Determinants to Behavioral Intentions of Job Recruiters to Distribute to E-Recruitment Systems

Abstract

Purpose: With the interest and ability in recruitment, it is totally worthy to discover how is labor market developed in Vietnam since the existence of the Internet. Considering employers as well as recruiters similar to jobseekers in term of technology adoption of job distribution, they should be treated fairly while performing their scopes of work. This paper aims to explore how employers perceive online recruiting platform as a supportive system, and this is to find out contributions of external factors regarding Job Boards behavioral intention to use. Research design, data and methodology: Conceptual framework tests the hypotheses involving in recruitment websites. Results: The results suggest that Board of Directors (BoD) or direct middle managers need to pay more attention on how their workforce performs through various channels of recruitment - both formal and informal, such as Linkedin or Facebook rather than a few compulsory job boards - no longer being accessed with high performance expectancy. However, the technology adaptation of recruitment is still a controversial issue due to its access to privacy of users. Conclusion: Therefore, employers should be more careful on how their personal contacts are going to be delivering without consent. Secondly, keeping systems up to date and "connected" is essential for both employees and employers.

1. Introduction¹

At any stage, technological applications and tools such as corporate websites, online job boards, or even social platforms are available to assist the recruiters, job distributing a more efficient and advanced process, called e recruitment that offers timely, cost-effective and interactive solutions to both recruiters and job seekers. Among of all mentioned digital recruitment platforms, job boards are still the most popular one (Nguyen, 2021a; Kim & Cho, 2015; Javanmard, 2016).

Recruitment is the procedure of interaction between candidates and employers through a specific channel or system. Therefore, it is necessary to consider employers as well as recruiters similar to jobseeker in term of technology adoption, the process to adopt innovation into their scope of work needed to be examined carefully (Nguyen & Tran, 2019; Nguyen, Nguyen, & Tran, 2021; Nguyen, 2019). In Vietnam, job boards are growing in popularity to both applicants and practitioners. There are just a small number of studies explained the other side of recruitment procedures – technology adoption of job distribution. It is necessary to consider employers as well as recruiters similar to jobseeker in term of technology adoption.

As the use of technology in recruitment is getting more and more common in developed countries, the recruitment process in Vietnam is evaluated as very old-fashioned and slow to adapt new technology that has become effective and efficient globally. Therefore, it is highly essential for recruiters to access and apply hi-tech tools to improve the effectiveness as well as efficiency of their work.

Aligning with the above-mentioned research statements, this study aims to identify the factors influencing recruiter’s Perceived ease of use regarding Job Boards application and to evaluating Perceived ease of use, Perceived Usefulness and Perceived Privacy Risk impact on recruiters’ behavioral intentions to use Job Boards application.

This study proposes that Computer Self Efficacy has a direct and positive impact on Perceived Ease of use, while Perceived Usefulness is directly and positively associated with Behavioral Intention to Use job boards. The paper also makes an argument about directions for future research regarding the effects of Internet on recruitment and how employers react to adopt this questionable development. Especially in Vietnam human resource market, practical application of this study’s findings is useful for managers to understand and encourage the use of e-recruitment in general and Job board in specific in their business.

2. Literature Review

2.1. Technology Acceptance Model (TAM)

Technology Acceptance Model is a well-known diagram which has been researched and combined by Davis (1993) in order to introduce the totally new perspective in how individual users take action and come up with an acceptance of latest technology.

In TAM model, there are two elements – perceived usefulness and perceived ease of use have significant relationship with individuals’ behavior using computer (Davis, 1993).

In this study, in order to focus more on the outcome of job distribution which is Behavioral Intention to Use job boards of recruiters or employers, the TAM model has been modified to be suitable.

2.2. E-Recruitment

E- Recruitment, as defined by Yoon Kin Tong, is the use of internet enabled technologies to attract and select candidates for a live vacancy existing in an organization (2009). “E-recruitment has seen phenomenal success within a very short period of time” (Galanaki, 2002).

E-recruitment applications can be divided into three types which commonly are Corporate (Company’s own) website for recruitment, Commercial Job Boards for posting jobs advertisements, Social Platforms for rapid interaction recruitment posts (Nguyen & Nguyen, 2020; Park, Chaffar, Kim, & Ko, 2017).

2.3. Job Boards

Jobs boards are online-platforms where organizations can pay to advertise an opening position. Organizations tend to use Commercial job boards to go along with the trend or to keep up with their components (Kim & Jung, 2012; Nguyen, 2021b).

2.4. Defined Terms Based on Previous Studies

Perceived privacy risk (PPR)

Perceived Privacy Risk (PR) is frequently considered as an issue. Due to Bauer’s study published in1960 and continuous development of internet in following years, Privacy Risks as "barriers" to computing innovation acceptance. Most customer Behavior writings assessed on PRs are in relation of money aside from Liebermann and Stashevsky (2002) discoveries on the legitimacy of individual data appropriation without owners’ consent. This issue relates not exclusively to jobseekers but also to employers.

Computer Self-Efficacy (CSE)

In this study, CSE is going to take into insightful analyses due to it broader range of effects which is characterized as an individual judgment of efficacy over different computers (Choi, 2013; Singh, 2016).

Perceived usefulness (PU)

Perceived Usefulness is mentioned as the expecting user's subjective probability that the usage of a specific structural supporting tool will improve his/her activity (Davis, 1993).

Perceived ease of use (PEOU)

Overall, an easy-to-use application requires as much effort as possible but subsequently improving the probability of high-level performance. On the other hand, the complicated one is hard to get into adoption due to large amount of practices, exertion and enthusiasm done by clients (Teo, 2001).

Behavioral Intention to use (BI)

Mentalities towards usage and intentions to involve in might be not well created or having confusion in the first place or just slowly figure out how to utilize the trending innovation. The connection among Perceived Usefulness (PU) and Behavioral Intention to Use (BI) was explained in Sanchez-Franco and Roldan (2005).

2.4. Conceptual Model and Hypotheses

Intentions to use E-Recruitment Applications (Figure 1) in which the hypotheses are as below:

Figure 1: Conceptual model: Factors That Impact Recruiters’ Behavioral

H1: Computer Self-Efficacy has a positive impact on Perceived Ease of Use

H2: Perceived Privacy Risk has a negative impact on Behavioral Intentions to Use

H3: Perceived Ease of Use has a positive impact on Behavioral Intentions to Use

H4: Perceived Usefulness has a positive impact on Behavioral Intentions to Use

3. Methodology

3.1. Research Method

Quantitative research method is going to be applied in this paper since its ability of measuring judgment in numerical digits to transformed into valuable and insightful outcomes which ensures the reliability and validity (Cooper & Schindler, 2006).

Firstly, the sample size of at least 250 is proposed in order to generate good research outcome (Comrey & Lee, 1992). This study has a total of 313 eligible survey answers. Therefore, it is qualified to proceed with data analysis.

Considering its advantages, convenience sampling was suitable for this research. Ho Chi Minh International University (HCMIU) HRM students of all classes; as well as Career builder Vietnam, Intel Products Vietnam, TMA Solutions, VNG Corporation, and iStar English center - my current and former workplaces were chosen to be sampling location. Convenience sampling is defined as a non probability sampling that any willingly available persons are considered as respondents (Cooper & Schindler, 2006). Considering its advantages, convenience sampling was suitable for this research. All of these target groups of respondents are familiar, interested or even experienced with e-recruitment, making them highly suitable and significant for the questionnaire.

Data for this study is collected in the form of questionnaire, through online survey (like Google Form), and conducted from middle of April to June in 2020.

3.2. Questionnaire Design

Five-point Likert scale (5 = strongly agree, 1 = strongly disagree) is applied to conduct this study. The questionnaire includes four parts. First part is the filtering session to ensure the qualification of the respondents. Moving on to the second part, the group of factors including four independent variables and one ultimate dependent variable will be in the third part. Lastly, the fourth part focuses on demographic questions to collect personal information of respondents’ demographic characteristics. Table 1 shows the measurement items for the research model.

Table 1: Measurement items for the research model

3.3. Data analysis method

Statistical Package for the Social Sciences (SPSS Statistics 20) software is used in the data analysis process, which starts with demographic statistics, to descriptive statistics and reliability test of all variables. Exploratory Factor Analysis (EFA) is also used to test if any items should be deleted due to low factor loadings. Multiple Regression analysis is also applied as the final test.

After collecting raw data from respondents, data will be coded and screened for errors. SPSS (Statistical Package for the Social Sciences) statistical software version 20 is applied in the data analysis. The process follows these major analyses:

Demographic statistics: Including frequency, percent, valid percent and cumulative percent

Descriptive statistics: General metrics of each variable

Reliability analysis: Cronbach’s Alpha is applied. Reliability Statistics present how well the items in a set are positively correlated to one another.

Correlation

The correlation coefficient Pearson (r) is used to measure the relationship between two or more variables. The absolute value of r approaches 1 when two factors have a close linear correlation

Exploratory Factor Analysis (EFA)

EFA is used to measure the value of the scale. When analyzing EFA, some criterion need to be considered are KMO, Bartlett’s test, Eigenvalue, cumulative percentage and factor loading, which classify the relationship between variables and determine the numbers of common factors that affect a set of measure. Factor loading is an indicator to ensure the practical significant level of EFA.

Multiple linear regression

Linear regression is a statistical approach allowing to summarize and examine relationships between response variables and predictor variables. This model helps us to understand how changes in the independent variables are correlated with changes in the dependent variable. Besides, we can also use regression model to make prediction thanks to the values of the explanatory variables.

4. Data Analysis and Finding

4.1. Demographic Statistics

Offline responses are eliminated because the respondents prefer online platform. As a result, the study is conducted on a set of 313 qualified responses, with a 100% response rate.

The largest portions taken by 18-24 years old group (group 1) occupying 44.4% which usually include recruitment internship or junior level. Accounting of 29.4% of total respondents is group 2 – people from 25-34 years old. The other two are under 18 groups with 20.4% of the total number and above 35 years old group with only 6% portion. This result was predictable due to limitation of resource of the research. Persons with high level position such as Human Resource Directors, Recruitment Managers, and so on frequently fall into these segments which are group 3 and 4are not willingly to response the survey to give a more insightful overview of the current situation. However, with around 74% of respondents of two other groups, the study is still able to generalize the whole population of employers and recruiters.

It can be seen that level 1 occupies largest number of respondents with 61.3% of total. This group includes positions from senior down to intern. As predicted in discussion of table of age distribution, with only 38.7%, middle level managers or above are not so willingly to do the research with due to the consideration of delivering sensitive information.

4.2. Reliability Analysis

Cronbach's Alpha test is necessary for internal consistency assessment. Unsuitable items are going to be eliminated one by one (George & Mallery, 2003).

Item-total correlation was taken carefully and has to be greater than 0.3. Reliability test was conducted for each factor and the result showed in the Table 2 below.

Table 2: Result of Reliability test

The above table summarizes the Reliability analysis statistics of all variables in the measurement. Accordingly, the Cronbach's Alpha of all variables is higher than 0.7, which is the benchmark value for “Good” internal consistency level. Moreover, no item has the Cronbach's Alpha if Item Deleted value lowers than 0.3, which proves that there is no need to eliminate any item at this step. The items within each factor also do not exceed its respective Cronbach’s Alpha value. These results strengthen the conclusion that there is a strong internal consistency in the established and well-designed measurement. And the scale is eligible to proceed with further data analysis.

Exploratory Factor Analysis

It can also be seen that the Rotation Sums of Squared Loadings’ Cumulative percentage is 66.870%, which is higher than the required level (50%), indicating that these two extracted factors give explanation for more than 50% of total variance in the 12 independent variables. Besides, the Eigenvalues of these two considered components are acceptable as they are greater than 1. In sum, this measurement is valid and eligible for further analysis.

Correlation and Pearson

According to the Table 3, PEOU relates to CSE since the significant outcome is smaller than 0.05. The similar situation happens with BI and other factors including PPR, PEOU, PU. Therefore, Multiple Regression step can be conducted.

Table 3: Correlation

**. Correlation is significant at the 0.01 level (2-tailed).

*. Correlation is significant at the 0.05 level (2-tailed).

Furthermore, as can be seen from the result, independents factors also have significant relationship with behavioral intentions, however, due to limitation of resources, this paper aim to factors provided of previous studies in spite of seeking for new variables. Future research can apply this outcome to conduct further explanation or exploration of new factors.

Multiple Regression

This process is going to be taken twice due to 2 level of simple dependent variables. The first level of relation includes proposed hypothesis of Computer Self-Efficacy and Perceived Ease of Use. The second level includes hypotheses of main dependent variable Behavioral Intentions to Use with two independent variables Perceived Usefulness and Perceived Privacy Risk and one minor dependent – Perceived Ease of Use.

Multiple Regression for PEOU

According to the result, Adjusted R squared is 0.368 which mean CSE explains 36.8% of the change of PEOU generally. However, this process of multiple regression only doing on one independent variable CSE, therefore, standardized coefficients can be used to explain the portion of effect which means Computer Self-Efficacy explains 60.8% of Perceived Ease of Use’s fluctuation from Table 4 to Table 7.

Table 4: Model Summary of Computer Self-Efficacy

a. Predictors: (Constant), ACSE

Table 5: Anova of Computer Self-Efficacy

a. Dependent Variable: APEOU

b. Predictors: (Constant), ACSE

Table 6: Coefficients of Computer Self-Efficacy

a. Dependent Variable: APEOU

Table 7: Collinearity Diagnostics of Computer

a. Dependent Variable: APEOU

Multiple Regression for BI

According to the results which include Adjusted R Square is 0.222 which means independents variables explain 22.2% of the change of BI.

In addition, the Coefficients table explains clearly that PU factor has major effect on Perceived Ease of Use which is 36.2% while Perceived Usefulness and Perceived Privacy Risk only describe 21.3% and 16.4%, respectively, the change of Behavioral Intention to Use factor. This outcome is not expected and predictable due to the significance and well tested result at reliability step. However, it is totally understandable because of different context of previous studies compared to this paper (from Table 8 to Table 11). The proposed hypotheses are supported as below:

Table 8: Model Summary of Behavioral Intentions to Use

a. Predictors: (Constant), APU, APEOU, APPR

Table 9: Anova of Behavioral Intentions to Use

a. Dependent Variable: ABI

b. Predictors: (Constant), APU, APEOU, APPR

Table 10: Coefficients of Behavioral Intentions to Use

a. Dependent Variable: ABI

Table 11: Collinearity Diagnostics of Behavioral Intentions to Use

a. Dependent Variable: ABI

5. Conclusions

5.1. Limitations

It provides some indications of the employers’ intent of job distribution technology adoption, which can be replicated in other countries using the same model and instrument to identify and consolidate employers’ perceptions and Behavior toward this technology adoption.

Furthermore, due to lack of time and resources, this study cannot cover all of independents factors which as mentioned above may contribute to the dependents.

On the other hands, sensitive information is not appropriate to collect by this type of survey or method without any support from well-known expertise which also leads to the doubtful reliability in several hypotheses.

Regarding the conflict with previous studies, this paper eliminated several hypotheses that are considerably unrelated or due to its complexity such as modifying Technology Acceptance Model, excluding Perceived Ease of Use to Perceived Usefulness.

5.2. Contributions

Taking advantage of the internet in the recruitment process is a rising trend for both employers and employees. This study provides empirical findings of the experience and perception of e-recruitment platforms in the modern search for job distribution of candidates. Accordingly, this paper demonstrates multiple key indicators to the adoption of e recruitment tools, which are essential to the human resources literature in general, and to recruitment in specific.

The construction of PEOU demonstrates that employers are capable of understanding and getting acquainted with the operation of e-recruitment technology over a short period of time. As the development of technology and internet, more and more intelligent labor force who are familiar with job boards created. Therefore, PEOU could not play an important role as it used to be in the past as CSE rapidly becomes a must-have skillset to compete in labor market. Considering the positive relationship between CSE and PEOU, a considerable effect on BI is statistically proven. In other words, the construction of online job distribution indicates that there is a great importance of communication in the process since it can affect applicant’s perception and feeling while choosing suitable vacancy. Moreover, timing is a matter of importance for both recruiters and employers, especially under the conditions that jobs are changing faster than ever before (Table 12).

Table 12: Summary results of hypotheses

Employers used to perceive usefulness in technology for hiring process is strongly important, and it indicates that using this system lead them to better out comes. In this research, it is clearly that Perceived Usefulness does not have much effect on BI. This may due to the lack of external variables which determine job board users’ decision. As an employer or recruiter, it is much more complicated to choose a specific job board for working purpose. There are several other limitations origin from work place such as corporation’s resources or strategy of direct management level which leads to undeniable choice for staff level to execute their tasks.

5.3. Conclusions and recommendations

The technology adaptation of job distribution is still a controversial issue due to its access to privacy of users. Therefore, employers should be more careful on how their personal contacts are going to be delivering without consent. Secondly, keeping your systems up to date and “connected” is essential for both employees and employers - an open gate for “uninvited guests” if cyber security not taken seriously.

For future studies, those external variables should be listed in conceptual framework for a more accurate result. In addition, Board of Directors or direct middle managers need to pay more attention on how their workforce performs through various channels of job distribution – both formal and informal, such as Linkedin, Facebook rather than a few compulsory job boards which, as the result explain, no longer being accessed with high performance expectancy. Behavioral Intention’s variance experiences an occupation of small portion explain the reason that job boards will not become the only method for candidates, especially recruiters to handle recruitment procedure but a combination with conventional ways of job distribution.