Chapter ⅢⅢⅢⅢ Methodology
Research Structure
Accounting measurements are considered as lagging indicators because they are just the result of management action and organizational performance, and do little to future organizational performance. In other words, most organizations emphasize financial performance too much so that they have an internal rather than an external focus. Based on this reason, the study did a survey of managers from insurance companies in Taiwan to develop key performance indicators on human resource management for insurance companies in Taiwan.
From the research of theses, literature review and other relevant study, most studies which related to performance evaluation mostly shown the relationship between some specific topics and performance evaluation, such as, training and performance evaluation, performance evaluation after companies merged, or other case studies. To the best of my knowledge, constructing an instrument of key performance indicators on human resource management for insurance companies in Taiwan has not been seen. Therefore, collect opinions from experts through Delphi study to review the whole KPIs developed previously will be conducted.
The questionnaire questions developed first by the relevant studies which discuss KPIs and HRM and to prove whether it is appropriately applied in insurance companies in Taiwan by Delphi technique. The first main categories need to be further developed are Labor Costs, Recruiting Indicator, Retention Indicator, HRD Indicator, Benefit Cost and Employee Relations. Figure 3.1 is the research structure of this study.
Figure 3.1 Research Structure KPIs
B. Recruiting Indicator
C. Retention Indicator
D. HRD Indicator
Delphi Technique with 12 experts
Test the instrument among insurance companies in Taiwan
An instrument of KPIs on HRM among insurance companies in Taiwan
E. Benefit Cost A. Labor Cost
F. Employee Relations Literature
Review
Research Methods
This study is going to construct an instrument of KPIs on HRM among insurance companies in Taiwan; therefore, research methods conducted literature analysis and Delphi to develop the study.
Literature Analysis
This study focused on the literature related to the development of insurance companies in Taiwan, the meaning and importance of key performance indicators, human resource management. It also gave an overview of the relevant dissertations and journals about these theories, research methods, processes and conclusions; this served as basis for this study. The sources of literature are listed below.
1. Journals and books.
Journals and books are related to the theory and development of insurance industry, KPIs, HRM, and performance measurement, and research method.
2. Domestic document and publication from government.
The document and publication are about the public data of domestic insurance companies.
3. Theses and dissertations.
Theses and dissertations are the relevant study about KPIs, the development of insurance industry and the study applying Delphi technique.
Delphi
The Delphi technique, which is used to measure the judgment of a group of experts, is a method of generating ideas and facilitating consensus among individuals who do not meet and who may be geographically distant (Polit & Beck, 2004). It is
useful for achieving consensus in areas lacking empirical evidence (Powell, 2003).
Besides, the Delphi technique has several advantages. It is considered time and cost efficient and, in the second and third survey rounds, enables participants to consider the opinions of other experts (Hardy et al, 2004). Because direct contact among experts is avoided, bias based on personality traits, seniority, and experience is minimized (Jairath & Weinstein, 1994). Therefore, the Delphi technique helps the group reach a consensus of opinion by incorporating all available data and disseminate those data among all participants. As for the quantity of experts, Tersine and Riggs (1976) recommended between 10 and 15 experts, provided that the group is homogeneous.
Besides, in order to avoid disturbance of group discussion, such as group polarization, groupthink and the bandwagon effect, the research method, Delphi is thought of the most appropriate one for this study. Since this group of experts is anonymous, and experts can communicate without seeing each other, the result of responses form questionnaire can be synthesized and highly possible to predict future circumstance. 12 mangers from HR department in insurance companies in Taiwan and which also are the suggested number of experts of Delphi.
Research Process
Based on research structure and methods, the research process is mainly as follows,
1. Relevant literature collection
Collect literature related to insurance company, KPIs and HRM to analyze them.
2. Identification of research topic and purpose
The research direction and scope were further clarified through discussing with advisor and literature review.
3. Literature review and analysis
Collect and analyze relevant literature to make them as the foundation of this study. Questions of interview will be mostly based on literature review.
4. Develop questionnaire
After collect relevant data through literature review and analysis, a questionnaire will be developed based on the literature review and pilot study.
The questions are KPIs which require Delphi experts to identify the importance of them by Likert 5-point, such as HR cost rate, Recruiting time.
5. First stage of questionnaire delivery Deliver the questionnaire to 18 experts.
6. Second stage of questionnaire delivery
After three weeks later, deliver the questionnaire to 12 experts based on the responses of first stage questionnaire delivery
7. Third stage of questionnaire delivery
Three weeks later, deliver the questionnaire to 12 experts based on the responses of second stage questionnaire delivery
8. Develop the instrument
Develop the instrument based on the data of response from questionnaire 9. Test the instrument
Test the instrument to 12 Delphi experts by developing the fourth round of questionnaire with questions ensure the instrument is understandable and appropriate among insurance companies in Taiwan.
10. Make conclusion and suggestion
According to literature analysis and questionnaire response from experts, make specific conclusion and suggestion.
The detail research process is as figure 3.2
Delphi Relevant literature collection
Literature review and analysis
Develop the instrument
Three stages of questionnaire delivery Peer Review of the questionnaire
Test the instrument
Make conclusion and suggestion Identification of research topic and purpose
Develop the questionnaire
Figure 3.2 research process
Sample
Delphi Sample
The researcher acquired Delphi sample by the name book of human resource management association (HRMA) of 2005, and pick up the ones who work in insurance companies. After having the list, the researcher conducted phone interview to check the correctness of the information and their willingness of completing questionnaire for this study.
There are 12 participants on the expert panel for this study, the recommended number meets the reliability; all of them are above Sr. supervisor level in insurance companies in Taiwan assure the validity of the sample. As for the background of the experts, 4 of them are with master degree, and 8 of them are with bachelor degree. As for total work years in HR department, 2 of them have over 16 years HR work experience, 3 of them have 11-16 years HR work experience, 5 of them have 6-10 years work experience, and 2 of them have 1-5 years HR work experience. As for participants’ age range, 1 of them is over 51 years old, 7 of them are in the range of 41-50 years old, and 4 of them are in the range of 31-40 years old. Detail information is listed in appendix A and table 3.1 to table 3.3.
Table 3.1 Educational Degree of the Experts
Table 3.2 Total Years of Work Experience in HR Department
Number of the Participants Year Range
2 Over 16 years
3 11-15 years
5 6-10 years
2 1-5 years
Number of Participants Degree Earned
4 Master
8 Bachelor
Table 3.3 Ages of the Experts
Number of the Participants Age Range
1 Over 51years old
7 41-50 years old
4 31-40 years old
Instrument
The applying instrument of this study is questionnaire via Delphi technique, the reliability and validity of the instrument are described as follows.
Reliability of the Instrument
Reliability refers to the demonstration that the operation and procedures of the research inquiry can be repeated by other researchers which then achieve similar findings, that is, the extent of findings can be replicated (Riege, 2003).
As for reliability of questionnaire, peer review was conducted to do the pre-test of questionnaire. The sample of peer review is 10 HR employees from 3 different domestic insurance companies and 10 HR-majored graduate students who study in the institute of international workforce education and development (IWED) in Nation Taiwan Normal University. Suggestion and recommendation will be the adapted to revise the questionnaire.
Validity of the Instrument
Validity evidences specifically refers to the empirical evidence and theoretical rationales that support the interpretations and actions that we take on the basis of the score or scores we get from an assessment procedure (Johnson & Christensen, 2004).
Hence, as for the validity of questionnaire, the questions of questionnaire are from relevant literature review and pilot study with 5 experts from 4 different insurance companies with average 10 work years in HR field from the expert panel what can reveal the real business circumstance with scholars’ professional point of views.
Data Collection and Analysis
This section is going to be divided into two parts, data collection and data analysis.
Data Collection
As for questionnaire, 12 experts were invited to complete the first round to third round of questionnaire, and in the fourth time of questionnaire, the researcher listed 9 questions to ensure the expert panel understands each item of the instrument and ensure the instrument is appropriate to use among insurance companies in Taiwan to meet the research purposes. The questions will be developed based on the analysis of previous round of questionnaire and develops questions in close-ended way.
Data analysis
The data analysis of this study is to analyze the responses of the questionnaire.
The questionnaire adopts Likert scales. The popularity of Likert scales can be traced to a number of factors, including ease of construction, intuitive appeal, adaptability, and usually good reliability (Babbie, 1998; Nunnally, 1978). The questionnaire will conduct 5-point Likert scale. The responses are equated with integers, including strongly agree=5, agree=4, neutral=3, disagree =2 and strongly disagree=1.
In this study, the researcher conducts mean score of minimum 3.75 to represent the important option because 3.75 represents the top 75% from mean score of number 1 to 5. Besides, to lend strength to the selected response and to the degree on consensus about the response, it was decided that a maximum of 0.6 of quartile deviation (Q.D.) of the responses had to be satisfied before the consensus was considered to be highly reached (Cheng, 2005).