• 沒有找到結果。

This chapter describes the research method of the study. Research design, research framework, research hypothesis, and research procedure are described first. Then, sample, data collection procedure and measurement are described. Data analysis will be described at the last in this chapter.

This study was conducted using quantitative method. Self-report questionnaires were adopted to collect data from 266 undergraduate students studying in business-related major in the colleges in Taiwan. A quasi-experimental research method was used to test the difference between pretest and posttest of organizational attractiveness before and after job seekers examine the website content. Moreover, this study also investigated the moderating effect of website content on the relationship of organization brand and organizational attractiveness.

Research Framework

The research framework is developed from the literature. Organization brand is selected to be the independent variable which includes two dimensions: organizational familiarity and organizational image. It is proposed to have a positive effect on organizational attractiveness.

Website contents including job and organization information is proposed to have a positive moderating effect on the relationship between organization brand and organizational attractiveness. The research framework is shown in Figure 3.1.

19 Figure 3.1. Research framework

Note. H2 refers to the difference between pretest and posttest of organizational attractiveness after job seekers examine the website contents.

Research Hypothesis

The hypotheses of the study are as follows:

Hypothesis 1. Job seeker’s perception of organization brand will positively predict

organizational attractiveness.

Hypothesis 2. There will be a significant difference between pretest and posttest of

organizational attractiveness after job seekers examine the website contents.

Hypothesis 3. Website content will have a positive moderating effect on the relationship between job seeker’s perception of organization brand and organizational attractiveness.

Research Procedure

This section describes the eight steps for the researchers to conduct this study. The first step helped the researchers to build up background knowledge, generate and narrow down interested research topics. A specific research topic was identified following the first step. Then, the research purpose and questions were developed. Based on the research purpose and questions, the researchers developed research framework and hypothesis. Then, the instrument was developed based on the literature. Pilot test was performed to ensure the validity and

20

reliability of the instrument. Then, the data was collected and analyzed to test the hypothesis and answer the research questions. Finally, the findings were concluded based on the results and literature. The procedure is shown in figure 3.2.

Figure. 3.2. Research procedure

Sampling and Data Collection

The sample population of this study were junior and senior undergraduate students and graduate students studying in departments from business schools in colleges in Taiwan.

Snowball sampling were used due to the lack of the sampling frame. The participants were

Identify Research Topic

Develop Research Purpose and Questions

Develop Research Framework and Hypotheses

Develop Instrument for Data Collection

Conduct Pilot Test

Collect Data

Analyze Data Review Literature

Conclude Research Findings

21

recruited from 22 different colleges in Taiwan and the sample size was 266. Junior and senior undergraduate students and graduate students in colleges are targeted as the sample because they are about to enter the workforce soon and are more likely to reflect the intentions of job seekers of the new generation. The reason for targeting on students in departments from business schools was because they have higher opportunities to find jobs in most kinds of industries.

Quasi-experimental approach was conducted in this study using survey questionnaires to collect self-report data from the samples. There were two kinds of data collection methods: via paper questionnaires and via online questionnaires. The measurement of research variables in the questionnaire were translated from previous researches such as Blau, 1994; Cober, Brown,

& Levy, 2004; Highhouse, Lievens, & Sinar, 2003; Lievens, Van Hoye & Schreurs, 2005, and all of the items underwent backward translation process to ensure the validity. For the paper questionnaires, the researcher used snowball sampling by contacting the teachers of classes consisted by either junior, senior or graduate students in colleges for permission to deliver the survey questionnaires in the classes. The data were collected from 6 classes in 3 different schools with the completed questionnaires in the number of 166 (62%) of total 266 samples.

For the online questionnaires, the researchers used snowball sampling by contacting friends and by sending email invitations to the secretary/assistant from departments of business schools in 32 different schools located in various areas in Taiwan. The total returned questionnaires were 100 (38%) of total 266 samples and were from 20 schools.

In the data collection procedure, every participant was randomly assigned an organization name from the list of “Top 100 Most Admired Employers by the New Generation”. First, a pre-test was conducted to assess participants’ perceptions of organization image and organizational attractiveness toward the assigned organization. After they finished the pretest, the students were asked to examine the website of the assigned organization from the perspective of a job

22

seeker and fill out the questionnaire on website content. Then, a post-test was conducted to assess their perceptions of organizational attractiveness after they had examined the organization website. Finally, the participants were asked to fill out the background information questions. The average time for the participants to complete the paper questionnaire was within 20 minutes. According to Allen et al., (2007), the time for examine the website would not affect the result, therefore, it was not included in the analysis.

Sample Profile

The researchers collected the questionnaire from 266 participants, including 76 males (29%) and 190 females (71%). There were two types of questionnaire source: 166 paper survey (62%) and 100 online survey (38%). Age of the participants ranges from below 20, 21~25, to above 26. Most of them are in the group of 21~25 (90%). The education level includes undergraduate and graduate school with the majority in undergraduate study (79%). Their working experience ranges from “have been working less than 1 year” to “have been working more than 6 years”. The largest group is “have been working less than 1 year” (40%) and the second largest group is “have been working 1~2 years” (38%). Please refer to Table 3.1 for detailed information.

Table 3.1.

Descriptive Statistics on Sample Characteristics

Item Frequency Percentage (%)

Gender

23

Independent-Samples T-test of the Samples

In order to establish the validity of the samples, the researcher conducted several independent-samples t-tests. First, because almost half of the participants come from National Taipei University of Business (NTUB), they were categorized as one group, and their responses were compared with the participants from other schools. The result shows that there is no significant difference between the two groups’ answers on all the research variables. For

24

instance, on the score of organization brand, there is no significant differences between students from NTUB (M=3.20, SD=.73, n=125) and students from other schools (M=3.20, SD=.79, n=141), t(263)=.06, n.s., two-tailed. Second, the researcher tested the difference between answering by paper questionnaires and answering by online questionnaires. The result shows that scores of most of the research variables have no significant differences. Only organizational attractiveness shows significant difference between paper questionnaires (M=3.14, SD=.76, n=166) and online questionnaires (M=3.40, SD=.79, n=100), t(201)=-2.68, p=.01, two-tailed. However, it would not affect the following analysis in testing the hypotheses.

Last, the researcher tested the difference between examining the website content through a computer and that through a mobile phone. The result shows that there is no significant difference on the scores of website content, t(144)=.26, n.s., two-tailed. between the former (M=.55, SD=.30, n=194) and the later (M=.54, SD=.25, n=71). Please refer to Table 3.2 for detailed results.

Table 3.2.

Results of Independent-Samples T-test (N=266)

School 95% CI

for Mean Difference Students from NTUB Students from other school

M SD n M SD n t df p

OB 3.20 .73 125 3.20 .79 141 -.18, .19 .06 263.40 .95 OA 3.15, .76 125 3.30 .80 141 -.34, .04 1.6 262.60 .11

WC .55 .26 125 .54 .30 141 -.06, .08 .20 263.79 .85

POA 3.39 .72 125 3.47, .67 141 -.24, .09 -.89 255.17 .38

Questionnaire type 95% CI

for Mean Difference paper questionnaires online questionnaires

M SD n M SD n t df p

25 Table 3.2. (continued)

Device to examine website content 95% CI for Mean Difference

Computer Mobile phone

M SD n M SD n t df p

WC .55 .30 194 .54 .25 71 .06, .08 .26 144.39 .80

Measurement Organization Brand

Organization brand is defined as component of a name, term, sign, symbol, design, or either combination of them (Kotler, 1979). This variable includes two dimensions:

organizational familiarity and organizational image. Participants were asked to score organization brand of an assigned company which was chosen from the list of " Top 100 Most Admired Employers by the New Generation” provided by Cheers Magazine (http://topic.cheers.com.tw/news/20150320.pdf). The measurement of organizational familiarity and image is derived from Lievens et al. (2005) and Lemmink et al. (2003). A 5-point Likert Scale was applied to measure the construct of the two dimensions, from 1 (strongly disagree) to 5 (strongly agree). Answers to all items from the two dimensions were summed to form a total score for organization brand.

Organizational familiarity. This study derived the measurement of organizational familiarity from Lievens et al., (2005). The internal consistency of this measurement was .81 (Lievens et al., 2005). This dimension includes three items. An example item of organization familiarity is “I am familiar with the organization as an employer”. Please refer to table 3.3 for details on measurement items.

26 Table 3.3.

Items of Organizational Familiarity

Item description Source

I am familiar with this organization as an employer.

Lievens et al. (2005) I have heard from others what goes on in this organization.

I have already heard many things about this organization

Organizational image. Lemmink et al. (2003) tested the reliability of the first-order construct of organizational image by its composite reliability. All values surpassed the recommended cut-off level of 0.60 (Jöreskog, 1971). This dimension is operationalized into two sub-dimensions and includes totally six items. An example item is “This organization is the first choice for high quality products/services”. Please refer to table 3.4 for details on measurement sub-dimensions and items.

Table 3.4.

Sub-dimensions and Items of Organizational Image Organizational

management

Buying company stock of this organization is a good investment This organization often introduces innovations

Higher management is committed to this organization Short-term

experience

This organization possesses appealing advertising of products/services I heard/experienced positive things about this organization

This organization is the first choice for high quality products/services Note. Adapted from “The Role of Corporate Image and Company Employment Image in Explaining Application Intentions,” by J. Lemmink, A. Schuijf, and S. Streukens. 2003, Journal of Economic Psychology, 24(1), p. 1-15. Copyright 2005 by Elsevier.

Website Content

The definition of website content in this study is the amount of job and organization information provided on the website. The measurement of this variable was derived from Cober et al. (2004). The questionnaire uses yes/no options to measure the presence or absence of job and organization information (yes=1, and no=0). Answers to all items were summed to form a

27

total score for website content. Higher score means more information are presented in the website. An example item is “Information presented regarding the culture, goals, values, or working environment of this organization”. Please refer to table 3.5 for details on measurement items.

Table 3.5.

Measurement of Website Content: Types of Job and Organization Information and Operational Definitions

Types of job &

organization information

Operational definitions

Employment-Related Information

Compensation Information found regarding salaries and/or benefits

Culture Information presented regarding the culture, goals, values, or working environment of an organization

Training Information about training and career development included on the Web site

Online Job Preview Information describing a typical day or responsibilities of an employee, other than that found in specific job advertisements, included on Web site

Fit Messages

Targeted Message Sections of the Web site specified for navigation by specific groups of job seekers (e.g., college students, experienced workers)

Work-Life Message Information found directly referring to organizational facilitation of employee balance between work and personal life

Diversity Message Web site contained messages promoting workforce diversity (e.g., racial, gender, sexual) beyond a statement about EEO compliance

Employee Testimonials Employee statements of their work experience found on the Web site; this excludes messages from the president/CEO or customers

Community Service Information about corporate philanthropy and/or employee community service found on employment Web site

Note. Adapted from “Form, Content, and Function: An Evaluative Methodology for Corporate Employment Web Sites,” by R. T. Cober, D. J. Brown, & P. E. Levy. 2004, Human Resource Management, 43(2/3), p. 201-218. Copyright 2004 by Wiley Online Library.

28

Organizational Attractiveness

Organizational Attractiveness refers to the degree of applicant attraction toward a certain organization (Turban & Greening, 1997). The measurement of this variable is derived from Highhouse et al. (2003). It includes three dimensions: general attractiveness, intentions to pursue, and prestige. The internal consistency of the three scales was measured by using Cronbach’s alpha value (general attractiveness=.88, intentions to pursuit=.82, and prestige=.83). A 5-point Likert Scale will be applied to measure the construct, from 1 (strongly disagree) to 5 (strongly agree). Answers to all items were summed to form a total score for organizational attractiveness. An example item is “this company is attractive to me as a place for employment”. Please refer to table 3.6 for details on measurement items.

Table 3.6.

Dimensions and Items of Organizational Attractiveness

General attractiveness

1. For me, this company would be a good place to work.

2. I would not be interested in this company except as a last resort.

3. This company is attractive to me as a place for employment.

4. I am interested in learning more about this company.

5. A job at this company is very appealing to me.

Intentions to pursue

6. I would accept a job offer from this company.

7. I would make this company one of my first choices as an employer.

8. If this company invited me for a job interview, I would go.

9. I would exert a great deal of effort to work for this company.

10. I would recommend this company to a friend looking for a job.

Prestige

11. Employees are probably proud to say they work at this company.

12. This is a reputable company to work for.

13. This company probably has a reputation as being an excellent employer.

14. I would find this company a prestigious place to work.

15. There are probably many who would like to work at this company.

Note. Adapted from “Measuring Attraction to Organizations,” by S. Highhouse, F. Lievens, and E. F. Sinar. 2003, Educational and Psychological Measurement, 63(6), p. 986-1001.

Copyright 2003 by Sage Publications.

29

Control variables

Based on previous research, some demographic variables were controlled, such as gender, race, work experience and years in school (Cable & Judge, 1994, 1996; Turban, 2001) in the regression analysis to predict attraction. These variables are chosen as control variables because individual’s demographic characteristics, work experience, and years in school would affect their applicant attractions toward the organization as job seekers (Turban, 2001). These variables were measured by the questionnaire in the background information part. Race is already controlled in this study. Gender will be a nominal measure with two options (male/female). Years in school was measured by education level question which has five choices from (1) below high school to (5) above graduate school. Work experience was an ordinal measured with five choices from (1) less than 1 year to (5) above 6 years. In addition, job search behavior was also chosen to be the control variable because job seeker’s job search experience may also have effects on applicant attraction.

Schwab et al. (1987) suggested that job search intensity was one of the factor in producing job alternatives. It referred to preparatory and active job search behavior (Blau, 1994; Schwab et al, 1987). In most cases, job seekers may experience the information collection and evaluation process when they are searching for a job. For example, if they had examined some organizations’ websites before this study, they may have different perceptions toward the organization from other participants who had never done it before. In addition, other factors in job search behavior may also affect job seekers original perceptions toward an organization.

For instance, the job seeker who had an interview experience with a certain organization may have different perceptions toward the organization than others (Turban, 2001). It was also suggested that recruiter's behaviors during the interview may affect applicant attraction (Harris

& Fink, 1987). Based on the literature, the researchers decided to combine these behaviors into one variable as job search behavior and add it as a control variable. The measurement of job

30

search behavior was derived from Blau (1994). Please refer to the appendix A and B for details on measurement items of the control variables.

Method of Data Analysis

In this study, the researchers conducted the analyses through statistical programs SPSS version 22 and SPSS AMOS. Descriptive statistics was applied in order to get a clear understanding of the important characteristics of the sample, as well as the mean, standard deviation and distribution of data.

Factor analyses were conducted to establish the construct validity. First, exploratory factor analysis (EFA) was performed through SPSS both in pilot study and in the main study. It was used to identify the properties of measured variables, to make sure of the factor loadings and if all items loaded onto the appropriate factors. When conducting EFA, Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy should be above .50 and Bartlett’s test of sphericity should be significant. Kaiser (1974) suggested when the KMO index reaches the .90s, it is

“marvelous”, in the .80s is “meritorious”, in the .70s is “middling”, in the .60s is “mediocre”, and in the .50s is “miserable”. Second, confirmatory factor analysis (CFA) was conducted through AMOS, a covariance-based structural equation modeling (SEM) technique to test the validity of measures. CFA is usually conducted based on established theories. It is used to observe the relationship among latent variables, and to see if it establishes goodness of fit with the original theoretical model. Commonly interested indices to determine model fit include chi-square divided by degree of freedom (2/df ), SRMR (standardized root mean square residual), GFI (goodness of fit), AGFI (adjusted goodness of fit), RMSEA (root mean square error or approximation). Please refer to Table 3.7 to see a summary of suggested model fit indices and their levels. In addition, researchers suggested that for appropriate construct validity, average variance extracted (AVE) should be above .50 and composite reliability (CR) should be above .80 (Fornell & Larcker, 1981).

31 Table 3.7.

Levels of Suggested SEM Model Fit Indices

Good fit Acceptable fit

2/df 2-5 (2008) (top row) and Schermelleh-Engel, Moosbrugger, and Müller (2003) (bottom row)

Moreover, Pearson’s Correlation Analysis was conducted to examine the linear relationship among variables and also to test the first hypothesis. T-test was used to compare the mean difference between two variables. It was used to test if there was significant difference between answers to organizational attractiveness in pretest and posttest (hypothesis 2). Lastly, multiple regression analysis was performed to test the moderating effect of website content on the relationship between organization brand and organizational attractiveness (hypothesis 3).

Reliability and Validity Analysis

Validity includes content validity and construct validity. All the measurements underwent expert review to ensure the content validity. A pilot study was conducted to ensure the validity and reliability of the measurement. It also helped to improve the design of the main study.

Exploratory factor analysis (EFA) was conducted to examine the factor structure both in the pilot test study and in the main study. Construct validity was established through the confirmatory factor analysis (CFA) in this study. The validity analysis, results of EFA and CFA, and also the reliability analysis, are presented in this section.

32

Pilot Study

A pilot study was conducted to ensure the validity and reliability of the measurement, and also to preliminarily understand the relationships among study variables. Samples of pilot study were 40 undergraduate senior students and graduate students from any department in colleges in Taiwan. Convenience sampling was used due to the lack of the sampling frame. The data was collected by online questionnaires from all the participants (Fang & Yeh, 2016). Pretest and posttest of organizational attractiveness were conducted to see the effect of website content on it. Please refer to Table 3.8 to see more detail of descriptive statistics on sample characteristics.

Table 3.8.

Descriptive Statistics on Sample Characteristics of Pilot Study Item Frequency Percentage

(%)

Item Frequency Percentage

(%)

33

Exploratory Factor Analysis was conducted to ensure the validity (Fang & Yeh, 2016).

Please refer to Table 3.9 for the factor structure. Extraction method was Principal Component Analysis. Rotation method was Varimax with Kaiser Normalization. First, Kaiser-Meyer-Olkin

Please refer to Table 3.9 for the factor structure. Extraction method was Principal Component Analysis. Rotation method was Varimax with Kaiser Normalization. First, Kaiser-Meyer-Olkin

相關文件