Abstract
Given organizations’ increasing reliance on career websites for recruiting job seekers, the experience a job seeker has on the organization’s website can influence the stream of job applicants and influence organizational success. This research, grounded in social cognitive theory, investigates the linkages between job seekers’ intentions, formed after using an organization’s recruitment website, and factors antecedent to these intentions. Data were collected by questionnaire from a sample of subjects who were representative of recent or current job seekers. The estimation of the proposed model was performed using a structural equations approach. The empirical results indicate that job seekers’ intentions are affected by their satisfaction with the recruitment website. Satisfaction is affected by job seekers’ perceived usefulness of the website; and perceived usefulness is positively affected by website ease of use, engagement, and social norms. Because the trend in recruiting is toward more impersonal contact (e.g., website) with job seekers prior to their decision to apply, it is essential that organizations design websites that increase job seekers’ intentions to apply and not to repel them. The study adds evidence that characteristics of the website communicate additional information to job seekers about the company that is used to form intentions.
Introduction
To develop and sustain a competitive advantage firms need resources that are valuable, rare, and hard to imitate [1]. One such resource is human capital [2]. The task of acquiring human capital is challenging because it requires firms to attract and hold the attention of high quality job seekers through a series of decisions by the firm and by the job seeker. The common image of the recruiting process is many applicants and a single open-position heavily weighted to the advantage of the firm. However, the success of the process hinges, at least initially, on a job seeker’s decision to apply or not to apply. Therefore, it is essential that firms understand how to influence job seeker’s application-decisions during the recruitment and selection process.
The recruitment and selection process can be conceptualized as a series of stages representing the development of a relationship between the job seeker and the organization [3, 4] – i.e., the generation stage, the maintenance stage and the job-choice stage. During the generation stage there is often a high level of uncertainty between the parties [5] and a low level of personal contact. If a job seeker chooses to withdraw from the relationship during this stage and prior to submitting an application then there is no opportunity for the organization to use more personalized recruitment activities that typically occur in later stages to influence the relationship [6–8].
Job seekers’ decisions during the three stages are largely affected by four types of information [9–13]: Information about the company (e.g., image, culture), the jobs (e.g., opportunities for growth), the recruiters’ behaviors (e.g., warmth, level of knowledge), and the selection process (e.g., perceived fairness, timeliness). A company’s website is often the initial mechanism through which job seekers find company and job information. If, as a result of this information, a qualified job seeker chooses not to apply then the organization misses an opportunity [6, 14], and may be left with a smaller and lower quality applicant pool as a result [15, 16]. Thus, companies must have a clear understanding of the factors that affect a job seekers’ level of satisfaction with the website and how this satisfaction affects job-pursuit intentions.
As stated above, the recruitment and selection process can be conceptualized in stages over which a relationship develops between the job seeker and the organization. However, the research presented here does not model how this relationship develops overtime through these stages. Rather, we study the end result of this relationship development process as opposed to the dynamic process itself. In this context, the purpose of the presented research is to explain the role of recruitment websites and applicant characteristics on an applicant’s job-pursuit intentions. These specific website characteristics are applicants’ perceptions of ease of use, engagement, usefulness, and satisfaction. The potential applicant characteristics include social norms, computer self-efficacy, and experience applying for jobs online. Furthermore, this research focuses on humans’ interactions with technology and their perceptions of the technology rather than the technology itself. The presentation of the research begins with the theoretical model and corresponding hypotheses followed by the empirical method used. The discussion of the empirical method includes presentations of the sample, measures, and the estimation of the model. A discussion of the empirical results comes next followed by conclusions from the research.
The model and hypotheses
The theoretical model and corresponding hypotheses are presented and discussed below. The model is also shown in Fig. 1 along with each hypothesis.
Job-pursuit intentions
There are several important recruitment-oriented outcomes; and they relate closely to the stage of the job seeker-employer relationship. Early in the relationship the focus of interest is on a job seeker’s job-pursuit intentions, which include such things as the job seeker’s intentions to submit an application or intentions to stay in an applicant pool [17]. Later in the relationship the focus of interest shifts to a job seeker’s acceptance intentions (e.g., “How likely are you to accept an offer?”) and job-choice decisions (i.e., to accept or decline an offer). Given the company website is often used by a job seeker early in the process as a source of information about the company and possible jobs, the most immediate outcome of a website search would be the job seeker’s intention to pursue a job with the company.
Aiman-Smith, Bauer & Cable (2005) differentiate job-pursuit intentions from organizational attractiveness, and argue that each are affected by different factors. They conceptualize organizational attractiveness as a “general positive affect toward an organization” (p. 221), contrasting this with job-pursuit intentions; describing it as an “intention to take action” (p. 221). Previous research has shown that job-pursuit intentions can be predicted by several factors, including job characteristics, organizational attributes, and organizational attractiveness (e.g. [19–22]). A company’s website is a mechanism through which job seekers can gather information about job characteristics and organizational attributes upon which to base their job-pursuit decisions[23, 24]. The website provides cues about organization characteristics that are unobservable by an outsider to help the outsider judge the extent to which he or she fits with the organization’s culture or a specific position [23, 25].
Evidence of a website’s value in affecting job seekers’ job-pursuit intentions can also be found in studies that have applied the elaboration likelihood model (ELM) to job seeker attraction settings (e.g., [26]). The model suggests that the level at which an individual processes information (i.e., deep or surface level) depends upon their ability and motivation to evaluate the information carefully [27]. When motivation and ability are high an individual is more likely to look deeply at the quality of the argument in a message. However, when motivation and ability are low individuals are more likely to rely on heuristics, the opinions of others, or simple cues. During the initial stage of recruitment a job seeker is more likely to be searching a number of potential employment opportunities and therefore be less committed to any single company; and because access to information about a company may be limited to such things as company websites, review sites (e.g., Glassdoor), and job advertisements, an individual’s ability to search deeply is also limited. In these instances, job seekers may rely more heavily on simple cues “such as a user-friendly website” [13, p. 601]. Thus, reactions to a recruitment website play a significant role on the extent to which a job seeker becomes attracted to an employer [28]. Job applicant’s percep-tions of website satisfaction have significant, positive influences on the applicants’ job-pursuit intentions to the company.
Perceived usefulness
Through the logic of social cognitive theory [29], the perceived usefulness of a company’s website can be seen as a factor that helps explain a job seeker’s willingness to continue the relationship with a potential employer because of satisfaction with its website. The theory has been used in the past to explain a number of continuance outcomes in technology-oriented settings, such as the use of a website [30–34]. The theory helps link an individual’s willingness to continue using a technology to two cognitive states [35], one behavioral and one affective. Outcome expectancy or perceived usefulness, the behavioral state, is the belief that a desired outcome will be achieved by performing a task. Job applicant’s perceptions of website perceived usefulness have significant, positive influences on website satisfaction.
Self-efficacy and personal mastery
Self-efficacy, the affective state, is the belief that an individual possesses the skills and abilities to successfully accomplish a specific task. Previous technology-oriented research built upon social cognitive theory has identified a user’s level of self-efficacy and their perceptions of personal mastery as primary drivers of a technology’s perceived usefulness [36]. When applied to a job seeker using a company website to gather information about the company and possible jobs, social cognitive theory helps determine a job seeker’s self-efficacy with accomplishing the task of searching for a potential employer. Job seekers develop perceptions about the usefulness of the website and his or her ability to accomplish the task. Self-efficacy often has a direct effect on perceived usefulness.
The final category of antecedents included in the study comes from personal mastery in the form of computer experience. The underlying concept for this antecedent is a job seekers’ prior level of computer experience influencing the use and perceptions of a website. The related hypotheses are listed below. Job applicant’s computer self-efficacy has significant, positive influences on website perceived usefulness. Job applicant’s experience applying online have sig-nificant, positive influences on website percei-ved usefulness.
Social norms
Social attitudes, derived from social cognitive theory, are particularly important in mandatory technology settings [37]. In the case of job seekers, the majority of companies require submission of an application through the recruiting website [38]. A job seeker perceives that others around them – friends, classmates, professors – have a role of importance and encourage the behavior of the job seeker to perform the desired task [39]. This perception becomes a social norm and is hypothesized below. Job applicant’s social norms about the web havesignificant, positive influe-nces on website perceived usefulness. Job applicant’s perceptions of social norms regarding important individuals have significant, positive influences on website perceived usefulness.
Website characteristics
The following antecedent relates to a company’s website ease of use and user engagement with the site. Ease of use and engagement positively influences a user’s perceived usefulness [23, 40]. The relationships between these antecedents and perceived usefulness are expressed below as hypotheses as well as noted on the model shown in Fig. 1.
Hypothesis Seven (H7): Job applicant’s perceptions of the recruitment website’s ease of use have significant, positive influences on website perceived usefulness.
Hypothesis Eight (H8): Job applicant’s perceptions of engagement with the recruitment website’s have significant, positive influences on website perceived usefulness.
The method
The sample
The data were collected by distributing an online questionnaire to potential respondents. Two distribution methods were used. The first method distributed questionnaires to undergraduate students enrolled in required business courses at a medium-sized university in the western United States who were preparing to graduate and therefore active job seekers. The second method was to contract with Mechanical Turk on Amazon.com to have their contract workers complete the questionnaire. The total sample size was 199 respondents with 99 questionnaires completed by students and 100 questionnaires completed by contract workers. The percentage of each respondent type in the sample is shown in Table 1.
There were several steps in the questionnaire administration process. First, respondents were asked to complete a subset of the questionnaire items. Second, each respondent was given the name of a company and the name of an open position at that company. There were a total of four such companies from which a specific assignment was made to each respondent. Respondents were instructed to go to the company website, find the position listing and initiate the job application process. They were then directed to search on the company website to identify another potential position of interest for the future. After completing these activities on the assigned website, respondents completed the remaining items on the questionnaire. While the names of the companies whose websites were used are not revealed for privacy reasons, the percentage of responses for each company included in the sample are shown in Table 1.
The sample demographic of the declared major for the students in the sample is also reported in Table 1. These percentages of majors ranged from 1.51% to 14.07%. An interesting reported value was that the largest response category was non-students and missing responses at 27.64%. From this result it appears that a number of the contract workers at Mechanical Turk are business students who reported their majors on the questionnaire.
The measures
The measures to empirically test the theoretical model were formed by responses to appropriate questionnaire items. These items or indicants of the measures are shown in Table 2. For all items, the respondents were given a statement and a scale upon which to respond. The response scale used was a five-point scale with anchors of 1-Strongly Disagree; 2- Disagree; 3-Neither Agree or Disagree; 4-Agree; and 5-Strongly Agree. All the items were previously published and modified to fit the current research. Also displayed in Table 2 are the standardized path coefficients of the indicants and psychometric properties of the measures generated from a confirmatory factor analysis using procedure Calis in PC SAS version 9.2.
All the measures in the confirmatory factor analysis were reflective in their indicants and allowed to pairwise correlate. The statistics summarizing the quality of the fit between the data and the estimated model were acceptable. The goodness of fit measure was 0.87 and adjusted for degrees of freedom it was 0.82. The parsimonious goodness of fit was 0.69. The root mean square residual and its standardized version were 0.042. The chi-square statistic was 386.78 with 239 degrees of freedom. It was statistically significant at a 1% level. The normed chi-square statistic was 1.62. The root mean square error of approximation was estimated to be 0.057 with a 90% confidence interval of 0.046 to 0.067. Bentler’s comparative fit index was 0.96. The incremental fit indexes(i.e., Bentler & Bonett’s non-normed and normed indexes; Bollen’s normed and non-normed indexes) ranged from 0.87 to 0.96.
Ease of use was measured using three items with standardized path coefficients of 0.78, 0.87, and 0.85. Its composite reliability was 0.87 and its shared variance extracted was 70%. Engagement was measured by two items with standardized path coefficients of 0.91 and 0.82. The composite reliability coefficient calculated from these path coefficients was 0.86 and the percentage of shared variance extracted was 75%. The social norms regarding important individual’s measure used three items with standardized path coefficients of 0.84, 0.95, and 0.92. Its calculated composite reliability coefficient was 0.93 and the percentage of shared variance extracted was 82%. Two indicants were used to measure computer self-efficacy. Its path coefficients were 0.84 and 0.85 with a reliability coefficient of 0.83 and shared variance extracted of 71%. The experience applying online measure consisted of two questionnaire items. The corresponding standardized path coefficients were 0.79 and 0.67. The resulting composite reliability measure was 0.70 with a shared variance of 54%. Five questionnaire items formed the job-pursuit intentions to the company measure. The standardized path coefficients for these items were 0.90, 0.89, 0.89, 0.90, and 0.88. The calculated composite reliability measure was 0.95 and its shared variance extracted was 80%. Two questionnaire items formed the social norms about the web measure. The standardized path coefficients were 0.79 and 0.83 with composite reliability of 0.79 and shared variance of 66%. Satisfaction with the website and website perceived usefulness were each measured using three questionnaire items. The satisfaction measure had standardized path coefficients of 0.82, 0.79, and 0.88 with a composite reliability coefficient of 0.87 and shared variance extracted of 69%. Website perceived usefulness had estimated path coefficients of 0.84, 0.84, and 0.72 and resulting reliability and shared variance extracted estimates of 0.84 and 64%, respectively.
Based on the magnitudes of the standardized path coefficients that ranged from 0.67 to 0.95, it can be argued that item reliability was satisfied [41]. In terms of composite reliability, all the measures demonstrated acceptable values based on the calculated reliability coefficients of 0.70 to 0.95 [41]. All these reliability measures equal or exceed the generally accepted cutoff level of 0.70 [42]. Additionally, all the shared variance extracted percentages were above 50%. The combination of these results indicates that the measures satisfy convergent validity [41, 43].
Discriminant validity was also examined by comparing, for each pair of measures, the squared correlation to the individual measures’ percentages of shared variance extracted. If discriminant validity is satisfied, the items within a measure share greater common variation among themselves than between the two measures. This is demonstrated when for each measure pair, the individual measures’ percentage of extracted shared variances are greater than the squared correlation between the two measures [44]. The squared correlation between two measures is in the spirit of a coefficient of determination. It shows the percentage of the total variation between these two measures captured or explained by the correlation between the two measures. All the squared correlations were calculated from the confirmatory factor analysis results and are reported in Table 3. The percentages of shared variance extracted were also calculated from the confirmatory factor analysis and are shown in Table 2. From these values, it is seen that discriminant validity was satisfied for all the measure. As a result, for all the measures, it can be concluded that construct validity issatisfied [45].
Estimation of the model
The estimation of the model and the corresponding tests of the hypotheses were performed using a structural equations approach. Specifically, the estimation technique was procedure Calis in PC SAS version 9.2 and maximum likelihood estimation. The statistics summarizing the quality of the fit between the model and the data indicate an acceptable fit [45, 46]. The goodness of fit index was 0.85, the adjusted goodness of fit was 0.80, and the parsimonious goodness of fit index was 0.71. The chi-square statistic was 464.89 with 253 degrees of freedom and was statistically significant at a 1% level. The resulting normed chi-square statistic was 1.84. The root mean square residual and its standardized counterpart were both 0.085. The root mean square error of approximation was estimated to be 0.066 with 90% confidence interval of 0.057 to 0.076. Bentler’s comparative fit index was 0.94 while the incremental fit indexes (i.e., Bentler & Bonett’s as well as Bollen’s normed and non-normed indexes) ranged from 0.085 to 0.94. All these values are displayed inTable 4.
The details of the model’s estimation are shown in Fig. 2 using standardized path coefficients. In the measurement model, all the paths which were free to vary were statistically significant at a 1% level. Five of the eight structural paths among the measures were statistically significant at a 1% level. All the significant structural paths had the hypothesized signs. The statistically insignificant paths were all to website perceived usefulness. Specifically, these were from computer self-efficacy, experience applying online, and social norms about the web. The corresponding significant paths to website perceived usefulness were from ease of use, engagement, and social norms regarding important individuals. Furthermore, the paths from website perceived usefulness to website satisfaction and then website satisfaction to job-pursuit intentions to the company were statistically significant at a 1% level.
Discussion
Crucial to organizational success is often a consistent source of qualified job applicants, some of which can hopefully be recruited to join the organization [47]. The increasing use of websites by organizations to recruit potential employees [48], indicates the importance of the recruitment website. The experience of job applicants on the organization’s recruitment website can affect, in a meaningful way, the stream of current and future job applicants [5, 48]. The job applicant’s experience with the recruitment website can be represented as a job-pursuit intention to the company. This intention is used to investigate factors influencing applicant recruitment and hence organizational success.
This research drew upon social cognitive theory to investigate the linkages between applicants’ job-pursuit intentions toward a company and factors antecedent to these intentions. Notice that the research interest is on the human interaction and perceptions in using recruitment websites and not in technology for technology sake. The empirical results indicate that job applicants’ satisfaction with the website has a meaningful impact on job-pursuit intentions toward the company. This may well imply that good, applicant-centered recruitment website design can affect applicants’ intentions to the company and therefore, the stream of current and future job applicants.
The next question is what influences website satisfaction? Based on our findings, the job applicants’ perceptions of recruitment website usefulness positively affects website satisfaction and ultimately job-pursuit intentions to the company. The implication is that if the recruitment website is perceived as useful in applying for a job, the applicant is more satisfied with the website and has higher intentions toward the company. The usefulness of the recruitment website can be affected by job applicant-centered website design to assure that the information provided and required of the job applicant is relevant and understandable.
The final question is which factors influence job applicants’ perceptions of website usefulness? Our empirical results indicate that if the recruitment website is seen by job seekers as easy to use and engaging, perceived website usefulness is enhanced. The practical implications of these results is that a recruitment website that has clear interactions and information that is easy to locate as well as being absorbing and holding applicants’ attention positively affects perceived usefulness. These are once again system design issues in terms of understanding the wants and needs of the company’s job seekers and developing the recruitment website with these in mind. The final factor influencing website perceived usefulness was found to be social norms regarding important individuals. This may well be a variable developed over a longer time period and from a perspective broader than website recruitment. These perceptions develop slowly overtime and may depend on a variety of sources, including public relations and communications.
Three other factors used in this research were found not to be statistically significant. Computer self-efficacy and experience applying online showed no effect on applicants’ perceptions of website usefulness and ultimately job-pursuit intentions through website satisfaction. Potential explanations for these insignificant results could be that completing forms and applications using a website are now common place for many of these job seekers. For other applicants, these could be uncommon activities. Across these two groups, if sufficiently large, the effect of computer self-efficacy and prior experience could be offset. This essentially implies that the attitudes regarding these two factors are cancelled out across these two groups. As a result, respondents’ computer self-efficacy and prior experience applying online levels, at least as measured in this study, are statistically insignificant in influencing website perceived usefulness. In order to investigate this possibility, the indicants forming each measure for each respondent were summed and the sums averaged and the standard deviation calculated for each measure. The standard deviations for computer self-efficacy and prior experience applying online had the two largest standard deviations. In fact, these standard deviations were approximately 90% larger than the measure with the smallest standard deviation, possibly explaining these insignificant results.
The final insignificant result was between the social norms about the web variable and website perceived usefulness. These social norms relate to effects of others encouraging the use of the web to search for and apply for jobs. The underlying assumption is that job seekers’ use of a website to find and apply for jobs is volitional; most likely that is not the case. Most companies require job-applicants to submit resumes’ via the company website making the application process non-volitional [23]. As a result, these social norm values have little to no impact on website perceived usefulness and ultimately job-pursuit intentions through website satisfaction.
As discussed earlier, this research focused on the end result of the relationship which develops between the job seeker and the organization facilitated by use of the recruitment website. Interesting research for the future could be to examine the dynamic process by which this relationship develops over time. Such research would require a dynamic analysis, recursive model, and longitudinal data.
Conclusions
Job seekers look for information about the company, the jobs, and the selection process from companies’ websites during the recruiting process. From the company’s perspective, finding a qualified job applicant can enhance the efficiency of the recruitment and selection process and ultimately enhance company performance by finding the right person who has a desire for the job. If a job applicant has a negative experience in the recruiting process – difficultly finding the jobs, lack of response from recruiters, hard to use website – the company’s efficiency will decline when the job applicant pool doesn’t have qualified applicants. During recruitment, and ultimately the selection process, it is important for the company to understand how to affect job seeker’s application-decisions. The use of websites for job seekers and job recruitment will continue to increase [49] and companies need to consider the job seeker satisfaction, perceptions of usefulness, ease of use, and engagement with the site.