Research Framework
Based on the review of literature, the framework consists of process virtualization requirements, IT Capability, Individual Attributes, and Behavioral Intention to use. With Behavioral Intention as the dependent variable, while process virtualization requirements, IT Capability and Individual Attributes act as independent variables.
36
Research Hypothesis
Following literature review, the hypotheses below were formed:
Hypothesis 1: Perceived Sensory Requirements are negatively related to Behavioral Intention to use E-HR system.
Hypothesis 2: Perceived Relationship Requirements are negatively related to Behavioral Intention to use E-HR system.
Hypothesis 3: Perceived Synchronism Requirements are negatively related to Behavioral Intention to use E-HR system.
Hypothesis 4: Perceived Identification and Control Requirements are negatively related to Behavioral Intention to use E-HR system.
Hypothesis 5: Perceived Representation Capability is positively related to Behavioral Intention to use E-HR system.
Hypothesis 6: Perceived Reach Capability is positively related to Behavioral Intention to use E-HR system.
Hypothesis 7: Perceived Monitoring Capability is positively related to Behavioral Intention to use E-HR system.
Hypothesis 8: Attitude toward Technology is positively related to Behavioral Intention to use E-HR system.
Hypothesis 9: Computer Self-efficacy is positively related to Behavioral Intention to use E-HR system.
Research Design
Different HR processes have been suspected to account for different degrees of process virtualizability. The study chose two HR processes that show opposite trend in degrees of virtualization as measurement context: (1) change of personal profile information and (2) performance appraisal. The chosen processes shall only act as a
skeleton for future explorations of other processes (i.e. the adoption of the measurement items to a specific process). The measures are tested on two "extreme examples", to see if they work out well in at least two different and distinct situations.
Similarly, the research controls the variation in sophistication of IT technology by conducting the research in an E-HR Lab where two "extreme examples" of IT are used to test virtualization requirements of different processes.
The “change of personal profile information” HR process was believed to be low in the process virtualization requirements because minimal social interaction is involved. During the general process, the applicant only had to request for a form to change his/her personal profile, fill it out and turn it in. Therefore, interaction with human being or object is assumed to be minimal. On the other hand, the “performance appraisal” HR process was believed to have a much higher process virtualization requirement because normally the performance appraisal process has to be done with a lot of forms and human interactions. The forms have to go back and forth between different departments, managers, and HR professionals. The process may also require face-to-face meetings between applicants and managers/HR professionals.
The experiment was conducted using one of the most sophisticated E-HR products, PeopleSoft. Due to the complexity of the PeopleSoft software, an imitation of the software was used on the subjects to provide control over the experiment and reduce time spent on training the subjects. Using an imitated "mock-up" version of PeopleSoft, all the participants were able to go through the same procedure without wandering off to other functions of the software. This also minimized the difference in familiarity with the software. The participants were still able to click and interact with the designated functions for the experiment. The idea was to make the experience as uniformed as possible, providing sufficient control over the process of
38
Figure 3.2 E-HR process for "Change of personal profile information"
Figure 3.3 E-HR process for "Performance Appraisal"
Experimental Procedure
1. The subjects were randomly divided into two groups with even numbers of participants. They were placed in separate sides of the lab to avoid interference between the students. Pictures demonstrating the experiment are shown in
appendix A.
2. The subjects were issued Part A of the questionnaire which contains measurement scales on Attitude toward Technology and Computer Self-efficacy.
3. Subjects were briefed on the definition and content of the selected HR processes.
The conceptual scripts are included in appendix B.
4. Subjects were instructed to complete part B of the questionnaire with the four process virtualization requirement measures.
5. The subjects were asked to watch two different tutorial videos for completing the two HR processes on the PeopleSoft software separately. The two HR processes were change of personal profile information and performance appraisal, respectively. They then went on with the hands-on experience of the two E-HR processes respectively as mentioned above.
6. After the subjects have experienced the virtualized process, they were asked to complete part C of the questionnaire which contains IT Capability, Behavioral Intention measures, and demographic information.
The entire process took an average time of approximately 40 minutes. A flow chart on the experimental procedure is shown in Figure 3.4.
40
Sample and Data collection
The sample for this research was targeted at college to graduate level students with business major. This is due to the fact that they are the soon-to-be potential workforce. Also, business major students may have a better idea of organizational processes. Convenience sampling was used to find sample students.
Due to the time it took to prepare, explain HR processes, and setup computer stations, the researcher‟s ability to acquire a large number of participants was limited.
The researcher contacted teachers of business major classes and asked for permission to come to their classes to solicit participation. At the end, the researcher paid a total of five visits to different college campuses in northern Taiwan, set up E-HR labs in their computer classrooms, and conducted the experiments. Some participants were invited to the NTNU E-HR lab for experiment. A total of 230 samples were collected from 6 different schools. 217 out of 230 samples were valid for analysis in the end, which was about 94%. The subjects ranged from college level to master level students.
Table 3.1.
Sample Profile
Feature Categories Count/Value Percentage (%)
Treatment Profile Change 112 51.6
Performance Appraisal 105 48.4
Gender Male 73 33.6
Female 144 66.4
Education Bachelor‟s Degree
program
190 87.6
Master Degree program 27 12.4
Computer
42
Table 3.1 (Continued).
Feature Categories Count/Value Percentage (%)
Schools National Taiwan normal
Tongnan University 33 14.3
According to the research design, the researcher tried to collect equal number of applicants for the 2 types of treatment. After deletion of invalid data, 112 were kept from profile change group and 105 were kept from performance appraisal group.
Due to the traditions in Taiwan, females would usually go to business school while males would pursue the engineer path. It is reflected in the sample collected in which 83.7% of the participants were female. The average age of the participants was around 22. Since most of the participants were born in the Net generation, a good number of participants were able to use computers at a very early age, so most participants had more than 6 years of computer experience. Considering that undergraduate students are a bigger population and easier to reach than graduate level students, most participants were undergraduate students.
Measurement Individual Attributes
The scale of Attitude toward Technology intends to measure an individual's overall affective reaction toward using online HR technology. Ten items were adapted from the scale Venkatesh et al. (2003) developed in order to measure Attitude toward
Technology.
The measurement was administered using a Likert type scale rating of 1 to 5, 1 would mean strongly disagree with the statement and 5 would be strongly agree with the statement. A higher score would mean a more positive Attitude toward Technology. Scores were aggregated in the end to sum up the results. Examples for the items include "The actual process of using online work-related technology is pleasant." and " I look forward to those aspects of my job that require me to use online work-related technology."
Computer Self-efficacy aims to measure the user‟s perception of his or her ability to use a computer to accomplish a specific task. A scale consisting of 9 items developed by Compeau and Higgins (1995) was used to measure the individual's Computer Self-efficacy.
The measurements was administered using a Likert type scale rating of 1 to 5, 1 would mean strongly disagree with the statement and 5 would be strongly agree with the statement. A higher score would mean a better self-efficacy to use computer technology. Scores were aggregated in the end to sum up overall Computer Self-efficacy. Items of self-efficacy include "I could complete the job using a computer software package if I had never used a package like it before." and "I could complete the job using a computer software package if someone showed me how to do it first."
Process Virtualization Requirements
A scale for measuring the process virtualization requirements from Overby and Konsynski (2010) was used as a starting point for developing measurement under HR context. The original measurement scale did not have enough items, and some of the items did not fit the HR context, therefore, some items for measuring the process
44
For the Sensory Requirements, the original concept from the theory was to measure the degree to which the participants needs to enjoy a full sensory experience of the process and the other process participants and objects. Another measurement scale by Schmitt, Zarantonello, and Brakus (2009) was supplemented in the development of measurement items.
The Relationship Requirements aim to measure the need for process participants to interact with one another in a social or professional context. Items from the scale of Lin, Lin, and Laffey (2008) was used to complete the scale for Relationship Requirements.
Synchronism Requirements refer to the need for a process to occur quickly with minimal delay. Items were adapted from Culiberg and Rojšek (2011) and Jun, Yang, and Kim (2004) to develop a scale to measure Synchronism Requirements under HR context.
The goal of the Identification and Control Requirements is to measure the need for a process to require unique identification of process participants and the ability to exert control over or influence their behavior. Items from Ho and Lee (2007) were adapted to complete the measurement scale for Identification and Control Requirements under HR context.
The measurement was conducted using a Likert type scale rating of 1 to 5, 1 would mean strongly disagree with the statement and 5 would be strongly agree with the statement. A higher score would mean a higher process virtualization requirement.
Scores were aggregated under each of the process requirements. A table showing sample items for each of the process virtualization requirements is shown in Table 3.2.
Table 3.2.
Sample Items for Process Virtualization Requirements Process Virtualization
Requirement
Sample items Sources
Sensory Requirements I like to physically see a real person/paperwork when conducting this HR function.
Overby and Konsynski (2010) This function requires visual sense or other senses to
complete.
Schmitt et al.
(2009) Relationship
Requirements
I enjoy the social aspects of interacting with the real people.
Overby and Konsynski (2010) I feel the need to express myself to the HR department Lin et al. (2008) Synchronism I need the service I requested to be delivered to me
within promised time.
Jun et al. (2004) Identification and
Control Requirements HR takes note on who request for this function. Overby and Konsynski (2010) I feel my privacy is protected in this particular HR
process.
Ho and Lee (2007)
IT Capability
Scale for measuring the IT Capability were adapted from Overby and Konsynski (2010) to start the development of measuring the capability of IT under HR context.
The original measurement scale did not have enough items and some of the items did not fit the HR context. Also, Overby and Konsynski's scale did not include items for Monitoring Capability. Therefore, some items for measuring the IT Capability were self-developed.
The Representation Capability of IT Capability attempted to measure IT's capability to present information relevant to a process and how easy it is for users to interact with them. Other than Overby and Konsynski's scale, scale by Barnes and Vidgen (2002) was used to develop measurement for the Representation Capability of IT.
46
The Reach Capability is to measure IT's capacity in allowing process participation across both time and space. In addition to Overby and Konsynski's scale, items from the scale developed by Goodhue (1998) were adapted to fit under HR context.
Monitoring Capability is to measure IT‟s capability to authenticate process participants and track activity. Items were adapted from Barnes and Vidgen (2002) to develop the scale measuring Monitoring Capability of IT.
The measurement was conducted using a Likert type scale rating of 1 to 5, 1 would mean strongly disagree with the statement and 5 would be strongly agree with the statement. A higher score would mean a better IT Capability. Scores were aggregated under each of the IT capacities. Table 3.3 shows the sample items for each of the IT Capability.
Table 3.3.
Sample Items for IT Capability
IT Capability Sample item Sources
Representation Capability
I can get the information I need from the online HR technology.
Overby and Konsynski (2010) I find the online HR technology easy to navigate Barnes and
Vidgen (2002) Reach Capability Using online HR technology, it is easy to get access to
data that I need.
Goodhue (1998) Using online HR technology allowed me access to HR
activities I otherwise couldn't complete.
Overby and Konsynski (2010) Monitoring
Capability
It feels safe to fill data into the online HR technology. Barnes and Vidgen (2002) I feel confident using the online HR technology.
Behavioral Intention
Behavioral Intention attempted to measure an individual's overall intention to use online HR technology. A scale from Hu, Chau, Sheng, and Tam (1999) was adapted in
development of the scale to measure Behavioral Intention.
The measurement was conducted using a Likert type scale rating of 1 to 5, 1 would mean strongly disagree with the statement and 5 would be strongly agree with the statement. A higher score would mean a higher Behavioral Intention to use online HR technology. Scores were aggregated in the end to sum up overall Behavioral Intention. Examples for this measurement scale include "Using online HR technology allowed me to conduct HR activities I otherwise could not complete." and "Using online HR technology, I can get information quickly and easily when I need it."
Control Variables
As seen in literature review, demographic variables such as gender, age, education, and years of computer experience have been known to affect Behavioral Intention to use technology; they are therefore included in the study as controls (Sun
& Zhang, 2006; Venkatesh et al., 2003).
Validity and Reliability Analysis Content Validity
In attempt to ensure the content validity of the measures, the entire questionnaire had gone through expert review. Seven experts were selected from different companies with criteria of a minimum one year of experience in E-HR software usage.
The software that was used in their company also had to include the functions conducted in this experiment. The researcher went on with face-to-face interviews with the experts one by one. A table showing the basic profiles of each of the experts is provided in table 3.4.
48
Table 3.4.
Experts Profile
Expert Industry E-HR experience(Years) E-HR software used by company
1 Advertisement 4 In-house development
2 Technology 3 In-house development
3 Retail 3.5 SAP E-HR system
4 Service 15 In-house development
5 Insurance 4 In-house development
6 Insurance 10 GSS(Galaxy software) E-HR system
7 Travel 1.5 In-house development
The experts provided helpful suggestions to make the wording smoother and easier to understand for participants. They also assured that the items measure what they were intended to measure. Since Chinese is a very sophisticated language, different wording may affect the ability to understand the measurement items.
Therefore, the experts went through every question to make sure they were easy to understand for the targeted participants. The original developed scale had 69 items, the seven experts went through and smoothed out the wording for 31 of them, one item was considered a duplicate and was deleted. After all seven of the experts were interviewed, the researcher produced a chart indicating all the changes by different experts and mailed them back to all the experts for cross-checking. The end result was 68 measurement items with agreement among all seven experts. The final measurement items are shown in appendix C, both the English and Chinese questionnaires.
Construct Validity
After data collection was completed, the researcher conducted exploratory factor
analysis (EFA) on all the measurement items in order to explore the factor structure of the measurement scales. The KMO value of the sampling adequacy was 0.86 which indicated the data was suitable to run factor analysis; the result of Bartlett‟s sphericity test was significant (p<.001), which meant the analysis outcome was valid to explain the structure. The items went through nine EFA runs to clarify the structure of constructs and to avoid cross loading. A total of 10 factors were extracted compared to the originally designed nine variables. The largest factor explained 25.82% of total variances; being less than 50%, the concern of common method variance was
minimized. In addition, all items with loading under .5 were deleted to ensure validity.
The following table demonstrates the EFA results.
Table 3.5.
Exploratory Factor Analysis Results
Factor Question Items Factor Loadings
Relationship and Sensory
Attitude toward Technology ATT3_3 .809
ATT2_2 .799
50
(Continued) Table 3.5 (Continued)
Factor Question Items Factor Loadings
Representation Capability (1) REP6_51 .843
REP5_50 .809
REP7_52 .807
REP4_49 .800
REP3_48 .539
Representation Capability (2) REP8_53 .736
REP9_54 .716
REP10_55 .558
Computer Self-efficacy COM2_12 .709
COM4_14 .705
Identification requirements IDEN6_42 .764
IDEN4_40 .754
IDEN7_43 .752
IDEN3_39 .701
IDEN5_41 .604
Monitoring Capability MON1_60 .833
MON2_61 .830
MON3_62 .592
MON4_63 .579
Synchronism Requirements SYNC3_34 .801
SYNC2_33 .750
SYNC4_35 .680
SYNC1_32 .643
Behavioral Intention (1) INT4R_67 .842
INT3R_66 .793
Behavioral Intention (2) INT5_68 .724
INT2_65 .696
INT1_64 .546
In the first run of EFA, the "Reach Capability" items were loaded along with the items of the dependent variable "Behavioral Intention" in the same factor, therefore, the "Reach Capability" items had to be dropped.
Contrary to the original theory, the process requirements of sensory and relationship merged into one dimension. There is a possible reason for this, in the original Process Virtualization Theory; it was mainly developed for e-commerce. In e-commerce, there is an actual physical product, which the process participant may want to physically touch and feel. However, when applying under HR context, the people offering HR services are the actual product. Therefore, the Sensory and Relationship Requirements were seen as the same concept by the participants.
Reliability
Cronbach‟s alpha value for each variable used in the study was calculated to ensure internal consistency reliability. In order to be acceptable, the Cronbach‟s alpha value should be over .70 (Nunnally, 1978). Cronbach‟s alpha values are shown in the following table.
Table 3.6.
Reliability Test Results
Variables Number of items Cronbach‟s Alpha
Relationship and Sensory
Requirements 10 .914
Attitude toward Technology 9 .906
Representation Capability 8 .870
Computer Self-efficacy 8 .870
Identification requirements 5 .847
Monitoring Capability 4 .822
Synchronism Requirements 4 .812
Behavioral Intention 5 .776
Since all variables had a Cronbach‟s alpha value of over .70, the scales are considered
52
to be reliable.
Revised Framework
The original research framework and hypotheses had to be revised because some items did not load as intended after EFA was conducted. The Sensory and
Relationship Requirements loaded together forming "Relationship and Sensory Requirements". While the "Reach Capability" measures loaded with the dependent variable and therefore had to be dropped. A revised framework is presented as Figure 3.5.
Revised Hypotheses
To reflect the changes in research framework as shown above, two hypotheses were dropped and the revised list of hypotheses is presented below.
Hypothesis 1: Perceived Sensory and Relationship Requirements are negatively related to Behavioral Intention to use E-HR system.
Hypothesis 2: Perceived Synchronism Requirements are negatively related to Behavioral Intention to use E-HR system.
Hypothesis 2: Perceived Synchronism Requirements are negatively related to Behavioral Intention to use E-HR system.