Social Networking Information and Pre-Employment Background Checks: The Mediating Role of Sustainable Organizational Reputation

Abstract

This study was aimed at identifying the role of social networking information (SNI) on pre-employment background checks (PBCs) of job applicants involving a single mediating variable, sustainable organizational reputation (SOR). In addition to that, we conducted a multi-group analysis to examine whether the direct and mediating effects were stronger in terms of gender difference. For investigation, a structured questionnaire was utilized to collect data from 1138 respondents working as hiring professionals in Bangladesh applying purposive sampling. Finally, 798 valid samples were considered. Structural equation modeling was applied using AMOS 24 to test the hypothesized relationships. The results indicated that SNI is positively and significantly related to PBCs. Additionally, it was revealed that SOR can partially mediate the direct relationship between SNI and PBCs. Furthermore, the study did not reveal any noteworthy gender differences influencing these relationships. The paper is expected to assist academicians involved in investigating pre-selection background checks and hiring professionals looking for further academic references to formulate a proper social media background policy.

1. Introduction

Competent employees play a vital role in achieving organizational performance and are considered as the basis for viable advantages over rivals [1]. For obvious reasons, organizations are eager to recruit the right people for the right positions. In this regard, social networking sites (SNSs) such as LinkedIn, Facebook, Twitter, and Google+ have been receiving special attention by academicians, practitioners, and human resources (HR) professionals [2]. SNSs are online communication and sharing platforms where users can share and comment on posts, photos, messages, videos, and events, and indicate “like/love.” These platforms are now increasingly being utilized as common platforms for some HR functions such as recruitment, selection, and, most recently, pre-employment background checks (PBCs) by potential employers [3].

The emergence of a new hyperlinked information world became evident after 2005 due to the rise of a number of SNSs such as Facebook, LinkedIn, Blogs, Google+, Twitter, Xing, etc. [4]. According to Internet Live Stats [5], user numbers total up to 1.5 million on Facebook, more than 4 million on Google+, up to 4 million on LinkedIn, and 3.3 million on Twitter. Due to the large volume of users, mostly young people, organizations have started to utilize SNSs for various purposes, such as promotion, selling, recruitment, and even background checks. In particular, SNSs have been widely extended to HR practices [6]. A study conducted by the Society for Human Resource Management (SHRM) [7] revealed that 81%, 75%, and 45% of employers used LinkedIn, Facebook, and Twitter, respectively, for personnel recruitment.

Although SNSs are becoming popular among HR professionals as recruitment and selection tools, academic research in this domain is still scant. Moreover, even though a handful of studies have been conducted investigating the role of SNSs inhuman resource management (HRM) activities such as recruitment and selection, almost all of them were from a Western perspective. Further, such studies were limited to recruitment and selection or investigating the perceptions of candidates, rather than looking at topics such as pre-employment background checks. There is a considerable absence of such studies in developing and underdeveloped nations. In particular, no studies regarding the role of social networking information (SNI) on PBCs has been conducted in South Asian countries.

Considering this gap, this study intends to identify the role of SNI in pre-employment background checks of job candidates in Bangladesh, the 8th most populated country in the world with more than 50% of young people, out of which the majority are young jobholders or are looking for jobs [8]. Moreover, this study adds a single mediator, sustainable organizational reputation (SOR), to examine how this variable influences the relationship between SNI and PBCs. This is the first study conducted in Bangladesh on SNI and PBCs and puts forward the following research questions:

RQ1: 

What is the role of social networking information in PBCs in Bangladesh?

RQ2: 

How can organizational reputation mediate the relationship between SNI and PBCs?

The paper has been divided into several parts. The first section introduces the study and puts forward the research questions. Section 2 reviews the literature, develops research hypotheses and theoretical framework. Section 3 discusses the research methods, while the next one discusses the results obtained after statistical analyses. Section 5 interprets the results and Section 6 highlights the rationale and significance of the study. The last two sections highlight the limitations and further scope of study and conclude the paper with some recommendations respectively.

2. Literature Review, Research Hypotheses and Theoretical Framework

2.1. Social Networking Sites and Social Networking Information

Social networking sites (SNSs) are online platforms that permit users to generate profiles, connect with other users, and share information [9]. SNSs incorporate, but are not limited to, blogs, picture-sharing platforms such as Tumblr, Internet forums and sites such as Facebook and Twitter [10]. In general, members of SNSs can generate individual profiles by signing up and uploading the contents such as biographical information, images, videos, posts and much more [9]. Users have their own networks where they have access to people they are connected with and can view information posted by them, popularly known as “friending” in the language of SNSs. The extent of information a user can view depends on the sharing friend’s privacy settings. Some information are publicly available (everyone can see such information even not connected to that user), some can be viewed only by friends while some contents are not viewable to others, other than the profile owner. SNSs can be recreational (Facebook), job-oriented (LinkedIn), or for posting statements (Twitter). These SNSs are a source of a huge amount of information, which is the central attraction of many third-party users such as organizations that want to promote their products or businesses [11,12], Governments that want to observe and control the flow of information and, very recently, HR professionals who want to recruit and select potential candidates as well as check their backgrounds.

2.2. Pre-Employment Background Checks: The Basics

An employee background check, also referred to as a pre-employment background check (PBC), is the process of looking up and compiling personal, criminal, financial and job records of a potential employee. Background checks in Western countries are often conducted by screening agencies or third parties that are especially efficient in this regard. In Bangladesh, for government organizations, criminal and personal records are screened by local police stations. Financial records are checked by financial bureaus. However, for private organizations, a PBC is mostly conducted by the organization offering the job. Presently, a PBC also includes personality, taste, working relationships, employment history and many other factors related to the personal and professional history of job candidates.

A PBC is necessary for several reasons. First of all, it allows an employer to justify the information that job applicants provide on their resumes. As the candidates might be future employees of the organization, it is imperative for hiring professionals to be confident that the provided information are correct. Second, a recruiter who is unsuccessful at executing a systematic PBC for a potential staff member might be accused of negligent hiring or service discrimination [13]. Recruiters might also be accused of negligent hiring if they are not able to perform background investigations on potential staff members who were involved in criminal activity, or who injured a client or third party member while performing their job responsibilities [13,14,15,16]. Finally, a handful of studies recommend that, without proper background screening, hiring professionals might use race, gender and additional apparent correlates of unlawful action to evaluate the likelihood of a candidate’s preceding convictions and use such judgments to influence employment decisions [13,17].

Although PBC does not guarantee that a candidate will not do something unusual or against the interest of the organization or perform well, a valid argument can be made that the candidate was properly screened and judged, which allows the hiring professional to avoid charges or accusations of negligent hiring.

2.3. The Role of Social Networking Information in Pre-Employment Background Checks

The main purpose of the present investigation is to explore the role of social networking information in pre-employment background checks of job applicants. Due to the rising attractiveness of SNSs such as Facebook, LinkedIn and Twitter, increasing numbers of HR professionals and hiring managers are interested in utilizing the information on these sites [9] in order to evaluate job candidates [9,18,19,20,21].

The practice of utilizing SNSs to improve hiring efficiency [22,23,24,25] is likely to be omnipresent among recruiters. Around 93% of hiring managers have utilized, or are currently utilizing, SNSs during recruitment and selection stages and 55% have reevaluated applications based solely on the contents published on applicants’ SNS profiles [26]. After scrutinizing such information, hiring professionals try to accumulate a complete summary of potential future employees that would otherwise be unavailable, particularly since such information is related to individual personalities [27,28]. Moreover, many hiring professionals evaluate job candidates using SNSs as a supplementary source to assess best matches with the organization or to identify “red flags” [23]. Such “red flags” may comprise SNS contents related to a job candidate’s personal life, such as consumption of alcohol or prohibited drugs, vulgarity, or involvement in sexually implicit actions [25]. Recruiters frequently use such information to remove candidates being considered in hiring pools [22]. Indeed, 89% of hiring managers said they would be very unlikely to employ individuals who have evidence of “unprofessional behavior” on their social networking profile [23].

Although organizations have accountability in terms of remaining fair and thorough in the hiring process in order to avoid negligent hiring, the likelihood of misusing the information obtained from SNSs is also high [18,19]. With the use of information from SNSs such as Facebook to evaluate job applicants, which is generally considered as personal social networking site, organizations could violate confidentiality rights by holding deceptive or wrong information regarding candidates, or potentially confidential information that hiring managers are unable to query because of pertinent service laws [9,18,19,20].

The use of SNSs by hiring managers to examine job candidates is a swiftly increasing practice [29], as recruiters require additional information beyond the resume and want to avoid negligent hiring [30]. In a survey of 300 recruiters conducted by Swallow [31], 69% of hiring professionals said they discarded job applicants based on SNS profile contents. Furthermore, numerous candidates were asked to give information that permitted potential employers to have complete access to SNS profiles [32]. In spite of the growing popularity of SNI for conducting PBCs, whether or not such practices are lawful or proper is still a matter of debate. Employers might hope for the possibility to review SNSs; however, the unregulated use of such a screening practice brings up numerous ethical issues that have to be addressed [9]. Many employers applaud a social networking PBC as it allows them to accumulate a great deal of information regarding job candidates, creating the possibility to successfully connect candidates and jobs. However, commentators have pointed out that employers must be careful in using SNI to screen applicants [33]. Due to the rising concerns regarding the trustworthiness and validity of SNS information, alongside the potential for abuse, privacy specialists persuade employers to look elsewhere for individual data rather than social networking sites [33].

Hypothesis 1 (H₁). 

SNI has a significant positive relationship with SOR.

Hypothesis 2 (H₂). 

SOR has a significant positive relationship with PBCs.

Hypothesis 3 (H₃). 

SNI has a significant positive relationship with PBCs.

2.4. Further Extension of the Study: Sustainable Organizational Reputation

This study employs a mediating factor, sustainable organizational reputation, to investigate how it influences the relationship between SNI and PBCs.

Organizational reputation as a theme has been defined and established from organizational branding theory, resulting from the application of ideas about affecting customers to HRM and its influence on potential and present workers [34]. Such reputation can sustain if the organizational goals and policies are compatible with the customers’ (potential employees) preferences, even if they face risks and uncertainty. The objective of such reputation management is to accumulate information and thoughts and to understand, forecast and deal with them in such a manner that sustainability and a strong corporate profile can be managed. There are growing insights regarding the outward impact of sustainable organizational reputation [35] in drawing the attention of potential employees [36], increasing hiring efficiency [37] and lowering the cost of hiring [38]. Studies have confirmed that how potential candidates perceive an organization’s corporate social responsibility (CSR) has an impact on its attractiveness to them [35]. Organizations have experienced that effective and strong brand reputation creates certain competitive advantages and contributes to easily attracting new and retaining existing employees [39]. Research further confirms that companies that undertake efforts to establish a positive image are seen by candidates as first-choice employers [40]. This signifies that a company has a strong and recognized sustainable organizational reputation and offers the possibility of first-rate employment [41].

Therefore, there are reasons to believe that an organization with a strong and sustainable brand/organizational reputation in the job market has more appeal to potential employees. As a result, SOR can have a positive influence on the use of SNI for PBCs. Employers can also use the privilege of asking for SNS information from candidates for background checks based on this strong influence of SOR.

Hypothesis 4 (H₄). 

SOR can mediate the relationship between SNI and PBCs.

2.5. Using SNI for Background Checks: The Bangladesh Context

Unfortunately, there is no previous research regarding the use of SNI in Bangladesh. Therefore, this research is the first of such empirical experiments. Bangladesh is a highly populated country, with a population of 170 million [42]. SNSs are becoming very popular in Bangladesh and are regularly used by most of the Internet generation (particularly young people) to communicate with families and acquaintances. In addition to individual use, companies seem to be enthralled with the supremacy of web-based social networking tools to create official communication and an online presence. In some cases, such sites, especially Facebook [43] and LinkedIn, are used for online marketing and by professionals in various organizations, respectively. However, in Bangladesh, studies, to date, have mostly been conducted on the use and influence of social networking by users, while limited studies [44] have been conducted on employer and employee perspectives, particularly in the context of Bangladesh.

In Bangladesh, SNSs are under the strong supervision of the Bangladesh Telecommunication Regulatory Authority (BTRA), a department of the Ministry of Postal and Telecommunication. In the Bangladesh Labor Act of 2012 and Bangladesh Information Act of 2017 (amended), there is nothing mentioned regarding the background checks based on SNI. Therefore, it is not illegal, as the sites are meant to be open for everyone who is connected (friends or groups). The practice is, in case of HR functions, to date, [44] limited to talent hunting and is not used for background or pre-employment history checks. However, very recently, some government professions such as the Bangladesh Civil Service, police, the mint service, the press and reputed foreign firms operating in Bangladesh have started background checks based on SNI, along with traditional methods such as local police verification [45].

Therefore, it is becoming a growing concern among young people (consisting more than 50% of the total population) in Bangladesh as to what extent and how the information in social networking sites should be used for background checks, addressing issues of ethics, fairness and privacy. The proposed study is aimed at finding the statistical relationship between SNI and PBCs, considering three major concerns among candidates.

In addition to study the above mentioned relationships, authors made a further attempt to identify whether there is any gender influence over these relationships. Therefore, three additional hypotheses were developed to be tested:

Hypothesis 5_A (H_5A). 

The positive correlation between SNI and SOR is stronger for women.

Hypothesis 5_B (H_5B). 

The positive correlation between SOR and PBCs is stronger for women.

Hypothesis 5_C (H_5C). 

The positive correlation between SNI and PBCs is stronger for women.

2.6. Theoretical Framework

The purpose of this empirical study is to investigate the role of social networking information on pre-employment background checks (PBCs) of job applicants using one mediator, sustainable organizational reputation (SOR). The theoretical model is presented in Figure 1.

3. Research Methods

3.1. Sampling Region

The investigation data were collected from hiring professionals working at different organizations in the Dhaka and Chittagong districts of Bangladesh. The reason for choosing those two cities was that most of the head offices of various organizations are located there.

3.2. Sampling Size and Data Collection Technique

The study used primary data accumulated with the help of a well-designed research instrument (questionnaire). By using the convenience sampling method, 1138 questionnaires were sent out to different organizations. Additionally, 321 volunteers have been recruited for this investigation having an active social networking account who agreed to let the researchers using their profiles. The respondents were requested to fill in the survey instruments only after visiting the profiles of those volunteers. However, after collecting the data, incomplete, biased, and abnormal questionnaires were discarded through a scrutinizing process and finally 798 valid samples were used for the analysis. For descriptive analysis, SPSS (Version 24) and for testing the hypothesized relationships, structural equation modeling (SEM), AMOS (Version 24) has been utilized. Before testing the relationships through SEM technique, a number of tests were conducted rather than just jumping to SEM.

3.3. Determination of Sample Size

The respondents were selected from different organizations in Bangladesh as the hiring professionals. The total number of respondents was calculated by adopting the following formula, recommended by Yamane [46]:

$$n=\frac{N}{\left(1+N{e}^2\right)}$$

Where n is the sample size, N is the population size, and e is the precision level (here = 5%). The following assumption was made while calculating the sample size:

n = 400 (here, n = 798). Population size (N) is >100,000 respondents.

3.4. Questionnaire Design

A well-structured survey questionnaire was used to collect data from respondents using a 5-point Likert scale, where 1 indicates strongly disagree and 5 indicates strongly agree. After conducting reliability and validity tests of the questionnaire, the numbers of items for each variable were determined, as shown in

Table 1. Numbers of items before and after reliability and validity test.

Variable	Number of Items
	Before Reliability and Validity Test	After Reliability and Validity Test
SNI	7	7
SOR	7	7
PBC	6	5
Total	20	19

SNI: social networking information; SOR: sustainable organizational reputation; PBC: pre-employment background check. Source: Survey instrument.

.

The interviewers were appropriately instructed regarding the questionnaire items before starting of the interview process.

3.5. Normality of Data

A fairly normal distribution of our indicators of the latent factors was observed in terms of skewness. However, we observed mild kurtosis for the indicators of the dependent variable (PBC) and mediator (SOR). These kurtosis values ranged from −0.90 to 2.5. While this violates strict rules of normality, it is within more relaxed rules, as suggested by Sposito et al. [47] who recommend 3.3 as the upper threshold for normality (

Table 2. Normality of data.

List of Factors	Descriptive Statistics
	N	Skewness	Kurtosis
The information in social networking sites is insufficient for background checks.	798	0.271	−0.426
People often provide inaccurate information on social media regarding age and job experience.	798	0.092	−0.840
I do not believe the information provided on social networking sites.	798	0.299	−0.480
The information on social networking sites can sometimes be misleading.	798	−0.262	−0.724
It is expected that Facebook will provide mostly behavioral information and that LinkedIn will provide mostly job-related information on candidates.	798	−0.078	−0.918
Social networking information can be used for background checks as a supplement to other measures.	798	−0.104	−0.927
It is somehow beneficial to use social networking information for background checks.	798	0.101	−0.616
My organization is reputed in the industry.	798	−0.698	1.004
Due to the sustained reputation, candidates are attracted to my organization as a place to work.	798	−0.709	1.365
Sustainable organizational reputation creates a favorable image for job candidates regarding the organization.	798	−0.600	0.929
Job candidates will be willing to provide their SNS username and password if the organization is a reputed one.	798	−0.598	1.095
As a reputed organization, I think it is justified to ask for candidates’ SNS ID and password for a proper pre-employment background check.	798	−0.544	0.724
I believe pre-employment background checks based on social networking information should be reliable and valid to avoid negligent hiring.	798	−0.879	2.046
As a reputed organization, users’ privacy and fairness concerns are important to us when using their social networking information.	798	−0.357	−0.078
As a reputed organization, we pay the highest attention to conducting pre-employment screening of future employees before offering jobs.	798	−0.739	0.452
Social media is a good platform to get ideas about candidates’ behavioral information that is not otherwise available.	798	−0.793	0.802
I am well aware of therisks of using information on social networking sites for background checks.	798	−0.779	0.771
My organization has established policies and guidelines for when social media checks will be conducted, by whom, on which sites, what information, and how that information will be evaluated.	798	−0.766	0.486
My organization prohibits anyone besides authorized employees from conducting background checks.	798	−0.575	0.053
My organization conducts social media background checks only after taking substantial precautions.	798	0.048	−0.693

SNS: social networking site. Source: Survey data.

).

The data showed a fairly normal distribution that permitted the researchers to continue the study.

3.6. Tests of Reliability and Validity

To examine the reliability of the data, this study adopted composite reliability, the average variance extracted (AVE) from the data, which is highlighted in

Table 3. Model validity measures.

	CR	AVE	MSV	MaxR(H)	F1	F2	F3
F1	0.911	0.596	0.045	0.921	0.772
F2	0.939	0.687	0.045	0.946	0.212 ***	0.829
F3	0.851	0.541	0.021	0.881	0.118 **	0.146 ***	0.735

Significance of correlations: p < 0.100, * p < 0.050, ** p < 0.010, *** p < 0.001. F1: Social networking information; F2: Sustainable organizational reputation; F3: Pre-employment background check. CR: composite reliability; AVE: average variance extracted; MSV: Miller-Siegmund values. Source: Test of discriminant validity.

and indicates that all data are within the cut-off value of acceptable limits, as per Hair et al. [48], Fornell and Larcker [49] and Henseler et al. [50].

In order to check discriminant validity, this study used Fornell and Larcker’s [49] principle of comparing the AVE value with equivalent correlation values with other variables. The square root values of AVE must be higher than the corresponding correlation values with other variables [51]. The discriminant validity of the factors is shown in Table 3.

4. Results and Discussion

4.1. Industry-Wise Sample Distribution

The respondents were selected intentionally from different organizations in a variety of sectors to make a sample group that well represented the whole job market in Bangladesh.

Table 4. Industry-wise distribution of respondents.

Industry	Number of Firms	Number of Respondents
Manufacturing	41	144
Telecommunication	34	122
Bank and insurance	33	105
Education	31	97
Information technology	27	86
Transport	25	69
Shipping	22	47
Postal service	19	43
Print and electronic media	17	32
Food and beverage	14	29
Hotel and tourism	11	24
Total	274	798

Source: Survey instrument.

highlights the industry-wise sample distribution for the investigation.

4.2. Demographic Information of the Respondents

Table 5. Demographic information on respondents based on factors in the questionnaire.

Particulars		Frequency	Percent
Age	18–23	40	5.01
	24–29	80	10.03
	30–35	287	35.96
	36 and above	391	49.00
	Total	798	100
Gender	Male	418	52.38
	Female	380	47.62
	Total	798	100
Social media background checks	1 (Yes)	785	98.37
	2 (No)	13	1.63
	Total	798	100
Frequency of social media background checks	1 (Often)	626	78.45
	2 (When needed)	102	12.78
	3 (Rarely)	70	8.77
	Total	798	100
Organizational social media account	1 (Facebook)	56	7.02
	2 (LinkedIn)	398	49.87
	3 (Both)	344	43.11
	Total	798	100
Separate HRM department	1 (Yes)	795	99.62
	2 (No)	3	0.38
	Total	798	100
Education level	1 (Undergraduate)	89	11.15
	2 (Graduate)	111	13.91
	3 (Postgraduate)	598	74.94
	Total	798	100

HRM: human resources management. Source: Demographic components of the questionnaire.

shows the distribution of sample respondents based on other demographic factors included in the questionnaire such as age, gender, education, possession of a social media account, and the kind of social media account (Facebook or LinkedIn), as well as the organization’s social media background checks and frequency of background checks and whether it has a separate human resources department.

Table 5 indicates that almost half of the participants (49%) were aged 36 or above and over half of them were men; 74.94% were postgraduates and all had organizational social media accounts (LinkedIn, Facebook, or both). A further investigation revealed that all of them conducted organizational social media background checks on job applicants; 78.45% of them did so frequently and only 8.77% did it rarely. Almost all of the organizations (99.62%) had a separate HR department.

4.3. Results of Exploratory Factor Analysis

To evaluate the exploratory factor analysis (EFA), four frequently used assumptions were adopted [48,52]: (1) a sampling adequacy (Kaiser–Meyer–Olkin) measure (greater than 0.5); (2) the lowest Eigen value for individual factors; (3) a sample size, where a factor loading of 0.40 for every item was treated as the threshold for retaining items to guarantee greater confidence; and (4) Promax rotation, as it is a good universal approach that simplifies the interpretation of factors [52].

Table 6. Exploratory factor analysis of social networking information in pre-employment background checks.

List of Variables	Pattern Matrix
	Factor
	1	2	3
The information on social networking sites is insufficient for background checks.	0.636
People often provide inaccurate information on social media regarding age and job experience.	0.846
I do not believe the information provided on social networking sites.	0.757
The information on social networking sites can sometimes be misleading.	0.762
It is expected that Facebook will provide mostly behavioral information and LinkedIn will provide mostly job-related information on candidates.	0.855
Social networking information can be used for background checks as a supplement to other measures.	0.883
It is somehow beneficial to use social networking information for background checks.	0.679
My organization is reputed in the industry.		0.808
Due to sustained reputation, candidates are attracted to my organization as a place to work.		0.854
Sustainable organizational reputation creates a favorable image for job candidates regarding the organization.		0.913
Job candidates will be willing to provide their SNS username and password if the organization is a reputed one.		0.932
As a reputed organization, I think it is justified to ask for candidates’ SNS ID and password for proper pre-employment background checks.		0.832
I believe pre-employment background checks based on social networking information should be reliable and valid to avoid negligent hiring.		0.765
As a reputed organization, users’ privacy and fairness concerns are important to us when using their social networking information.		0.715
As a reputed organization, we pay the highest attention to conducting pre-employment screening for any future employee before offering a job.			0.820
Social media is a good platform to get ideas about candidates’ behavioral information that is not otherwise available.			0.787
I am well aware of the risks of using information on social networking sites for background checks.			0.845
My organization prohibits anyone besides authorized employees from conducting background checks.			0.667
My organization conducts social media background checks only after taking substantial precautions.			0.515
Extraction method: Maximum likelihood. Rotation method: Promax with Kaiser normalization.
a. Rotation converged in four iterations.
KMO = 0.903; Bartlett’s test = 0.00001059; Sig.: 0.000.

KMO: Kaiser–Meyer–Olkin. Source: Exploratory factor analysis.

shows the EFA results. Hair et al. [53] suggested that factor analysis can be performed when the Kaiser–Meyer–Olkin (KMO) test and Bartlett’s test of sphericity are significant. An index of the KMO measure of sampling adequacy (overall MSA = 0.903) and Bartlett’s test of sphericity (p = 0.000) showed that the factor analysis was suitable for further analysis of the data in this study (Table 6). After examining the EFA pattern matrix, we found that all items had factor loadings higher than 0.50.

According to the results, the factor analysis is appropriate. After confirming research constructs, maximum likelihood (ML) and the promax method (PM) were used to extract factors from 19 items. Hair et al. [53] recommended that item factor loading must have a value more than 0.50 to be considered highly significant. On the basis of an Eigen value greater than 1, a three-factor model was identified that explains 65.954% of the total variance of the dataset (

Table 7. Total variance explained.

Factors	Initial Eigen Values			Extraction Sums of Squared Loadings
	Total	Variance (%)	Cumulative (%)	Total	Variance (%)	Cumulative (%)
F1	6.124	30.621	30.621	5.610	28.048	28.048
F2	3.936	19.679	50.300	3.712	18.560	46.608
F3	3.131	15.654	65.954	2.717	13.585	60.193

F1:Social networking information; F2:Sustainable Organizational reputation; F3:Pre-employment background checks. Source: Exploratory factor analysis.

). As a whole, 19 items were grouped (based on Eigen value) into the three factors (SNI, SOR, and PBC) by the analysis. The EFA results also showed that 0.515 was the lowest and 0.932 was the highest factor loading of the items, and that all factors, as a whole, were acceptable for further analysis (Table 6).

4.4. Results of Confirmatory Factor Analysis

Confirmatory factor analysis (CFA) is a statistical procedure that can confirm the factor composition of a set of observed variables, enabling researchers to assess whether an association exists between observed variables and their core underlying constructs. The results of CFA have been highlighted in Figure 2 and

Table 8. Model fit indices and their acceptable thresholds.

Goodness of Fit Index	Value	Level of Acceptance	Reference
Chi-square/Df	2.036	<5.0	Marsh and Hocevar (1985)
CFI	0.986	>0.90	Bentler (1990)
RMR	0.016	<0.08	Hu & Bentler (1998)
GFI	0.963	>0.90	Joreskog & Sorbom (1993)
AGFI	0.951	>0.85	Anderson & Gerbig (1984)
RMSEA	0.036	<0.08	Browne & Cudeck (1993)
SRMR	0.0278	<0.08	Browne & Cudeck, 1993

Df: degree of freedom; CFI: comparative fit index; RMR: root mean square residual; GFI: goodness of fit index; AGFI: adjusted goodness of fit index; RMSEA: root mean square error of approximation; SRMR: standardized mean square residual. Source: Literature review.

. The relative chi-square for this model was 2.036, which is smaller than the 5.0 recommended by Marsh and Hocevar [54], while other fit indices also showed a superior fit for the measurement model. The goodness of fit index (GFI) of the model was 0.963, which is more than the recommended value of 0.90 suggested by Joreskog and Sorbom [55]. In the present study, the adjusted goodness of fit index (AGFI) was found to be 0.951, which meets the recommended value (>0.85) and hence was deemed to be a good fit and acceptable, as supported by Anderson and Gerbig [56]. Additionally, the non-incremental comparative fit index (CFI) was 0.986, which exceeds the suggested cut-off level of 0.90 [57]. In CFA, the root mean square residual (RMR) value was found to be 0.016, which is less than 0.08, commonly recommended as acceptable [58].

The root mean square error of approximation (RMSEA) was 0.036, which is also less than the recommended value for proper fit [59]. Finally, the standardized mean square residual (SRMR) was 0.0278, which is less than the 0.08 recommend by Browne and Cudeck [59]. A summary of the analysis is shown in Table 8. Overall, all fit indices represent an adequate model fit to the data.

4.5. Common Method BiasTest

Harman’s single factor test can be utilized to identify whether the majority of variance can be explained by a single factor. If a single factor cannot explain the majority of variance, it can be assumed that the common method bias (CMB) test has not been met.

Table 9. Common method bias test results.

Total Variance Explained
Factor	Initial Eigen Values			Extraction Sums of Squared Loadings
	Total	% of Variance	Cumulative %	Total	% of Variance	Cumulative %
1	6.124	30.621	30.621	5.610	28.048	28.048

Source: Common method bias test.

indicates that one factor can explain only 30.621% variance in this study, which means that the CMB test did not occur.

4.6. Multicollinearity Test

Variance inflation factors (VIFs) range from 1 to 10 and upward. The VIF tells us what percentage of variance is inflated for each coefficient. In interpreting the variance inflation factor, generally 1 represents not correlated, 1–5 moderately correlated, and greater than 5 represents highly correlated.

To examine the influence of multicollinearity among the variables, VIFs were calculated and found to have a maximum value of 3.3 (

Table 10. Variance inflation factor (VIF) and tolerance in multicollinearity.

Coefficients
Model		Unstandardized Coefficients		Standardized Coefficients	t	Sig.	Collinearity Statistics
		B	Std. Error	Beta			Tolerance	VIF
1	(Constant)	2.823	0.162		17.378	0.000
	SNI	0.084	0.031	0.096	2.701	0.007	0.963	1.038
	SOR	0.140	0.038	0.132	3.707	0.000	0.963	1.038
a. Dependent variable: PBC

Source: VIF and collinearity statistics.

), which is at a permissible limit as recommended by Hair et al. [48].

4.7. Results of Structural Equation Model

A multivariate analysis technique (covariance-based structural equation modeling) was adopted to spot the noteworthy relationships among the three constructs—SNI, SOR, and PBC. From the direct effect model, we observed that a significant positive connection exists between SNI and PBC, because the beta coefficient was 0.09 and the p-value was less than 0.05, supporting H₃. In addition, from the direct effect model, we observed a highly significant positive association between SNI and SOR, since the beta coefficient was 0.123 and the p-value was less than 0.01, supporting H₁. Furthermore, a highly significant positive relationship was evident between SOR and PBC, as the beta coefficient was 0.09 and the p-value was lower than 0.01, supporting H₂. The regression weights for each relationship are shown in

Table 11. Regression weights (Group 1, the default model).

Hypothesis	Relationship	Estimate	S.E.	C.R.	p	Comment
H1	F2←F1	0.21	0.029	7.24	***	Significant
H2	F3←F2	0.123	0.039	3.149	0.002	Significant
H3	F3←F1	0.09	0.030	3.00	0.025	Significant

CR: Combined regression; F1:SNI; F2:SOR; F3:PBC. Significance of correlations: *** p < 0.001. Source: Regression weights.

and Figure 3 and Figure 4.

4.8. Results of Mediating Effects

This study has one mediator variable, organizational reputation. The result of the mediating effects is highlighted in

Table 12. User-defined estimands (Group 1, the default model).

Parameter	Estimate	Lower	Upper	p	Comment
SNI→OR→PBC	0.020	0.009	0.035	0.004	Partially Mediated

F1:SNI; F2:SOR; F3:PBC. Source: User-defined estimands.

.

When we considered the mediating factor, we observed that it can partially mediate SNI and PBC, as all path coefficients were significant at a 1% level. Therefore, H₄ is supported.

4.9. Moderating Effects of Gender

We examined whether there is any gender influence on the variables and on the explained relationships by selecting gender as the moderator, and thus developed three hypotheses accordingly. The aim of this effort was to identify any existing influence of gender on the direct and mediating relationships among the three variables. The moderating effects derived from multi-group analysis are shown in

Table 13. Significant differences between men and women for assuming unconstrained model.

Relationships	Model	Hypothesis	DF	CMIN	p	Decision
Assuming unconstrained model to be correct (for SNI and SOR)	Measurement weights	H5A	1	0.018	0.893	Not supported
Assuming unconstrained model to be correct (for PBC and SOR)	Measurement weights	H2B	1	0.446	0.504	Not supported
Assuming unconstrained model to be correct (for SNI and PBC)	Measurement weights	H3C	1	0.002	0.969	Not supported

DF: degree of freedom; CMIN: cumulative measurement index number. Source: Chi-square test.

.

A chi-square difference test was conducted with the unconstrained and constrained (individual) path models, and no significant difference was found between SNI and SOR, as the p-value was higher than 0.05with one degree of freedom. Thus, H_5A is not supported. Here, beta = 0.139 for men and 0.213 for women, indicating that a positive relationship between SNI and SOR was not stronger for women. For the path PBC and SOR, no significant difference was found, as the p-value was higher than 0.05 with one degree of freedom. Therefore, H_5B is not supported. Here, beta= 0.144 for men and 0.095 for women, indicating that a positive relationship between SNI and SOR was not stronger for women. Further, regarding the path of PBC and SNI, no significant difference was found, since the p-value was higher than 0.05 with one degree of freedom, thereby rejecting H_5C. In this case, beta = 0.065 for men and 0.069 for women, which means a positive relationship between SNI and PBC was not stronger for women. In conclusion, the three latent variables (SNI, SOR, and PBC) cannot be moderated by multi-group moderating variables such as gender.

5. Interpretation of Results

As we can observe, PAI has a noteworthy positive correlations with PBC accepting of the first hypothesis. This result is consistent with the previous findings since numerous authors have reported that hiring managers seek more information (than it is available in resume/CV) available from SNSs [21,30,31,32,33]. The study results indicate that the respondents were already using or were interested in using social networking information for pre-employment background checks at different organizations in Bangladesh. Further, sustainable organizational reputation can partially mediate the relationship between SNI and PBCs. Such results indicate that the more reputed the organization, the more it will use SNI. We note that gender had no influence on such practices. The results of this empirical investigation are consistent with previous findings and support previous arguments, although such investigations, particularly on this specific issue, are lacking in academia.

Notably, using social networking information for background checks has started to gain attention as a new practice and has not yet matured as it has in organizations in Western countries. It will be interesting to observe the future direction of this interesting phenomenon in this highly populated country.

6. Rationale and Significance of the Study

The practice of using social networking information for background checks of job applicants is a fairly new phenomenon, particularly in South Asian countries such as India, Bangladesh, Pakistan, and Sri Lanka. Although there are some practices evident in using SNI for recruitment and selection in some Western countries, such case is very rare in South Asia. Although the practice is becoming more popular and growing around the world due to easy availability and cost effectiveness, there are very few published studies regarding the use of SNI specifically for background checks. Particularly in Bangladesh, where more than 50% people are young and have at least one social media account, such studies are expected to be carried out often due to the rapid expected future utilization of SNI. Although this investigation has been conducted in a single country, the results can be helpful worldwide due to the utilization of social networking ratio (4 out of 5).

This empirical research is expected to fill the gap and further widen the possibility of using such big data for unique purposes, e.g., as pre-employment background checks. Further, it is expected to contribute to this rare field of theoretical knowledge and provide information for organizations to formulate and execute proper codes of conduct for employers and job candidates based on the practical findings. The different stakeholders involved in this paper (such as HR professionals and job candidates) may also consider the results of this study as the baseline for proper utilization. For example, HR professionals and other policymakers can use the results to formulate a transparent and coherent social media policy (SMP), while job candidates can be aware of their rights and responsibilities as future employees.

7. Limitations and Further Scope of Study

This study has several limitations that need to be pointed out. First, it was conducted in a particular culture where the unemployment rate is very high; a cross-cultural study might provide different and more representative results. Second, the practice of using SNI has just begun in Bangladesh. Therefore, organizations are not mature enough to handle such big data in an efficient manner. Further investigations should consider such limitations and try to reduce the gap in theory and practice, especially in conducting empirical studies employing different variables and techniques.

8. Recommendations and Concluding Comments

The authors would like to make some recommendations in the light of results. First of all, before using the information for any purpose at all, the employers should take the written permission from the owners in order to avoid any future complications. Secondly, only the job related information should be investigated and utilized for conducting PBC for any candidate. Finally and most importantly, a clear and transparent social media policy should be formulated and implemented for those organizations wishing to utilize SNI for PBC. The hiring professionals must be trained and the candidates must be informed regarding the possible utilization of such information for hiring purposes.

To date, although the relationship between social networking information and the process of pre-employment background checks has not been established empirically, the use of such information is becoming popular among hiring professionals due to its easy availability and cost friendliness. However, the practice has given rise to evolving and severe controversy among academicians, candidates, and practitioners. Particularly, the issue of mining social networking information by employers has led to legal and political controversy in some countries due to privacy and ethical concerns. However, it is apparent that such practice by employers continues to increase. Therefore, organizations are strongly recommended to protect themselves by complying with existing national and local laws regarding this issue. In addition, organizations are strongly urged to formulate and execute a precise social media policy (SMP) and communicate with job applicants from the beginning of the recruitment and selection process in order to avoid any unexpected outcomes.

Despite all of the criticisms regarding the use of social media information by HRM, particularly for selection and background checks, there is a growing possibility of using such a big data in the future. The proper ethical and effective use of such information will largely depend on controlled and well-guided policies and practices prepared by organizations and governments together. Organizations can work toward this purpose with the Ministry of Information of Bangladesh, which is basically responsible for guiding social network policies.