Number
of
Support All three dimensions of question items 12 0.88
Note: Adapted from: “The multidimensional scale of perceived social support” by Zimet, G., Dahlem, N., Zimet, S. & Faley, G. (1988), Journal of Personality Assessment, 52(1), 30-41.
Table 3.5
Items and Reliability Coefficient of Social Support Measure
Construct Item
I feel that I can talk freely with my coworkers If I face any difficulties with my students I can depend on my school/organizations/company for help
3 0.826
Note: adapted from” The relationship of previous international experience, pre-departure cross cultural preparation, personality, social support and multicultural competence” Pineda, A. (2011), (Unpublished Master’s thesis), Graduate of IHRD, Taipei, 64-65.
Affective Commitment
The affective commitment items in questionnaire were adopted from study conducted by Zain and Gill’s (1999) study, which included 8 items. The scale was rated from 1 (strongly disagree) to 5 (strongly agree) (See table 3.6). Reliability Coefficient of Organizational Commitment Measure was also shown in table 3.6.
Table 3.6
Items and Reliability Coefficient of Affective Commitment Measure Construct Item
I feel as if this organization’s problems are my own.
I do not think I could become as attached to another organization as I am to this one.
Organization
I feel like “part of the family” at any organization.
I feel emotionally attached to this organization.
This organization has personal meaning to me. organizational commitment measure” by Zain, A. & Gill, R (1999), Malaysian Management Review. Retrieved from:
http://www.eurojournal.com/ejss_13_4_12.pdf.
Instrument Design of Questionnaire
Quantitative method was conducted in this research to investigate the relationship among job stress, social support and affective commitment.
Questionnaires were used to obtain the data from the different areas of employees in Taiwan’s nonprofit organizations. The research selected three variables of questionnaires which were developed by previous studies, and then revised them for this study. A job stress (role ambiguity, role overload and role conflict) items were adopted from study conducted by Zhao and Rachid (2010). The role ambiguity and role conflict items were initially developed by Glazer & Beehr’s (2005) study , and the role overload items were initially developed by Peterson et al. (1995). The affective commitment items were adopted from study conducted by Zain and Gill (1999). The instrument of affective commitment was adopted from organizational commitment developed by Meyer and Allen (1991). The social support items were adopted from Multidimensional Scale of Perceived Social Support (MAPSS; Zimet et al., 1988). This questionnaire includes three dimensions of perceived social support (1) Significant others (2) Family (3) Friends. Also, Pineda (2011) adopted this questionnaire and added the fourth dimension (Organization support). Therefore, the study will adopt fourth dimension of perceived social support to the questionnaire. Five-point Likert scale was applied. The scale was from 1 (Strongly disagree), 2(Disagree), 3 (Neural), 4 (agree) to 5 (strongly agree).
In addition to background information, in this study, the questionnaire included three parts, job stress, social support and affective commitment. Then, the researcher adopted three steps to build the final version of questionnaire. First of all, all items in the questionnaire were revised based on literature reviews in the study. Secondly, in order to test the content validity, expert reviews (expert validity) and peer reviews would be conducted in this study;
Finally, the study have conducted the pilot study to make sure that Cronbach alpha coefficient for all the items in the questionnaire reached an acceptable level (Nunnallly, 1978) The Design of Chinese Questionnaire Scale
Figure 3.2 The process of designing formal questionnaire
Based on above figure 3.2, the researcher initially checked the literature reviews and adopted the questionnaire scales which were conducted by prior studies in this research. After implementing the forward translation and back translation through peer reviews, the researcher needed to synthesize the questionnaire check, and took it to two experts adjusting questionnaire items. Therefore, the research also continued to examine the pilot study and adapt the questionnaire in order to build the formal Chinese questionnaire.
English Translation and Back Translation
All items in the questionnaire were written in English. The researcher conducted English translation and back translation and used five steps to adjust the questionnaire based on the literature reviews. The researcher initially translated English version into Chinese version.
After translating the items; the study continued to invite two English teachers to retranslate
mutual evaluation with bilingual versions. The fourth step is to use peer and expert reviews to examine the Chinese version of content in the questionnaire. Finally, the researcher conducted the pilot study for the Chinese version of questionnaires in this study.
Data Collection
For delivery of questionnaires, two ways of data collection were used to deliver questionnaires through envelopes and emails in this study. On the one hand, the researcher used convenience sampling to ask the nonprofit organizations for deciding to deliver them to the organizations under the permissions. On the other hand, the researcher used snowball sampling to contact with the employees of non-profit organizations to collect the data in the different types of Taiwan’s nonprofit organizations. 156 copies of the questionnaire were analyzed in this study.
Also, all the questionnaires were adopted completely by anonymous and confidential. After collecting the data, the researcher continued to code and transformed the received data into the statistic data through using the SPSS 19.0(See the figure 3.3).
Figure 3.3 Data collection and analysis procedure Select questionnaires and deliver the
questionnaires to the participants
Collect the data
Transform the received data into the statistic data (Coding)
Run the statistic analysis (SPSS 19.0)
For the process of preparation, the data was collected from March 16 to April 30. Before conducting this study, potential participants filled in the questionnaire which was asked about job situation of employees in nonprofit organizations. This questionnaire assured us to certify that the participants met the selection criteria and to gather additional data conducting subsequent analysis.
Subjects responded individually through paper printing and electronic files. The procedure took approximately 10 minutes for each subject. Also, the time of data collection occurred over one and half months, and received totally 176 copies of questionnaires.
The researcher prepared a package including the cover letter, copies of questionnaires, outline of this study, and sent these documents to the headquarter or branch of nonprofit organizations respectively. Participants were initially contacted by the Executive Director of the nonprofit organizations through phoning and informing them of the process of the nature of the study. Accordingly, each participant received a questionnaire including hardcopy or electronic version through post or email. Questionnaires were sent directly to the participants, and returned to the researcher through the stamped addressed envelope within two weeks. For some of collected questionnaires, the researcher also conducted the online questionnaire survey system to collect 26 copies of questionnaire through different channels of community websites.
There are 35 NPOs in the contact list. The researcher tried to contact with those workers of headquarters or local branches based on assistances of friends and “Taiwan NPO Information Platform. For final data collection, 22 NPOs agreed to assist researcher in conducting this quantitative research. It includes 17 social welfare service foundations, four environmental protection foundations and one cultural education foundation. Therefore, out of 260 copies of the questionnaires, 176 were received, ineffective copies were 20. Effective copies were 156. The return rate of questionnaire was 74.91%.
Validity and Reliability
The study chose initially the questionnaire scales that have been examined in previous research. Also, the research conducted content analysis through peer reviews and expert reviews in order to examine and enhance appropriateness of the item contents (Kerkinger &
Lee, 2000)
For reliability analysis, the research used the Cronbach’s alpha to test reliability on the questionnaire item. The qualified Cronbach’s alpha value in this study was .7 (Guieford, 1965). When the value of Cronbach’s alpha was higher than .7, the reliability of variables were accepted in this study.
Expert Review
The researcher conducted the expert review and revised the questionnaire items in this research. The researcher invited three experts to review the adequacy of the content. One is a HR professor, and the other two are NPO’s senior practitioners (CEO and social work mentor) who currently work in the nonprofit organizations. After reviewing each item by experts, there are two places that needed to check and readapt in this questionnaire. Firstly, the senior practitioners suggested that the mentoring can be substituted for coaching in this questionnaire. That is because the meaning is suitable for the practical area of nonprofit organizations. On the other hand, the HR professor adjusted the meaning for some of items in order to make the sentence more smooth and increase content validity in the questionnaire.
The suggestions of expert review were shown in the appendix B.
Peer Review
The study invited three people for peer review in English translation and back translation. Based on their educational background, one is foreign graduate student who majors in Human resources development, and the other two are local Taiwanese graduate students who major in foreign language and social work. First of all, the original questionnaire was reviewed and translated into the Chinese version by two local Taiwanese friends, and conducted English back translation again after Chinese translation. Accordingly, because of bicultural difference, the researcher took the revised questionnaire to the foreign peer reviewer who examined the meaning of accuracy on each item. This result of peer review was listed in the appendix C.
Pilot Test
Since each questionnaire items in the research has initially developed and conducted by prior researchers, and also the validity and reliability for the items was good in the prior srudies (Meyer & Allen, 1991; Pineda, 2011; Zhao & Rachid, 2010; Zimet et al. 1988).
However, due to difference of research area, the researcher must adopt the pilot test and examine on this questionnaire. Therefore, the researcher conducted a pilot study and adjusted the content of questionnaires. The sections of questionnaire were composed of 39 items in the pilot study, including job stress (16 items) social support (15 items) and affective commitment (8 items). The sample size of pilot test was 30 employees who come from six distinct of social welfare foundations in Taiwan.
All the Cronbach’s alpha value for each section in this questionnaire was acceptable(α > .7 ) in this pilot test (Nunnallly, 1978). (Job Stress = .785, RA=.806, RO= .922 and RC= .808), Affective Commitment = .818 and Social Support = .818, SOSC=.
834, FMSC= .884, FRSC2= . 898 and OSC= .819).
Independent Variable: Job Stress
The job stress section of questionnaire was consisted of 16 questions. For this section, and the result of pilot study was shown in the table 3.7
Table 3.7
Reliability Analysis: Job Stress (N=30)
Scale Number of Items Cronbach’s α for
Pilot Study
Role Ambiguity 6
(RA1-RA6)
.806
Role Overload 5
(RO7-RO11)
.922
Role Conflict 5
(RC12-RO16)
.808
Job Stress 16 .785
Each of three dimensions (Role ambiguity, Role overload and Role conflict) in the job stress section had Cronbach alpha of .806, .922 and .808 respectively, and the total of Cronbach alpha for the pilot test was .785. Therefore, the same items were kept for job stress section in the final questionnaire.
Dependent Variable: Affective Commitment
The affective commitment section of questionnaire was consisted of 8 questions. For this section, and the result of pilot study was shown in the table 3.8
Table 3.8
Reliability Analysis: Affective Commitment (N=30)
Scale Number of Items Cronbach’s α for
Pilot Study Affective
Commitment
8 (AC1-AC8)
.818
The affective commitment section was consisted of 8 questions. The total Cronbach alpha for pilot study was .818. After completing this section of pilot study, the same items were kept for affective commitment section in the final questionnaire.
Moderator: Social Support
The social support section of questionnaire was composed of 15 questions. For this section, the results was shown in the table 3.9
Table 3.9
Reliability Analysis: Social Support (N=30)
Scale Number of Items Cronbach’s α for
Pilot Study
Significant Others 4
(SOSC1-SOSC4)
.834
(Continued)
Table 3.9 (continued)
Scale Number of Items Cronbach’s α for
Pilot Study
Family 4
(FMSC5-FMSC8)
.884
Friends 4
(FRSC9-FRSC12)
.898
Organizational 3
(OSC13-OSC15)
.819
Social Support 15 .818
The section was consisted of four different types of support as well as significant others, family, friends and organizational support in the questionnaire. Each of four dimensions in the social support section had Cronbach alpha of .834, .884, .898 and .819 respectively. The total of Cronbach alpha for the pilot test was 0.818, therefore, for this section, the same items were kept for social support section in the final questionnaire.
Data Analysis
In the process the research used SPSS 19.0 to analyze the dada of questionnaires through different types of statistical analysis. The processes of data analysis included item analysis, reliability analysis, validity analysis, descriptive statistics, correlation analysis, hierarchical regression of analysis and independent sample T test.
Descriptive Statistics
In this step, this study reported standard variance and mean of each variable in the descriptive statistic. The researcher will use the information of descriptive statistics to infer the important meaning of final data.
Correlation Analysis
The research will conduct the correlation analysis to understand the correlations among job stress, social support and affective commitment. And Pearson coefficient will be used to test these correlations in order to continue to run next steps in process of statistic analysis.
Hierarchical Regression of Analysis
Multiple regression analysis will be used to evaluate the relationship between a set of independent variables and a dependent variable. Therefore, hierarchical regression analysis will be conducted to examine the relationship between independent variables and a dependent
Multiple regression analysis will be used to evaluate the relationship between a set of independent variables and a dependent variable. Therefore, hierarchical regression analysis will be conducted to examine the relationship between independent variables and a dependent