Next Article in Journal
EC-MFR: A Hierarchical Edge–Cloud Collaborative Framework for Multimodal Fact-Checking
Previous Article in Journal
Application of Blockchain Technologies and Smart Contracts for the Storage and Verification of Academic Transcripts in the Higher Education Systems
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Information
Volume 17
Issue 5
10.3390/info17050479
information-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

User Acceptance of a Proposed Static-Dynamic Employment Recommendation Approach Among Computer Science Graduating Students

by
Huafeng Qu
1,2,*,
Shafrida Sahrani
3,*,
Fariza Fauzi
1,
Xiacheng Song
3 and
Yanfeng Zhao
4
1
Faculty of Information Science and Technology, Universiti Kebangsaan Malaysia, Bangi 43600, Selangor, Malaysia
2
School of Information and Intelligent Engineering, Yunnan College of Business Management, Kunming 650300, China
3
Institute of Visual Informatics, Universiti Kebangsaan Malaysia, Bangi 43600, Selangor, Malaysia
4
School of Information Engineering, Xi’an Fanyi University, Xi’an 710105, China
*
Authors to whom correspondence should be addressed.
Information 2026, 17(5), 479; https://doi.org/10.3390/info17050479
Submission received: 2 April 2026 / Revised: 28 April 2026 / Accepted: 12 May 2026 / Published: 13 May 2026
(This article belongs to the Section Information Systems)
Browse Figures
Versions Notes

Abstract

Employment recommendation services are increasingly used to support graduate job search. However, limited research has examined how graduating computer science students perceive a proposed employment recommendation approach that combines static profile-based matching with dynamic interactive functions. Drawing primarily on the Technology Acceptance Model (TAM), with selected dimensions of the Information System (IS) Success Model used as supplement, this study conducted an exploratory questionnaire-based survey of 386 graduating students. The respondents evaluated existing employment recommendation systems and provided open-ended comments, and the findings show that only 38.3% of respondents reported willingness to use existing employment recommendation systems for job hunting. The main reported problems were delayed matching to individual qualifications (71.0%), information lag (55.4%), and jobs not matching students’ majors (54.1%). In contrast, respondents expressed relatively favorable attitudes toward the proposed static-dynamic approach: 67.6% indicated willingness to use it and 59.6% indicated willingness to recommend it to others. Exploratory subgroup analyses further suggested that positive evaluations of the proposed approach were higher among students from emerging computing fields and those with more active job-seeking engagement (p < 0.05). Overall, the findings provide exploratory evidence that graduating computer science students may respond more positively to employment recommendation concepts that integrate profile-based matching with dynamic interaction. However, it is a proposed design concept, not an implemented system, evaluated by the respondents. Therefore, the results should be interpreted as perceptions and stated intentions, instead of evidence of actual adoption or real-world system effectiveness.

Graphical Abstract

1. Introduction

In recent years, the rapid expansion of higher education and the increasing complexity of labor markets have intensified competition in graduate employment worldwide [1,2]. Against this global backdrop, China provides a representative large-scale case, where higher education has rapidly transitioned from an elite system to mass and universal participation, resulting in a remarkable expansion in scale over the past two decades [3]. According to statistics released by the Ministry of Education, the total number of graduates from regular colleges and universities across the country has continued to climb from about 1 million at the beginning of this century [4] to nearly 12.22 million in 2025 [5], and is estimated to reach 12.70 million in 2026 [6]. This historic transformation has not only provided a large-scale supply of high-quality human resources for national development but has also put unprecedented supply pressure on the labor market [7]. In the context of the evolution of the global economic pattern, the upgrading of domestic industrial structure and the new technological revolution, the demand-side structure of the labor market is also undergoing profound and complex adjustments [8,9]. The transformation of traditional occupations and the emergence of new job profiles are occurring simultaneously, while labor market demand is becoming increasingly dynamic in terms of skill combinations, task complexity, and adaptability [10]. The continuous high level of supply and the rapid evolution of demand structure have jointly led to the prominence of structural contradictions in the employment market. The phenomena of “difficulty in finding employment” and “difficulty in recruiting” coexist, and the employment competition faced by graduates is becoming increasingly fierce and complex [11]. Against this macro background, how to break through the limitations of the traditional information matching model and achieve efficient, accurate and dynamic person-job fit with the help of technology, has become not only a technical optimization problem, but also a core issue concerning the career development path of millions of graduates, the efficiency of human resource allocation, and even the long-term stable development of the social economy [12,13].
To address this challenge, employment recommendation systems [14] based on big data and intelligent algorithms have emerged and are gradually becoming an important tool to assist students in job hunting [15]. These systems analyze students’ professional background, skills, practical experience and behavioral preferences, while integrating corporate recruitment needs and job characteristics, to provide students with personalized job recommendations, thereby effectively broadening job search channels, reducing information search costs, and improving the targeting and success rate of job hunting [16]. Studies have shown that effective employment recommendation systems can not only optimize matching efficiency, but also alleviate the problem of information asymmetry in the job search process to a certain extent [17].
In the Chinese graduate employment context, students commonly encounter digital recruitment platforms and services that combine vacancy browsing with recommendation, filtering, matching, or platform-mediated communication functions. Examples include widely used recruitment platforms such as Boss Zhipin, Zhaopin, 51job, and Lagou, all of which provide representative examples of the broader recruitment platform ecology in China [18,19,20,21]. Although these platforms differ in scope and interface design, they broadly represent the types of “existing employment recommendation systems” referred to in this study. This clarification is important because the practical form of employment recommendation services may vary across regional labor markets, and international readers may not be familiar with the Chinese recruitment platform context [22].
However, existing employment recommendation systems rely excessively on static historical data (such as resume keywords) for one-way matching in practical applications, lacking the ability to integrate and respond to dynamic real-time information [23]. This has led to many problems, such as “low accuracy in matching majors with positions”, “delayed information updates” and “inability to adjust based on job seekers’ real-time feedback” [24]. For fields such as computer science, where technology is developing rapidly and emphasizes practical operation and instant communication, the static characteristics of existing systems are difficult to meet graduates’ needs for in-depth information, such as detailed technical stack requirements, team culture, and real-time Q&A. This may in turn be associated with lower perceived credibility and lower stated willingness to use such systems [25].
A proposed static-dynamic employment recommendation approach has been increasingly discussed as a promising direction for improving digital job-search support [26]. In the present study, “static” refers to relatively profile-based matching processes that generate initial recommendations from user background information, resumes, skills, and job databases. By contrast, “dynamic” refers to more interactive and adaptive support processes, such as real-time communication, iterative preference adjustment, feedback-oriented clarification, and other forms of bidirectional interaction during job exploration and evaluation [27]. Compared with traditional one-way recommendation logic that relies mainly on historical profile data and static matching rules, this proposed approach emphasizes a more responsive and interactive process. Conceptually, such integration may improve perceived relevance, communication support, and transparency, and may therefore be associated with more favorable evaluations of employment recommendation services [28].
Although the concept of “static-dynamic fusion” has potential in the field of recommendation systems [29], there is currently no research on the acceptance of such systems by specific user groups, especially computer science students, because the technology in this field is updated rapidly and the demand for actual communication is strong, which better reflects the value of dynamic interactive systems [30]. Existing studies have mostly focused on algorithm optimization or macro-level performance evaluation [16], with limited attention to how specific user groups evaluate such systems in terms of perceived usefulness and behavioral intention [27]. Therefore, this study aims to investigate graduating computer science students’ awareness of, and stated willingness to use, existing employment recommendation systems, and to examine their perceptions of and behavioral intentions toward a proposed static-dynamic job recommendation approach. Rather than evaluating an implemented system, this study focuses on how respondents assess the perceived usefulness and stated acceptance of a proposed design concept on the basis of questionnaire descriptions. Accordingly, the study does not examine actual system use, observed trust behavior, usability outcomes, or real-world recommendation effectiveness.
Accordingly, this study addresses the following research objectives:
RO1: To examine respondents’ perceived usefulness of a proposed employment recommendation approach that integrates static matching and dynamic interaction.
RO2: To examine respondents’ behavioral intentions toward a proposed employment recommendation approach that integrates static matching and dynamic interaction.
The remainder of this paper is structured as follows: Section 2 outlines the research methods, theoretical framework, and questionnaire designed based on that framework. Section 3 presents the research findings. Section 4 discusses the findings. Finally, Section 5 summarizes the paper.

2. Materials and Methods

2.1. Ethical Considerations

Ethical approval was obtained from the Institutional Review Board of Yunnan College of Business Management, China (protocol code YCBM-2025-012; approval date: 22 November 2025). All participants provided written informed consent prior to participation and were informed that participation was voluntary and that they could withdraw at any time without penalty. Survey responses were anonymized and analyzed in deidentified form, and data were stored securely with access restricted to the research team.

2.2. Participants and Data Collection

In this study, “graduating computer science students” refers to final-year undergraduate students in computer science-related disciplines who are about to graduate. This study was conducted at two higher education institutions, surveying a total of 442 graduating students majoring in computer-related fields, of which 386 valid responses were retained after excluding questionnaires with missing responses or multiple selections in single-choice items, resulting in an effective response rate of 87.3%. Equation (1) shows that the target sample size was determined with reference to Cochran’s formula for large populations [31]. The formula was applied using a 95% confidence level (Z = 1.96), maximum variability (p = 0.5), and a 5% margin of error (e = 0.05). Under these assumptions, the recommended minimum sample size is approximately 384. Accordingly, the final sample of 386 valid responses satisfied the minimum sampling requirement for this study.
n 0 = z 2 · p · 1 − p e 2 ,
Participants were all graduating undergraduate students, with the following majors: Computer Science and Technology (165, 42.7%), Artificial Intelligence (78, 20.2%), Internet of Things Engineering (29, 7.5%), Software Engineering (23, 6.0%), Virtual Reality Technology (22, 5.7%), Network Engineering (21, 5.4%), Data Science (18, 4.7%), and other majors (30, 7.8%). Of all participants, 365 (94.6%) planned to enter the workforce directly after graduation, while 21 (5.4%) had no immediate plans for employment. All participants possessed basic skills in using digital tools and online platforms. In this study, sociodemographic characteristics (gender, major) and job-seeking status were treated as background or grouping variables; the main survey outcomes included participants’ awareness of employment recommendation systems, willingness to use them, functional evaluations, and stated acceptance of, and behavioral intentions toward, new recommendation systems integrating static matching and dynamic interaction.

2.3. Theoretical Framework

The theoretical framing of this study is informed primarily by the Technology Acceptance Model (TAM) [32], with selected dimensions of the Information System (IS) Success Model [33,34] used as supplementary interpretive lenses. TAM was chosen because the present study focuses specifically on two acceptance-oriented dimensions that align closely with the questionnaire design and the exploratory purpose of the investigation: perceived usefulness (PU) and behavioral intention (BI). Although extended variants such as TAM2 may offer stronger explanatory power for predictive modeling of technology use, the present study was not designed to test a more complex causal model involving social influence, cognitive instrumental processes, or structural relationships among multiple latent constructs [35]. Instead, the aim was to provide an exploratory, questionnaire-based assessment of how graduating computer science students evaluate a proposed employment recommendation approach that integrates static matching with dynamic interaction. The IS Success Model was therefore used only as a supplementary interpretive framework when discussing respondents’ evaluations of information quality, responsiveness, service reliability, and related concerns such as trust and privacy [36]. Accordingly, this study should be understood as theory-informed rather than theory-testing.

2.4. Existing Employment Recommendation Systems and the Proposed Static-Dynamic Approach

In this study, the term “existing employment recommendation systems” was used in a broad applied sense to refer to digital job-search platforms or services that provide recommendation, matching, filtering, or platform-mediated recruitment support. In the Chinese context, this may include widely used recruitment platforms such as Boss Zhipin, Zhaopin, 51job, and Lagou. Their official websites are available at https://www.zhipin.com/, https://www.zhaopin.com/, https://www.51job.com/, and https://www.lagou.com/, respectively. The survey did not restrict respondents to one specific platform. The proposed approach was presented as a two-stage service concept: static matching first provided initial job recommendations based on student profiles and job information, whereas dynamic interaction then allowed users to clarify job requirements, adjust preferences, and receive feedback during job exploration. Compared with traditional one-way recommendation systems that rely mainly on historical profile data, resume keywords, and static matching rules, this approach emphasizes a more responsive and interactive recommendation process.

2.5. Questionnaire Design and Measurement

Questions related to perceived usefulness, such as “Do you think improving the accuracy of job matching in the employment recommendation system will help you find a suitable job? [37]” directly reflect students’ perception of whether the function helps improve job search efficiency, which is a judgment of the system’s “usefulness” [38]. Questions like “Do you think adding real-time chat interaction to the employment recommendation system will help you find a suitable job?” assess the perceived practicality of dynamic interactive functions in promoting information communication and improving matching effectiveness [39].
Behavioral intention was assessed using intention-oriented items such as “Would you use such an employment recommendation system?” and “Would you be willing to recommend such a system to your classmates or friends?”, which reflect adoption and recommendation intentions [40,41].
Questions related to system quality and information quality: For example, when investigating the limitations of existing systems [42], options such as “false information [43],” “information lag [44],” and “job mismatch with major [45]“ directly point to students’ evaluation of the quality (authenticity, timeliness, accuracy) of the information provided by the system [22,23]. “Unable to find a job in a timely manner based on one’s qualifications” involves the perception of system quality (algorithm matching ability [46], responsiveness [47]).
Questions related to service quality and perceived risk: In the open comments, students commonly mentioned suggestions such as “strengthening corporate vetting” [48], “protecting personal privacy” [49], and “avoiding fake job postings” [50], reflecting their concerns about service quality (reliability and security) and also involving the impact of perceived risk (information risk and privacy risk) on their acceptance [51].
The question “Do you think this combination of static and dynamic employment recommendation systems can help you find a job?” was designed to capture respondents’ overall perceived usefulness of this proposed integrated approach [38], while also reflecting their general evaluation of its potential service value [52].
In addition to the structured items, the questionnaire also included a final open-ended question inviting respondents to provide supplementary comments and suggestions regarding the proposed static-dynamic employment recommendation approach. Responses were first screened to remove non-substantive entries such as “none,” “no comment,” or equivalent brief responses. The remaining substantive comments were then reviewed repeatedly and grouped into broad recurring issue categories through an iterative descriptive comparison process [53]. The initial descriptive grouping was conducted by the first author and subsequently reviewed by a second researcher to improve consistency in category interpretation. Any differences in category interpretation were discussed until agreement was reached. This procedure was intended to support descriptive interpretation of recurring concerns and expectations rather than to generate a formal qualitative coding framework or theory-driven qualitative model.

2.6. Research Design and Data Analysis

This study employed a cross-sectional descriptive survey design to investigate respondents’ awareness of existing employment recommendation systems and their acceptance of a proposed static-dynamic job recommendation approach. Data were collected on-site using a structured questionnaire. The responses were used to descriptively examine perceived usefulness, behavioral intention, and selected user evaluations relevant to employment recommendation service design.
The questionnaire consisted of four parts: Part One covered respondents’ socio-demographic characteristics and basic understanding of employment recommendation systems; Part Two investigated their user experience and functional evaluation of existing employment recommendation systems; Part Three explored the main limitations of existing systems; and Part Four focused on the acceptance, willingness to use, and recommendation intentions of the combined static and dynamic job recommendation system, along with related suggestions.
The questionnaire data were collected on-site through paper-based questionnaires, yielding a total of 386 valid responses. After being processed and coded, data were statistically analyzed using SPSS 28 software [54]. In addition to descriptive statistics of demographic characteristics and basic cognition, the reliability and construct validity of the scales were systematically evaluated through reliability and validity analyses [55].
The questionnaire explicitly examined respondents’ awareness of, prior exposure to, and evaluations of existing employment recommendation systems. However, with regard to the newly proposed static-dynamic recommendation approach, respondents did not interact with a working prototype. Their judgments were therefore formed on the basis of the functional descriptions provided in the questionnaire, together with their prior understanding of and experience with existing job recommendation platforms. Accordingly, the findings should be interpreted as respondents’ perceptions of and stated intentions toward a described design concept, rather than as evidence of actual system use, observed trust behavior, usability outcomes, or real-world recommendation effectiveness.

2.7. Methodological Limitations

Despite yielding meaningful findings, this study has several limitations. First, the sample consisted entirely of computer science students at two higher education institutions located in the same geographical region, limiting its representativeness and geographical coverage. Accordingly, the findings should be interpreted with caution when generalized to broader student populations. Second, although the study participants were all computer science students, their understanding of the job market, job-seeking experience, and familiarity with recommender systems varied, which may have influenced their evaluation and acceptance of the system’s functionality.

3. Results

3.1. Sociodemographic Characteristics

This study included 386 graduating computer science students as a valid sample, of whom 222 were male (57.5%) and 164 were female (42.5%). In terms of major distribution, Computer Science and Technology had the largest number of participants (165, 42.7%), followed by Artificial Intelligence with 78 participants (20.2%). Other majors included Internet of Things Engineering (29, 7.5%), Software Engineering (23, 6.0%), Virtual Reality Technology (22, 5.7%), Network Engineering (21, 5.4%), Data Science (18, 4.7%). An additional 30 participants (7.8%) were from other related majors. Regarding job-seeking intentions, the vast majority of participants (365, 94.6%) planned to work after graduation, with only 21 (5.4%) indicating no immediate plans for employment. For job-seeking status, the largest group was actively searching but had not yet received an interview opportunity (198, 51.3%). This was followed by students in internships (76, 19.7%), those not currently looking for work (93, 24.1%), and those who had received a job offer (19, 4.9%) (Table 1).

3.2. Reliability and Validity Analysis

To examine the measurement quality of the survey instrument, internal consistency and sampling adequacy were assessed using SPSS 28. As shown in Table 2, the overall five-item scale yielded a Cronbach’s α of 0.818, indicating good internal consistency. Since the subsequent exploratory factor analysis suggested a two-factor structure, reliability was also examined separately for the two subdimensions. The perceived usefulness (PU) subscale showed excellent internal consistency (Cronbach’s α = 0.928), whereas the behavioral intention (BI) subscale showed acceptable internal consistency for exploratory research (Cronbach’s α = 0.630). Overall, these results provide preliminary support for the reliability of the instrument, although given the small number of items in the BI subscale, its internal consistency estimate is naturally more conservative [56,57].
Regarding the preliminary structural pattern of the questionnaire, the KMO value was 0.735 and Bartlett’s test of sphericity was significant (p < 0.001) (Table 3), indicating that the correlation matrix was suitable for principal component analysis.
Principal component analysis (PCA) with varimax rotation extracted two components with eigenvalues greater than 1, explaining 81.99% of the total variance [58]. The rotated component loadings shown in Table 4 indicated a clear two-component structure, with all loadings exceeding 0.50. Based on the loading pattern, Component 1 was interpreted as reflecting perceived usefulness (PU), whereas Component 2 was interpreted as behavioral intention (BI).
These results provide preliminary evidence for the two-dimensional structure of the five acceptance-related items, although the structure should be interpreted cautiously given the exploratory nature of the study and the limited number of items.

3.3. System Awareness and Willingness to Use

This study surveyed 386 graduating computer science students. Of these, 286 (74.1%) were aware of the job recommendation systems, but only 148 (38.3%) were willing to use the existing systems for job hunting. This gap suggests that awareness did not necessarily translate into stated willingness to use these platforms and may reflect possible barriers to acceptance. This pattern provides an initial basis for examining respondents’ subsequent evaluations of existing employment recommendation systems and their perceived limitations.

3.4. Evaluation of the Existing Employment Recommendation Systems

Regarding the extent to which existing employment recommendation systems support users in finding suitable jobs, the respondents’ evaluations are distributed as follows: 37 people (9.6%) expressed “strongly agree,” 56 people (14.5%) expressed “agree,” 106 people (27.5%) expressed “neutral,” 112 people (29.0%) expressed “disagree,” and 75 people (19.4%) expressed “strongly disagree” (Table 5). The total percentage of respondents holding a positive attitude (“strongly agree” and “agree”) was 24.1%, while the total percentage of respondents holding a negative attitude (“disagree” and “strongly disagree”) was 48.4%. This distribution suggests that many respondents held reservations about existing employment recommendation systems. From the respondents’ perspective, these platforms may not fully meet the job-seeking needs of graduating computer science students.

3.5. Limitations of Existing Employment Recommendation Systems

Table 6 summarizes responses to the multiple-choice question, “What limitations do you see in the current job recommendation systems?” The results indicate that perceived problems cluster in the following areas.

3.5.1. Severe Deficiencies in Matching Functionality

A high percentage of respondents (71.0%, n = 274) felt that existing employment recommendation systems “cannot match jobs promptly based on their qualifications.” This response pattern suggests that respondents perceived existing platforms as insufficient in personalization and responsiveness, especially in relation to their expectations for more dynamic and accurate matching.

3.5.2. Multiple Challenges to Information Quality

Over half of the respondents perceived problems with the information quality of existing employment recommendation platforms. Specifically, 55.4% (214 people) selected “information lag,” highlighting insufficient timeliness. For graduating computer science students, information lag is not merely an inconvenience but a critical limitation. Due to the fast pace of technological change in the IT and AI sectors, employment information quickly loses relevance as emerging technologies alter required skills and job specifications [59].

3.5.3. Significant Mismatch in Professional Suitability

“Job not matching major” was specifically pointed out by 54.1% of respondents, further confirming the inherent limitations of general recommendation algorithms in understanding and matching majors such as computer science, which require specialized knowledge and skills.

3.5.4. Other Issues

A small number of respondents (2.1%) provided supplementary opinions in the “Other” option. However, most of the responses simply stated “none,” while a few respondents indicated that they were “not sure” or had never used the system. These responses did not reveal additional limitations beyond the categories already identified above.

3.6. Respondents’ Acceptance of the Proposed Static-Dynamic Job Recommendation Approach

To examine respondents’ attitudes toward a proposed employment recommendation approach that combines static profile matching with dynamic real-time interaction, this study assessed responses at two levels: perceived usefulness and behavioral intention. Overall, respondents reported generally favorable evaluations of this proposed design concept.
Figure 1 shows that regarding functional value recognition (perceived usefulness), respondents showed high levels of acceptance for both core functions, with particular emphasis on the dynamic interaction function. Specifically, regarding “improving job matching accuracy,” a total of 59.6% of respondents (230 people) held a positive attitude (135 people “strongly agree,” accounting for 35%; 95 people “agree,” accounting for 24.6%), 18.4% (71 people) held a neutral attitude, while a total of 22% (85 people) held a negative attitude. The acceptance of the “real-time chat interaction” feature further increased, with a total of 65.3% of respondents (252 people) believing it would be helpful for job hunting (139 people “strongly agree,” accounting for 36%; 113 people “agree,” accounting for 29.3%). The proportions holding neutral and negative attitudes were 17.1% (66 people) and 17.6% (68 people), respectively. When asked whether the system as a whole would be helpful for job hunting, 63.2% of respondents (244 people) gave a positive evaluation (98 people “agree,” accounting for 25.4%; 146 people “strongly agree,” accounting for 37.8%), showing recognition of the overall value of this integrated model.
Figure 2 presents responses at the behavioral intention level, indicating that respondents showed a relatively strong willingness to use and recommend the proposed system. First, regarding personal usage intention, 67.6% of respondents (261 people) indicated a willingness to use such a system (84 people "agree," accounting for 21.8%; 177 people "strongly agree," accounting for 45.9%). Second, regarding the willingness to recommend to others, although the positive percentage was slightly lower, 59.6% of respondents (230 people) still expressed a willingness to recommend (106 people "agree," accounting for 27.5%; 124 people "strongly agree," accounting for 32.1%). At the same time, 21.3% of respondents (82 people) selected “disagree” or “strongly disagree,” suggesting that recommendation intention was somewhat lower than personal usage intention.

3.7. Qualitative Summary of Open-Ended Feedback and Design Suggestions

Open-ended responses to the final questionnaire item were used as supplementary descriptive material to contextualize the quantitative findings. After excluding brief non-substantive entries such as “none,” “no comment,” or equivalent responses, 84 substantive responses remained for descriptive review. These responses were examined iteratively and clustered into four broad issue areas: information governance and trust building, algorithm optimization and personalized services, privacy, security, and algorithm transparency, and functionality improvement and experience refinement. Because this component was intended to supplement the survey findings rather than to constitute a formal qualitative study, the resulting categories should be understood as descriptive groupings of recurring concerns rather than as fully developed qualitative themes.

3.7.1. Information Governance and Trust Building

A recurring concern in the open-ended comments related to information governance and trust building, with respondents mentioning terms such as “fake,” “verification,” and “authentic.” Some respondents questioned the reliability of information on existing platforms, citing concerns such as “most job search platforms are actually intermediary companies, falsely advertising recruitment information” and “too many fake jobs.” They explicitly demand that the platform fulfill stricter verification responsibilities, “strengthening official verification and conducting background checks on companies,” and establishing a corporate credit endorsement mechanism to “eliminate fake job postings” and “ensure the authenticity of recruiting companies.” This reflects that information quality is not only a functional issue but also a fundamental problem affecting system credibility and user engagement.

3.7.2. Algorithm Optimization and Personalized Services

Users generally pointed out that the current recommended “jobs do not match their majors” (54.1% of the quantitative results echo this), hoping that the matching logic can go beyond keyword comparison and achieve a deeper understanding. Suggestions include “matching based on individual professional strengths” and “testing job matching. Recommending jobs based on work experience.” A further demand is for the system to provide personalized diagnostics and development suggestions, such as “helping us match ourselves with job requirements and analyzing the reasons for any mismatches and gaps.” This indicates that users expect the system to transform from a passive information filter into a proactive career development consultant.

3.7.3. Privacy, Security, and Algorithm Transparency

With increased awareness of data security, this topic was explicitly mentioned. Users are concerned about “the protection of personal information privacy” and suggested “viewing the platform’s privacy policy.” Simultaneously, they expressed unease about the “black box” algorithm, implicitly demanding the right to know about the recommendation logic, which is directly related to the suggestion to “improve privacy protection and algorithm transparency to enhance user trust.” This suggests that perceived risk (privacy risk and uncertainty risk) may be an important concern shaping respondents’ stated willingness to use such systems.

3.7.4. Functionality Improvement and Experience Refinement

Users put forward many specific suggestions for functional enhancement, mainly including Enhancing information dimensions and interactivity: Suggestions include “adding multi-person public evaluations from companies” and building an integrated “online interview” process. Enhance tool functionality: Calls for the provision of “resume correction” or “intelligent optimization” services. Optimize interaction and filtering mechanisms: Demands simplification and fewer processes, addressing issues such as “duplicate recommendations” and the inability to “block uninteresting positions.”

3.8. Exploratory Chi-Square Subgroup Analysis

For subgroup analysis, job search status was recoded into an active job-seeking group (actively searching or in an internship) and a less active job-seeking group (not actively looking or having received a signed offer). Professional orientation was recoded into Traditional Computing (Computer Science and Technology, Software Engineering, Network Engineering, and Internet of Things Engineering) and Emerging Computing Fields (Artificial Intelligence, Data Science, and Virtual Reality Technology), with respondents in Other excluded. The five user-acceptance items described in Section 3.6 were dichotomized into Positive and Non-positive for Pearson’s chi-square analysis [60]. Because this recoding reduces the ordinal information in the original Likert-scale responses, the subgroup results should be interpreted as exploratory.
As shown in Table 7, the active job-seeking group reported significantly higher positive response rates than the less active job-seeking group across all five evaluation items: improving job matching accuracy helps job search (63.5% vs. 50.0%, χ2 = 6.020, p = 0.014); real-time chat interaction helps job search (69.3% vs. 55.4%, χ2 = 6.862, p = 0.009); Combination of static and dynamic employment recommendation helps job search (66.8% vs. 54.5%, χ2 = 5.193, p = 0.023); intention to use the system (71.2% vs. 58.9%, χ2 = 5.439, p = 0.020); and intention to recommend the system (63.1% vs. 50.9%, χ2 = 4.951, p = 0.026).
As shown in Table 8, respondents in Emerging Computing Fields also reported significantly higher positive response rates than those in Traditional Computing for all five items: improving job matching accuracy helps job search (67.8% vs. 56.7%, χ2 = 4.044, p = 0.044); real-time chat interaction helps job search (73.7% vs. 61.8%, χ2 = 5.013, p = 0.025); Combination of static and dynamic employment recommendation helps job search (71.2% vs. 59.7%, χ2 = 4.518, p = 0.034); intention to use the system (77.1% vs. 63.4%, χ2 = 6.775, p = 0.009); and intention to recommend the system (70.3% vs. 54.6%, χ2 = 8.108, p = 0.004). Overall, positive evaluations of the proposed system were consistently higher among respondents with more active job-search engagement and among those from Emerging Computing Fields.
To preserve the ordinal information of the original Likert-scale responses, Mann–Whitney U tests were further conducted on the composite PU and BI scores [52]. As shown in Table 9, significant subgroup differences were observed for both professional orientation and job-seeking status. Respondents in Emerging Computing Fields showed significantly higher mean ranks than those in Traditional Computing on both PU (U = 12,215.000, Z = −2.018, p = 0.044) and BI (U = 11,631.500, Z = −2.674, p = 0.007). Similarly, the active job-seeking group showed significantly higher mean ranks than the less active group on both PU (U = 13,170.500, Z = −2.206, p = 0.027) and BI (U = 12,404.500, Z = −2.996, p = 0.003). These results further support the subgroup differences identified in the chi-square analysis.

4. Discussion

4.1. A Differential Analysis Based on Job-Seeking Status and Professional Orientation

Based on respondents’ job-seeking status and professional orientation, this study further compared the acceptance of the proposed static-dynamic job recommendation approach across different groups [23,61].
Regarding job-seeking status, the results suggest subgroup differences in the evaluation and acceptance of the proposed system functions between the two job-seeking groups. Specifically, respondents in the active job-seeking group (including those currently in internships or actively seeking employment) showed positive response rates that were generally about 12 to 14 percentage points higher across the five evaluation items than those in the less active job-seeking group. For example, regarding intention to recommend the system, the positive response rate in the active job-seeking group reached 63.1%, compared with 50.9% in the less active group. Similarly, for intention to use the system, the corresponding figures were 71.2% and 58.9%, respectively. These findings suggest that respondents with stronger immediate job-search involvement reported higher acceptance of the proposed system functions and more favorable overall evaluations of the proposed approach.
Regarding professional orientation, the observed differences were mainly reflected in the perceived value of specific functions and in overall behavioral acceptance. Based on the chi-square subgroup analysis, respondents in Emerging Computing Fields showed consistently higher positive response rates than those in Traditional Computing across all five evaluation items. In particular, students in Emerging Computing Fields reported higher positive evaluations for improving job matching accuracy (67.8% vs. 56.7%), and real-time chat interaction (73.7% vs. 61.8%) both with statistically significant subgroup differences (p < 0.05). They also showed higher positive response rates for overall system helpfulness, intention to use, and intention to recommend. One possible explanation is that students in fast-changing fields, such as artificial intelligence, data science, and virtual reality, may face faster knowledge updates and more specialized skill requirements. Therefore, they may place greater value on interactive communication and precise matching functions [47]. By contrast, students in more traditional computing-related fields may perceive relatively less urgency for such dynamic support. However, as these subgroup analyses were exploratory, the findings should be interpreted with caution. These exploratory differences suggest that the functional design of job recommendation systems may need to consider the differentiated communication and matching needs of job seekers in different sub-fields.

4.2. Research Findings and Theoretical Alignment

This study examined how graduating computer science students viewed existing employment recommendation systems and how they responded to a proposed static-dynamic job recommendation approach. The results offer exploratory, survey-based evidence that respondents generally perceived the proposed approach as useful and reported relatively favorable behavioral intentions toward it. In this sense, the findings are broadly consistent with the Technology Acceptance Model (TAM) and can also be interpreted with reference to selected dimensions of the Information Systems (IS) Success Model, particularly perceived usefulness (PU), behavioral intention (BI), information quality, and service-related concerns.
An important pattern in the present findings is the gap between relatively high awareness of existing employment recommendation systems and much lower willingness to use them. This suggests that familiarity with such platforms does not automatically translate into acceptance [41]. In this context, awareness may reflect market exposure, whereas willingness appears to depend more strongly on perceived relevance, information credibility, and confidence in the practical value of platform recommendations. This interpretation is supported by the open-ended comments, in which concerns about fake postings, delayed information updates, and weak major-position matching repeatedly emerged [44].
On the one hand, the findings suggest relatively favorable perceived usefulness toward the proposed approach. The quantitative results show a clear contrast. Only 24.1% of respondents believed that existing systems helped them find suitable jobs. However, 63.2% gave positive evaluations of the proposed static-dynamic approach as a potentially helpful direction for job-search support. This pattern is consistent with prior literature suggesting that interactive recommender features, such as user control and real-time feedback, may enhance the perceived relevance and personalization of recommendations [39,62]. The real-time chat function received particularly favorable ratings (65.3%). This suggests that direct communication is important in technically complex job-search contexts such as computer science. In such contexts, users may need to clarify specific skill requirements and project backgrounds.
For computer science graduates, the appeal of dynamic interaction may stem from the fact that technical positions are often difficult to evaluate through static vacancy descriptions alone. Job seekers may need to clarify technology stacks, project requirements, team expectations, internship conversion pathways, and skill fit. One-way recommendation interfaces may not easily support these needs. In this sense, the favorable evaluation of real-time interaction in the present study may reflect not only a preference for communication, but also a need for uncertainty reduction during job evaluation [44].
On the other hand, the findings indicate relatively favorable stated behavioral intentions toward the proposed approach. A substantial proportion of respondents reported willingness to use the system (67.6%) and willingness to recommend it to others (59.6%), suggesting a generally positive orientation toward possible adoption. The slightly lower recommendation intention may indicate that respondents were somewhat more cautious when considering whether to endorse such a system to peers, particularly with regard to credibility and practical readiness [63]. Overall, this pattern offers a more nuanced understanding of how users may respond to proposed employment recommendation designs before actual system implementation.

4.3. Practical Implications for System Design

Respondents’ feedback suggests that information governance should be treated as a foundational design priority rather than as a secondary enhancement. If users question the authenticity of job information, the practical value of even highly sophisticated recommendation algorithms may be undermined. Future employment recommendation services may therefore benefit from prioritizing company verification, clearer posting provenance, stronger anti-fraud mechanisms, and more transparent explanations of why specific jobs are recommended [23,43].
The findings also suggest that dynamic interaction should not be treated as a generic function that is uniformly added across the entire job-search process. Instead, it may be especially valuable at stages where job seekers need clarification, comparison, or uncertainty reduction, such as shortlist evaluation, fit assessment, or decision support. This implies a more stage-sensitive service design logic, in which interactive features are deployed to support specific decision moments rather than simply increasing the amount of communication available [62].
Beyond Keyword Personalization: Criticism of “jobs not matching majors” (54.1%) calls for moving beyond simple keyword matching. For computer science positions, recommendation mechanisms may need to better account for skill equivalence, portfolios, and the rapidly evolving technological landscape [64]. Integrating user-suggested features, such as skills gap analysis and personalized skills enhancement suggestions, may expand the system’s role beyond simple job listing functions.
Finally, the subgroup findings indicate that future systems may need to provide differentiated support for different user groups. More active job seekers may benefit from deeper filtering, faster response mechanisms, and more direct communication channels, whereas users at earlier exploratory stages may require lighter support and broader opportunity scanning. Similarly, students in rapidly evolving computing fields may place greater value on skill-sensitive matching and more adaptive communication support. These patterns suggest that employment recommendation design may benefit from adjustable interaction depth rather than a one-size-fits-all interface [65].

4.4. Limitations

This study provides valuable empirical evidence on graduating computer science students’ acceptance of a combined static and dynamic job recommendation system. However, several limitations should be considered when interpreting and generalizing the findings.
First, the representativeness and breadth of the sample are limited. All participants in this study came from two higher education institutions in the same region. Although the sample size (N = 386) is acceptable, their geographical location, institution type, and cultural background are relatively homogeneous. This limits the extrapolation validity of the findings to graduates from different regions and levels of institutions across the country (e.g., research universities, vocational colleges) and a wider range of academic disciplines. Future research needs to sample across a broader geographical and institutional scope to verify the generalizability of the study’s conclusions [31].
Second, the findings are based primarily on self-reported perceptions rather than direct interaction with an implemented system [66]. Although many respondents had prior awareness of, and in some cases experience with, existing employment recommendation systems, they did not directly interact with the newly proposed static-dynamic employment recommendation approach in this study. Their evaluations were therefore formed on the basis of the functional descriptions presented in the questionnaire and their prior experience with related platforms. Accordingly, the present findings should not be interpreted as evidence that such a system would necessarily be trusted, adopted, or used in the same way under real-world conditions [63]. Rather, they indicate how respondents evaluated a proposed design concept in survey form. Future research should incorporate functional prototypes, usability testing, and objective behavioral data (e.g., system logs or task-based performance indicators) to strengthen the robustness of the conclusions [67].
In addition, the behavioral intention dimension was measured with only two items, and although its internal consistency was acceptable for exploratory research, this relatively brief measurement structure may limit the depth and stability of interpretation [56,57]. Future studies should employ more comprehensive multi-item scales to strengthen construct measurement.
Finally, the study did not fully account for other potentially relevant background variables that may also be associated with user acceptance. Although the study analyzed differences in professional direction and job-seeking status, other factors could potentially influence acceptance were not adequately included in the model, such as individual levels of technology anxiety, prior experience using similar tools, specific job-seeking strategies, and the overall state of the job market. These uncontrolled variables may have unknown confounding effects on the research results [41].

4.5. Future Directions

Based on the findings and limitations of this study, future research should go beyond evaluating proposed designs through questionnaires alone and begin to test them in more practical settings. This includes developing a functional prototype, carrying out usability testing, and examining actual user behavior through experimental or quasi-experimental studies [68]. On this basis, future research can be further advanced in the following directions.
One important direction is to overcome the problems of high sample homogeneity and limited representativeness. Future research should systematically expand the sampling scope. Specifically, future comparative studies could include graduates from computer science and related majors across different types of institutions, such as research universities, applied undergraduate institutions, and vocational colleges. These studies could cover different economic regions in China, including eastern, central, and western China, and could also be extended to other countries, such as Malaysia, Thailand, and the UK. Such research could test whether the findings remain stable across different educational and cultural contexts. It could also examine how institution type and regional job market differences affect users’ functional needs and acceptance patterns. This would help develop a more general but still differentiated theoretical framework and practical guide.
Another important direction is to compensate for the shortcomings of relying solely on self-reported data [66]. Future research should adopt a mixed approach. While collecting subjective attitude data such as questionnaires and interviews, objective behavioral data should be actively integrated. For example, quasi-experimental studies can be conducted in cooperation with recruitment platforms to analyze users’ real behavioral logs (such as click-through rate, chat interaction depth, resume modification frequency, and application conversion rate) when using prototypes with “static and dynamic” functions. Researchers could link subjective perceived usefulness with objective usage behavior and task completion outcomes [69]. This would improve the validity and persuasiveness of the findings, and also help explain the relationship between system design features and user experience more accurately.
At the same time, future research may also incorporate additional psychological, experiential, and situational variables to better explain the mechanisms underlying user acceptance. To gain a more comprehensive and in-depth understanding of the complex mechanisms influencing user acceptance, future research needs to construct and test a more integrated theoretical model. Based on existing variables, it is necessary to systematically incorporate and measure individual psychological factors (such as technology anxiety, privacy concerns, and algorithmic trust), prior experience factors (such as proficiency in using similar tools in the past), and dynamic situational factors (such as the intensity of specific job-seeking strategies and immediate perception of job market pressure) [70]. Using advanced statistical methods such as structural equation modeling, it is possible to analyze how these variables act as moderating or mediating variables influencing the “perception-intention-behavior” path, thereby revealing the underlying motivations and boundary conditions behind acceptance formation, and providing a precise theoretical basis for refined service design tailored to different user profiles [65,71].

5. Conclusions

This study explored graduating computer science students’ acceptance of a proposed employment recommendation approach that integrates static matching with dynamic interaction through a questionnaire survey of 386 respondents. The results suggest that although respondents evaluated existing employment recommendation systems relatively critically, they expressed more favorable perceptions of the proposed static-dynamic approach, particularly with regard to matching support, real-time interaction, and overall helpfulness. The findings also indicate exploratory subgroup differences, with more active job seekers and respondents from emerging computing fields reporting relatively stronger positive evaluations. Overall, this study provides exploratory survey-based evidence that graduating computer science students may respond positively to employment recommendation concepts that combine profile-based matching with more dynamic interactive support. The findings further suggest that information credibility, responsiveness, communication support, and skill-sensitive matching may be important considerations in future employment service design. However, respondents evaluated a proposed design concept rather than an implemented system. Therefore, the results should be interpreted as perceptions and stated intentions. They should not be treated as evidence of actual adoption or real-world recommendation effectiveness.

Author Contributions

Conceptualization, H.Q. and S.S.; methodology, H.Q. and S.S.; validation, H.Q. and S.S.; formal analysis, H.Q.; investigation, H.Q. and Y.Z.; data curation, H.Q.; writing—original draft preparation, H.Q.; writing—review and editing, S.S., F.F., X.S. and Y.Z.; visualization, H.Q.; supervision, S.S. and F.F.; project administration, H.Q.; funding acquisition, H.Q. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by the Yunnan Provincial Department of Education Science Research Fund Project, grant number 2026J1468.

Institutional Review Board Statement

The study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Review Board of Yunnan College of Business Management, China (protocol code YCBM-2025-012; approval date: 22 November 2025).

Informed Consent Statement

Informed consent was obtained from all subjects involved in the study.

Data Availability Statement

The data presented in this study are available on reasonable request from the corresponding author. The data are not publicly available due to privacy and ethical restrictions.

Acknowledgments

The authors would like to thank the participating students for their time and cooperation in completing the questionnaire survey. ChatGPT-5 (OpenAI) was used only for language polishing, grammar correction, and wording refinement during manuscript preparation. It was not used for data generation, analysis, interpretation, or scientific conclusions. All content was carefully reviewed and verified by the authors.

Conflicts of Interest

The authors declare no conflicts of interest.

References

  1. Portnoi, L.M.; Bagley, S.S.; Rust, V.D. Mapping the terrain: The global competition phenomenon in higher education. In Higher Education, Policy, and the Global Competition Phenomenon; Springer: New York, NY, USA, 2010; pp. 1–13. [Google Scholar] [CrossRef]
  2. Günay, A. Challenges for higher education graduates in the post-pandemic labor market. In International Perspectives and Strategies for Managing an Aging Workforce; Ince, F., Ed.; IGI Global: Hershey, PA, USA, 2022; pp. 144–159. [Google Scholar] [CrossRef]
  3. Xiang, B.; Wang, H.; Wang, H. Is There a Surplus of College Graduates in China? Exploring Strategies for Sustainable Employment of College Graduates. Sustainability 2023, 15, 15540. [Google Scholar] [CrossRef]
  4. Ministry of Education of the People’s Republic of China. 2001 Statistical Communiqué on National Education Development. Available online: https://www.moe.gov.cn/jyb_sjzl/sjzl_fztjgb/tnull_844.html (accessed on 15 March 2026).
  5. Ministry of Education of the People’s Republic of China. 2025 College Graduates Expected to Reach 12.22 Million. Available online: https://www.moe.gov.cn/jyb_xwfb/s5147/202411/t20241115_1163118.html (accessed on 15 March 2026).
  6. Xinhua News Agency. The Number of 2026 College Graduates is Expected to Reach 12.70 Million. Available online: https://www.news.cn/20251120/ead0f25dff2948dfa7f01fa78f207882/c.html (accessed on 15 March 2026).
  7. Hang, T.; Zhou, Y. Chinese University Major Decision and Its Effect on Wages: Modeling Interaction Between Major Specificity and Education-Job Relevancy Using Machine Learning Approaches. J. Chin. Political Sci. 2025, 30, 475–500. [Google Scholar] [CrossRef]
  8. Montobbio, F.; Staccioli, J.; Virgillito, M.E.; Vivarelli, M. The empirics of technology, employment and occupations: Lessons learned and challenges ahead. J. Econ. Surv. 2024, 38, 1622–1655. [Google Scholar] [CrossRef]
  9. Pan, Y.; Gao, F.; Xing, Z. The impact of skill-biased technological change on urban-rural income inequality: Evidence from China. Appl. Geogr. 2026, 187, 103891. [Google Scholar] [CrossRef]
  10. Bone, M.; González Ehlinger, E.; Stephany, F. Skills or degree? The rise of skill-based hiring for AI and green jobs. Technol. Forecast. Soc. Change 2025, 214, 124042. [Google Scholar] [CrossRef]
  11. Lynn, T.; Rosati, P.; Conway, E.; van der Werff, L. The Future of Work: Challenges and Prospects for Organisations, Jobs and Workers; Palgrave Macmillan: Cham, Switzerland, 2023. [Google Scholar] [CrossRef]
  12. Feng, J.; Yang, J.; Li, S.; Miao, Q.; Xi, Y.; Xia, Z. Enhancing person-job fit through multi-temporal career trajectory modeling. Expert Syst. Appl. 2026, 300, 130413. [Google Scholar] [CrossRef]
  13. Xiong, Y.; Yu, J.; Wu, H. The relationship between psychological resilience and employability among higher vocational college students: The chain mediating effects of perceived social support and career decision-making self-efficacy. Front. Psychol. 2026, 16, 1707655. [Google Scholar] [CrossRef] [PubMed]
  14. Zayed, Y.; Salman, Y.; Awad, M.; Hasasneh, A. Employment Recommendation System for Graduates Using Machine Learning. Int. J. Eng. Appl. 2023, 11, 308–317. [Google Scholar] [CrossRef]
  15. Shi, Q. The Employment Management for College Students Based on Deep Learning and Big Data. IEEE Access 2023, 11, 115627–115634. [Google Scholar] [CrossRef]
  16. Ricci, F.; Rokach, L.; Shapira, B. Recommender Systems: Techniques, Applications, and Challenges. In Recommender Systems Handbook; Springer: New York, NY, USA, 2022; pp. 1–35. [Google Scholar] [CrossRef]
  17. Wang, Y. Design and implementation of student job matching system based on personalized recommendation algorithm. Syst. Soft Comput. 2025, 7, 200302. [Google Scholar] [CrossRef]
  18. Zhipin, B. BOSS Zhipin Official Website. Available online: https://www.zhipin.com/ (accessed on 25 April 2026).
  19. Zhaopin. Zhaopin Official Recruitment Platform. Available online: https://www.zhaopin.com/ (accessed on 25 April 2026).
  20. 51job. 51job Official Recruitment Platform. Available online: https://www.51job.com/ (accessed on 25 April 2026).
  21. Lagou. Lagou Official Recruitment Platform. Available online: https://www.lagou.com/ (accessed on 25 April 2026).
  22. Çelik Ertuğrul, D.; Bitirim, S. Job recommender systems: A systematic literature review, applications, open issues, and challenges. J. Big Data 2025, 12, 140. [Google Scholar] [CrossRef]
  23. Mashayekhi, Y.; Li, N.; Kang, B.; Lijffijt, J.; De Bie, T. A Challenge-based Survey of E-recruitment Recommendation Systems. Acm Comput. Surv. 2024, 56, 252. [Google Scholar] [CrossRef]
  24. Huang, X.; Lian, J.; Lei, Y.; Yao, J.; Lian, D.; Xie, X. Recommender AI Agent: Integrating Large Language Models for Interactive Recommendations. ACM Trans. Inf. Syst. 2025, 43, 96. [Google Scholar] [CrossRef]
  25. Konstan, J.; Terveen, L. Human-centered recommender systems: Origins, advances, challenges, and opportunities. AI Mag. 2021, 42, 31–42. [Google Scholar] [CrossRef]
  26. Tang, F.; Zhu, R.; Yao, F.; Wang, J.; Luo, L.; Li, B. Explainable person–job recommendations: Challenges, approaches, and comparative analysis. Front. Artif. Intell. 2025, 8, 1660548. [Google Scholar] [CrossRef] [PubMed]
  27. Knijnenburg, B.P.; Bostandjiev, S.; O’Donovan, J.; Kobsa, A. Inspectability and control in social recommenders. In Proceedings of the Sixth ACM Conference on Recommender Systems, Dublin, Ireland, 9–13 September 2012; pp. 43–50. [Google Scholar] [CrossRef]
  28. He, X.; Liu, Q.; Jung, S. The impact of recommendation system on user satisfaction: A moderated mediation approach. J. Theor. Appl. Electron. Commer. Res. 2024, 19, 448–466. [Google Scholar] [CrossRef]
  29. He, C.; Parra, D.; Verbert, K. Interactive recommender systems: A survey of the state of the art and future research challenges and opportunities. Expert Syst. Appl. 2016, 56, 9–27. [Google Scholar] [CrossRef]
  30. Roppelt, J.S.; Schuster, A.; Greimel, N.S.; Kanbach, D.K.; Sen, K. Towards effective adoption of artificial intelligence in talent acquisition: A mixed method study. Int. J. Inf. Manag. 2025, 82, 102870. [Google Scholar] [CrossRef]
  31. Cochran, W.G. Sampling Techniques, 3rd ed.; Johan Wiley & Sons: New York, NY, USA, 1977. [Google Scholar]
  32. Davis, F.D. Perceived Usefulness, Perceived Ease of Use, and User Acceptance of Information Technology. Manag. Inf. Syst. Q. 1989, 13, 319–340. [Google Scholar] [CrossRef] [PubMed]
  33. DeLone, W.; McLean, E. Information Systems Success: The Quest for the Dependent Variable. Inf. Syst. Res. 1992, 3, 60–95. [Google Scholar] [CrossRef]
  34. DeLone, W.H.; McLean, E.R. The DeLone and McLean model of information systems success: A ten-year update. J. Manag. Inf. Syst. 2003, 19, 9–30. [Google Scholar] [CrossRef]
  35. Venkatesh, V.; Davis, F.D. A theoretical extension of the technology acceptance model: Four longitudinal field studies. Manag. Sci. 2000, 46, 186–204. [Google Scholar] [CrossRef]
  36. Xu, Y.; Chen, L. Personalized recommendations and consumer trust: The role of perceived control and locus of control. Acta Psychol. 2025, 261, 105936. [Google Scholar] [CrossRef]
  37. Wang, Z.; Wei, W.; Xu, C.; Xu, J.; Mao, X. Person-job fit estimation from candidate profile and related recruitment history with co-attention neural networks. Neurocomputing 2022, 501, 14–24. [Google Scholar] [CrossRef]
  38. Akram, S.; Buono, P.; Lanzilotti, R. Recruitment chatbot acceptance in a company: A mixed method study on human-centered technology acceptance model. Pers. Ubiquitous Comput. 2024, 28, 961–984. [Google Scholar] [CrossRef]
  39. Gao, C.; Lei, W.; He, X.; de Rijke, M.; Chua, T.-s. Advances and Challenges in Conversational Recommender Systems: A Survey. AI Open 2021, 2, 100–126. [Google Scholar] [CrossRef]
  40. Mensah, I.K.; Zeng, G.; Luo, C. Determinants of social commerce purchase and recommendation intentions within the context of swift guanxi among Chinese college students. Sage Open 2023, 13, 21582440231175370. [Google Scholar] [CrossRef]
  41. Marikyan, D.; Papagiannidis, S.; Stewart, G. Technology acceptance research: Meta-analysis. J. Inf. Sci. 2023, 52, 429–450. [Google Scholar] [CrossRef]
  42. Al Naqbi, S.H. A mixed-method approach to post-implementation success of technology performance in UAE universities: Assessing DeLone and McLean IS success model. Sage Open 2024, 14, 21582440241240827. [Google Scholar] [CrossRef]
  43. Kingsley, S.; Silberman, M.S.; Wang, C.; Lambeth, R.; Zhi, J.; Eslami, M.; Li, B.; Bigham, J. ‘Your Duties Are To Sweep A Floor Remotely’: Low Information Quality in Job Advertisements is a Barrier to Low-Income Job-Seekers’ Successful Use of Digital Platforms. In Proceedings of the 3rd Annual Meeting of the Symposium on Human-Computer Interaction for Work, Newcastle upon Tyne, UK, 25–27 June 2024; pp. 18:11–18:20. [Google Scholar]
  44. Kissi, P.S. Job seekers satisfy or dissatisfy with the existing electronic recruitment: A theoretical and empirical investigation. Cogent Bus. Manag. 2023, 10, 2278233. [Google Scholar] [CrossRef]
  45. Du, Y.; Liu, H.; Zhu, H.; Song, Y.; Zheng, Z.; Wu, Z. Quasi-Metric Learning for Bilateral Person-Job Fit. IEEE Trans. Pattern Anal. Mach. Intell. 2025, 47, 3947–3960. [Google Scholar] [CrossRef]
  46. Gong, Z.; Song, Y.; Zhang, T.; Wen, J.-R.; Zhao, D.; Yan, R. Your Career Path Matters in Person-Job Fit. Proc. AAAI Conf. Artif. Intell. 2024, 38, 8427–8435. [Google Scholar] [CrossRef]
  47. Saito, Y.; Sugiyama, K. Multi-Behavior Job Recommendation with Dynamic Availability. In Proceedings of the 1st ACM SIGIR-AP Conference on Research and Development in Information Retrieval in the Asia Pacific Region, Beijing, China, 26–29 November 2023; pp. 264–271. [Google Scholar] [CrossRef]
  48. Ogunniye, G.; Legastelois, B.; Rovatsos, M.; Dowthwaite, L.; Portillo, V.; Vallejos, E.; Zhao, J.; Jirotka, M. Understanding User Perceptions of Trustworthiness in E-Recruitment Systems. IEEE Internet Comput. 2021, 25, 23–32. [Google Scholar] [CrossRef]
  49. Boldi, A.; Silacci, A.; Rapp, A.; Caon, M. Designing for transparency: A web job board for e-recruitment to explore job seekers’ privacy behaviours. Behav. Inf. Technol. 2025, 44, 3038–3063. [Google Scholar] [CrossRef]
  50. Amaar, A.; Aljedaani, W.; Rustam, F.; Ullah, S.; Rupapara, V.; Ludi, S. Detection of Fake Job Postings by Utilizing Machine Learning and Natural Language Processing Approaches. Neural Process. Lett. 2022, 54, 2219–2247. [Google Scholar] [CrossRef]
  51. Hosain, M.S.; Amin, M.B.; Debnath, G.C.; Rahaman, M.A. The use of Artificial Intelligence (AI) in the hiring process: Job applicants’ perceptions of procedural justice. Comput. Hum. Behav. Rep. 2025, 19, 100713. [Google Scholar] [CrossRef]
  52. Siro, C.; Aliannejadi, M.; De Rijke, M. Understanding and Predicting User Satisfaction with Conversational Recommender Systems. ACM Trans. Inf. Syst. 2024, 42, 55. [Google Scholar] [CrossRef]
  53. Naeem, M.; Ozuem, W.; Howell, K.; Ranfagni, S. A step-by-step process of thematic analysis to develop a conceptual model in qualitative research. Int. J. Qual. Methods 2023, 22, 1–18. [Google Scholar] [CrossRef]
  54. Verma, J.P. Data Analysis in Management with SPSS Software; Springer Science & Business Media: Berlin/Heidelberg, Germany, 2012. [Google Scholar] [CrossRef]
  55. Tavakol, M.; Dennick, R. Making sense of Cronbach’s alpha. Int. J. Med. Educ. 2011, 2, 53–55. [Google Scholar] [CrossRef]
  56. Nunnally, J.C.; Bernstein, I.H. Psychometric Theory, 3rd ed.; McGraw-Hill: New York, NY, USA, 1994. [Google Scholar]
  57. Cortina, J.M. What is coefficient alpha? An examination of theory and applications. J. Appl. Psychol. 1993, 78, 98–104. [Google Scholar] [CrossRef]
  58. Hair, J.F.; Black, W.C.; Babin, B.J.; Anderson, R.E. Multivariate Data Analysis, 8th ed.; Cengage Learning: Andover, UK, 2019. [Google Scholar]
  59. Nigar, M.; Juli, J.F.; Golder, U.; Alam, M.J.; Hossain, M.K. Artificial intelligence and technological unemployment: Understanding trends, technology’s adverse roles, and current mitigation guidelines. J. Open Innov. Technol. Mark. Complex. 2025, 11, 100607. [Google Scholar] [CrossRef]
  60. Agresti, A. Categorical Data Analysis, 3rd ed.; John Wiley & Sons: Hoboken, NJ, USA, 2013. [Google Scholar]
  61. Armstrong, L.; Metaxa, D. Navigating automated hiring: Perceptions, strategy use, and outcomes among young job seekers. Proc. ACM Hum.-Comput. Interact. 2025, 9, CSCW140. [Google Scholar] [CrossRef]
  62. Jannach, D.; Manzoor, A.; Cai, W.; Chen, L. A survey on conversational recommender systems. ACM Comput. Surv. 2021, 54, 105. [Google Scholar] [CrossRef]
  63. Deriu, V.; Pozharliev, R.; De Angelis, M. How trust and attachment styles jointly shape job candidates’ AI receptivity. J. Bus. Res. 2024, 179, 114717. [Google Scholar] [CrossRef]
  64. Smaldone, F.; Ippolito, A.; Lagger, J.; Pellicano, M. Employability skills: Profiling data scientists in the digital labour market. Eur. Manag. J. 2022, 40, 671–684. [Google Scholar] [CrossRef]
  65. Ma, Y.; Ziegler, J. Investigating meta-intents: User interaction preferences in conversational recommender systems. User Model. User-Adapt. Interact. 2024, 34, 1535–1580. [Google Scholar] [CrossRef]
  66. Podsakoff, P.M.; Podsakoff, N.P.; Williams, L.J.; Huang, C.; Yang, J. Common method bias: It’s bad, it’s complex, it’s widespread, and it’s not easy to fix. Annu. Rev. Organ. Psychol. Organ. Behav. 2024, 11, 17–61. [Google Scholar] [CrossRef]
  67. Knijnenburg, B.P.; Willemsen, M.C.; Gantner, Z.; Soncu, H.; Newell, C. Explaining the user experience of recommender systems. User Model. User-Adapt. Interact. 2012, 22, 441–504. [Google Scholar] [CrossRef]
  68. Jin, Y.; Chen, L.; Cai, W.; Zhao, X. CRS-Que: A user-centric evaluation framework for conversational recommender systems. ACM Trans. Recomm. Syst. 2024, 2, 2. [Google Scholar] [CrossRef]
  69. Jannach, D. Evaluating conversational recommender systems: A landscape of research. Artif. Intell. Rev. 2023, 56, 2365–2400. [Google Scholar] [CrossRef]
  70. Chen, A.; Han, F.; Zhang, X.; Lu, Y. Cracking the AI recruitment code: Striving for transparency in finding the right person–job fit. Inf. Manag. 2025, 62, 104156. [Google Scholar] [CrossRef]
  71. Niessen, A.S.M.; Neumann, M. Transparency and Trust in Simple Algorithmic Hiring Procedures. Int. J. Sel. Assess. 2026, 34, e70045. [Google Scholar] [CrossRef]
Information 17 00479 g001
Figure 1. Descriptive analysis of perceived usefulness (N = 386).
Figure 1. Descriptive analysis of perceived usefulness (N = 386).
Information 17 00479 g001
Information 17 00479 g002
Figure 2. Descriptive analysis of behavioral intention (N = 386).
Figure 2. Descriptive analysis of behavioral intention (N = 386).
Information 17 00479 g002
Table 1. Sociodemographic characteristics and job-seeking status of the respondents (N = 386).
Table 1. Sociodemographic characteristics and job-seeking status of the respondents (N = 386).
CharacteristicNumber (n)Percentage (%)
Gender
Male22257.5
Female16442.5
Major/Specialization
Computer Science and Technology16542.7
Artificial Intelligence7820.2
Internet of Things Engineering297.5
Software Engineering236.0
Virtual Reality Technology225.7
Network Engineering215.4
Data Science184.7
Other307.8
Post-graduation Intention
Plan to seek employment36594.6
No immediate employment plans215.4
Current Job Search Status
Actively searching, no interview yet19851.3
Currently in an internship7619.7
Not actively looking9324.1
Received job offer (signed contract)194.9
Table 2. Internal consistency analysis of the questionnaire items.
Table 2. Internal consistency analysis of the questionnaire items.
ItemCorrected Total Correlationα Coefficient with Item DeletedCronbach Alpha
Overall scale  0.818
Do you think improving the accuracy of job matching in the employment recommendation system will help you find a suitable job?0.7060.752
Do you think adding real-time chat interaction to the employment recommendation system will help you find a suitable job?0.7640.738
Do you think this combination of static and dynamic employment recommendation systems can help you find a job?0.7680.733
Would you use such an employment recommendation system?0.4550.827
Would you be willing to recommend such an employment recommendation system to your classmates or friends?0.3910.842
Perceived usefulness (PU)  0.928
Do you think improving the accuracy of job matching in the employment recommendation system will help you find a suitable job?0.7730.961
Do you think adding real-time chat interaction to the employment recommendation system will help you find a suitable job?0.9010.860
Do you think this combination of static and dynamic employment recommendation systems can help you find a job?0.8920.864
Behavioral intention (BI)  0.630
Would you use such an employment recommendation system?0.460/
Would you be willing to recommend such an employment recommendation system to your classmates or friends?0.460/
Table 3. KMO and Bartlett’s test results.
Table 3. KMO and Bartlett’s test results.
TestValue
KMO0.735
Bartlett’s Test of Sphericity1269.598
df10
Sig.0.000
Table 4. Rotated component loadings and communalities.
Table 4. Rotated component loadings and communalities.
ItemComponent 1Component 2 Communality
Do you think improving the accuracy of job matching in the employment recommendation system will help you find a suitable job?0.8670.2090.796
Do you think adding real-time chat interaction to the employment recommendation system will help you find a suitable job?0.9470.1580.923
Do you think this combination of static and dynamic employment recommendation systems can help you find a job?0.9400.1800.916
Would you use such an employment recommendation system?0.2170.8180.716
Would you be willing to recommend such an employment recommendation system to your classmates or friends?0.1200.8570.749
Table 5. User evaluation of the existing job recommendation systems.
Table 5. User evaluation of the existing job recommendation systems.
OptionsNumber (n)Percentage (%)
Strongly agree379.6
Agree5614.5
Neutral10627.5
Disagree11229.0
Strongly disagree7519.4
Table 6. Reported limitations of existing employment recommendation systems.
Table 6. Reported limitations of existing employment recommendation systems.
LimitationsNumber (n)Percentage (%)
Severe deficiencies in matching functionality and responsiveness27471
Multiple challenges to information quality21455.4
Significant mismatch in professional suitability20954.1
Other issues82.1
Table 7. Positive response rates and chi-square test results by job-seeking status.
Table 7. Positive response rates and chi-square test results by job-seeking status.
ItemActive Job-Seeking
(n, %)		Less Active Job-Seeking (n, %)χ2p-Value
Do you think improving the accuracy of job matching in the employment recommendation system will help you find a suitable job?174/274 (63.5%)56/112 (50.0%)6.020 0.014
Do you think adding real-time chat interaction to the employment recommendation system will help you find a suitable job?190/274 (69.3%)62/112 (55.4%)6.8620.009
Do you think this combination of static and dynamic employment recommendation systems can help you find a job?183/274 (66.8%)61/112 (54.5%)5.193 0.023
Would you use such an employment recommendation system?195/274 (71.2%)66/112 (58.9%)5.439 0.020
Would you be willing to recommend such an employment recommendation system to your classmates or friends?173/274 (63.1%)57/112 (50.9%)4.951 0.026
Note: Positive responses include “agree” and “strongly agree.” The valid sample size for this comparison was N = 386.
Table 8. Positive response rates and chi-square test results by professional orientation.
Table 8. Positive response rates and chi-square test results by professional orientation.
ItemEmerging Fields
(n, %)		Traditional Computing
(n, %)		χ2p-Value
Do you think improving the accuracy of job matching in the employment recommendation system will help you find a suitable job?80/118 (67.8%)135/238 (56.7%)4.044 0.044
Do you think adding real-time chat interaction to the employment recommendation system will help you find a suitable job?87/118 (73.7%)147/238 (61.8%)5.013 0.025
Do you think this combination of static and dynamic employment recommendation systems can help you find a job?84/118 (71.2%)142/238 (59.7%)4.518 0.034
Would you use such an employment recommendation system?91/118 (77.1%)151/238 (63.4%)6.775 0.009
Would you be willing to recommend such an employment recommendation system to your classmates or friends?83/118 (70.3%)130/238 (54.6%)8.108 0.004
Note: Positive responses include “agree” and “strongly agree.” Valid N = 356; respondents in Other were excluded.
Table 9. Mann–Whitney U test results by subgroup.
Table 9. Mann–Whitney U test results by subgroup.
Grouping VariableComparisonDimensionGroup 1 Mean RankGroup 2 Mean RankUZp-Value
Professional orientationEmerging Fields (n = 118) vs. Traditional Computing (n = 238)PU193.98170.8212,215.000−2.0180.044
Emerging Fields (n = 118) vs. Traditional Computing (n = 238)BI198.93168.3711,631.500−2.6740.007
Job-seeking statusActive job-seeking (n = 274) vs. Less active job-seeking (n = 112)PU201.43174.0913,170.500−2.2060.027
Active job-seeking (n = 274) vs. Less active job-seeking (n = 112)BI204.23167.2512,404.500−2.9960.003
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Qu, H.; Sahrani, S.; Fauzi, F.; Song, X.; Zhao, Y. User Acceptance of a Proposed Static-Dynamic Employment Recommendation Approach Among Computer Science Graduating Students. Information 2026, 17, 479. https://doi.org/10.3390/info17050479

AMA Style

Qu H, Sahrani S, Fauzi F, Song X, Zhao Y. User Acceptance of a Proposed Static-Dynamic Employment Recommendation Approach Among Computer Science Graduating Students. Information. 2026; 17(5):479. https://doi.org/10.3390/info17050479

Chicago/Turabian Style

Qu, Huafeng, Shafrida Sahrani, Fariza Fauzi, Xiacheng Song, and Yanfeng Zhao. 2026. "User Acceptance of a Proposed Static-Dynamic Employment Recommendation Approach Among Computer Science Graduating Students" Information 17, no. 5: 479. https://doi.org/10.3390/info17050479

APA Style

Qu, H., Sahrani, S., Fauzi, F., Song, X., & Zhao, Y. (2026). User Acceptance of a Proposed Static-Dynamic Employment Recommendation Approach Among Computer Science Graduating Students. Information, 17(5), 479. https://doi.org/10.3390/info17050479

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop