BUSI 643 Week 7 Literature Review Example

Asynchronous Video Interviews in Employee Selection: A Review of Recent Research on Design, Applicant Reactions, Bias and Integrity

Introduction

The employment interview is among the most widely used selection methods, and the selection chapters of the Heneman staffing text devote considerable attention to structuring it well. Over the past decade, a new format has spread quickly: the asynchronous video interview, in which applicants record answers to preset questions on their own schedule and recruiters, or increasingly algorithms, evaluate the recordings later. Employers adopt AVIs because they let a small recruiting team screen many applicants without scheduling calls, they present every applicant with identical questions and they create a record that several raters can review. Those features align with the structured interviewing principles the textbook recommends. The format also raises new concerns: applicants may find it impersonal, raters see applicants' homes as well as their answers, scoring may be automated and applicants can now read answers generated by artificial intelligence.

This review examines what peer-reviewed research published since 2021 shows about AVIs. I searched Business Source Complete, PsycINFO and ABI/INFORM through the Jerry Falwell Library using the terms "asynchronous video interview," "one-way video interview" and "digital interview," limited to peer-reviewed journal articles from 2021 onward, and screened results for empirical studies of employee selection. Thirteen articles met these criteria. They are organized below into five themes: how employers use AVIs in practice, how design features affect applicant reactions, what automated scoring can do, how visual background information can bias evaluations and how impression management and AI assistance threaten the format's integrity.

How Employers Use Asynchronous Video Interviews

Research on AVIs has often studied designs that researchers chose, so it is useful to begin with what employers actually do. Dunlop et al. (2022) analyzed archival data from an Australian AVI vendor covering more than 2.5 million responses from 627,999 candidates to 52,623 questions in 12,105 interview templates. The typical AVI was used for a small applicant pool, a median of ten candidates, and most comprised four or five questions. Candidates were usually given 30 seconds to prepare and two minutes to record each answer, and they were rarely allowed to preview questions or rerecord. The settings employers used were highly consistent with the platform's defaults.

Two implications follow. First, many employers are not using AVIs to screen thousands of applicants, as is often assumed, but to interview a handful, which weakens the efficiency argument and raises the question of whether a live conversation would serve better. Second, because employers tend to accept vendor defaults, the design decisions that shape applicant experience are often made by vendors rather than by staffing professionals. Defaults are not neutral. A vendor that sets a short preparation window and prohibits rerecording may do so to make the platform look rigorous, but as the next section shows, those very settings are associated with worse applicant reactions, and an employer that never revisits them inherits the cost without knowing it. The Heneman text's emphasis on deliberate staffing system design applies here: an organization adopting an AVI should treat each setting as a choice.

Design Features and Applicant Reactions

The most consistent finding across studies is that applicants react somewhat less favorably to AVIs than to live interviews, and that design features can narrow the gap. The largest field study, by Tilston et al. (2024), analyzed the experiences of 27,809 real applicants from 33 companies in 11 countries. Allowing more preparation time and permitting applicants to rerecord answers were associated with more favorable reactions, while including more questions was associated with more negative ones. The effects differed by group: applicants over about 30 reacted especially negatively to AVIs with many questions, younger applicants preferred longer maximum response times and women reacted more positively than men to additional preparation time.

Experimental work explains why preparation time matters. Basch et al. (2021) found in a pilot study and a main study that longer preparation time improved interview performance, and that the effect operated through more active preparation of responses rather than through reduced strain or more impression management. Preparation time did not directly change fairness perceptions, though it improved them indirectly by encouraging honest self-presentation. Set against the 30-second norm reported by Dunlop et al. (2022), this research suggests that common practice may be leaving performance and goodwill on the table.

Not every intuitively appealing design change helps. Niemitz et al. (2024) tested in an actual hiring process whether presenting questions by video, with a person asking them, rather than as text would increase applicants' sense of social presence and improve their reactions. They found no evidence that video questions affected social presence, reactions, impression management or interview performance, although their sample of 76 applicants was small. Patel et al. (2025) tested a more substantive change, adding structured follow-up questions, or probes, to an AVI. In their first experiment, probing improved applicants' perceived opportunity to perform and their interview ratings, but the effect on ratings was largely explained by longer responses; when a second experiment matched response time, probing's advantage shrank.

Finally, the comparison point matters. Basch et al. (2022) noted that AVIs are usually compared with live interviews, although they are typically used earlier, at the preselection stage. When the authors compared AVIs with other preselection tools, online cognitive ability tests and online application documents, potential applicants did not perceive AVIs as less fair, though perceptions differed on usefulness, ease of use, privacy concerns and organizational attractiveness. Read together, these studies suggest that AVIs are not inherently unpopular; reactions depend heavily on design and on what the AVI replaces. That conclusion matters for how the Heneman text's concern with applicant reactions should be applied: an organization that replaces a phone screen with a well-designed AVI may see little change in how applicants view it, while one that replaces a live first interview with a poorly configured AVI may lose candidates who have other offers. Measuring reactions and completion rates after a change, rather than assuming them, is the practical lesson.

Automated Scoring: Promise and Applicant Response

Many vendors now offer automated scoring of AVIs using machine learning applied to applicants' words, voices and facial expressions. Koutsoumpis et al. (2024) examined the psychometric properties of such scoring with 710 participants who completed a mock AVI of eight personality-related questions. Their algorithm explained substantially more variance in observer-rated extraversion and conscientiousness and in interview performance ratings than in self-reported personality. Consistent with trait activation theory, prediction was stronger when questions were relevant to the trait. However, test-retest reliability over seven months fell below the standards expected in personnel selection, and the authors found some evidence of bias related to age, gender and attractiveness. An algorithm that predicts what raters think is not the same as one that predicts how people will perform on the job.

Applicants' reactions to automated scoring add a further cost. Oostrom et al. (2024) informed participants in two experiments that either an algorithm or a recruiter had evaluated their AVI and personality inventory. Algorithm-based evaluation increased feelings of emotional creepiness in the first study and, in both studies, reduced perceived fairness and predictive validity, which in turn lowered feedback acceptance, organizational attraction and intentions to accept a job. Contrary to expectations, the effect did not depend on whether the applicant was selected. Falls et al. (2025) asked whether explanations could soften these reactions. Participants who completed an automatically scored AVI and were then rejected received one of several explanations; none had the hypothesized direct effects, but a combined explanation emphasizing both the consistency of the process and the opportunity it gave applicants improved organizational attraction and related outcomes indirectly, through perceptions of procedural justice and interpersonal treatment.

The automated scoring literature therefore points in a cautious direction. Algorithms may reproduce human raters' judgments efficiently, but their reliability, fairness across groups and acceptance by applicants are not yet established well enough to justify relying on them for consequential decisions.

Bias From What the Camera Shows

Traditional interviews take place in the employer's space; AVIs take place in the applicant's. That difference introduces information that interviewers never used to see. Roulin et al. (2023) conducted three experiments in which applicants' video backgrounds signaled parental status, sexual orientation or political affiliation. Effects varied: applicants shown as parents were seen as warmer and received higher interview ratings without being penalized on competence, sexual orientation cues had no effect, and applicants who appeared to support the evaluator's political party were rated warmer and higher on interview and expected job performance. Several of these characteristics are not legally protected everywhere, which makes the risk easy to overlook.

Springle and Bourdage (2025) extended this work to socioeconomic status. In an experiment with 260 American raters, evaluators could recognize differences in applicants' socioeconomic status from their backgrounds, and the authors induced high or low cognitive load to mimic a busy hiring manager. Neither status cues nor load directly changed perceived hireability, and evaluators' own characteristics did not directly influence ratings either, although the authors reported some more complex patterns. The two studies together suggest that background bias is real for some cues and not others, and that its effect depends on the rater, the cue and possibly the context. For practice, the safest response is the one Roulin et al. (2023) recommended: encourage or require neutral or virtual backgrounds and train raters to focus on answer content. A further option is to separate what raters see from what they score, for instance by having raters evaluate transcripts of answers before viewing the videos, a design that would remove background cues entirely for the content rating. No study in this review tested such a procedure, which is itself a useful direction for future research.

Impression Management and AI Assistance

Because applicants control the recording environment, AVIs may invite more polished self-presentation than live interviews. Arseneault and Roulin (2024) studied impression management in AVIs among 582 participants from 10 countries completing an eight-question interview for a bank manager role. Use of impression management differed only slightly across countries and was seldom related to cultural dimensions. Consistent with in-person research, honest tactics such as self-promotion were positively related to interview performance, while deceptive tactics such as extensive image creation were negatively related to it.

Generative AI creates a more serious integrity problem. Canagasuriam and Lukacik (2025) randomly assigned 245 participants to complete an AVI on their own, by reading ChatGPT-generated answers verbatim or by adapting ChatGPT answers personalized with their résumés. Both ChatGPT groups received considerably higher overall and content scores than the unassisted group, though their delivery ratings did not differ and they received lower honesty ratings. They also rated the AVI lower on procedural justice. The finding is troubling for the format: if an applicant can raise interview scores substantially by reading AI-generated answers, the AVI measures access to and skill with a chatbot as much as the qualities the employer cares about. The short preparation windows that Dunlop et al. (2022) found to be standard may limit this risk, but longer preparation time, which improves reactions and performance (Basch et al., 2021), may increase it, creating a design trade-off the research has only begun to address.

Gaps in the Research

Four gaps stand out. First, almost no recent study links AVI scores to subsequent job performance in real organizations; most outcomes are interview ratings or applicant perceptions, so the format's predictive validity rests largely on assumptions borrowed from structured interview research. Second, many experiments use online participants rather than real applicants, and the field studies that exist, such as Tilston et al. (2024) and Niemitz et al. (2024), sometimes produce weaker or different effects. Third, research on automated scoring has examined prediction of rater judgments more than fairness across protected groups in real hiring. Fourth, the AI-cheating problem is new, and research has not yet tested countermeasures such as follow-up live interviews, randomized questions or shorter preparation windows.

Implications for Staffing Practice

Despite these gaps, the evidence supports several practical conclusions for an organization using or considering AVIs. Settings should be chosen deliberately rather than accepted from the vendor: moderate preparation time, the option to rerecord and a limited number of questions are associated with better reactions. Raters should evaluate answers against structured, job-related scoring guides, as the Heneman text recommends for all interviews, and applicants should be encouraged to use neutral backgrounds. Automated scoring should be used, if at all, as one input reviewed by people, with applicants told how decisions are made and given a full explanation when rejected. Finally, because AI assistance can inflate scores, an AVI should be followed by a live interview for finalists in which applicants discuss their recorded answers.

Legal considerations reinforce these conclusions. Illinois's Artificial Intelligence Video Interview Act, in effect since 2020, requires employers that use AI to analyze video interviews to notify applicants, explain how the technology works and obtain consent, and other jurisdictions have followed with rules on automated employment decisions. Federal antidiscrimination law, which the Heneman text covers in its legal compliance chapter, applies to an AVI as to any selection procedure: if scores screen out a protected group disproportionately, the employer must be able to show that the procedure is job related. The background-bias studies suggest that adverse impact could arise from sources no one intended, such as cues about family or social class, so monitoring pass rates by group belongs in any AVI program from the start.

Conclusion

Asynchronous video interviews have become a common screening tool faster than research could evaluate them. The thirteen studies reviewed here show that the format's reputation for poor applicant reactions depends largely on design, that automated scoring is promising but not yet reliable or accepted enough for high-stakes decisions, that the applicant's background can introduce bias the traditional interview never faced and that generative AI poses a new threat to the format's validity. Used deliberately, as one structured step in a broader selection system, AVIs can serve employers and applicants well. Used by default, they risk measuring the wrong things.

References

Arseneault, R., & Roulin, N. (2024). Investigating impression management use in asynchronous video interviews across 10 countries. International Journal of Selection and Assessment, 32(3), 461-477. https://doi.org/10.1111/ijsa.12476

Basch, J. M., Brenner, F., Melchers, K. G., Krumm, S., Dräger, L., Herzer, H., & Schuwerk, E. (2021). A good thing takes time: The role of preparation time in asynchronous video interviews. International Journal of Selection and Assessment, 29(3-4), 378-392. https://doi.org/10.1111/ijsa.12341

Basch, J. M., Melchers, K. G., & Büttner, J. C. (2022). Preselection in the digital age: A comparison of perceptions of asynchronous video interviews with online tests and online application documents in a simulation context. International Journal of Selection and Assessment, 30(4), 639-652. https://doi.org/10.1111/ijsa.12403

Canagasuriam, D., & Lukacik, E. (2025). ChatGPT, can you take my job interview? Examining artificial intelligence cheating in the asynchronous video interview. International Journal of Selection and Assessment, 33(1), Article e12491. https://doi.org/10.1111/ijsa.12491

Dunlop, P. D., Holtrop, D., & Wee, S. (2022). How asynchronous video interviews are used in practice: A study of an Australian-based AVI vendor. International Journal of Selection and Assessment, 30(3), 448-455. https://doi.org/10.1111/ijsa.12372

Falls, B., Willis, C., & Liff, J. (2025). The impact of explanations on applicant reactions to automated asynchronous video interviews. International Journal of Selection and Assessment, 33(2), Article e70009. https://doi.org/10.1111/ijsa.70009

Koutsoumpis, A., Ghassemi, S., Oostrom, J. K., Holtrop, D., van Breda, W., Zhang, T., & de Vries, R. E. (2024). Beyond traditional interviews: Psychometric analysis of asynchronous video interviews for personality and interview performance evaluation using machine learning. Computers in Human Behavior, 154, Article 108128. https://doi.org/10.1016/j.chb.2023.108128

Niemitz, N., Rietheimer, L., König, C., & Langer, M. (2024). The (null) effects of video questions on applicant reactions in asynchronous video interviews: Evidence from an actual hiring context. International Journal of Selection and Assessment, 32(3), 421-428. https://doi.org/10.1111/ijsa.12470

Oostrom, J. K., Holtrop, D., Koutsoumpis, A., van Breda, W., Ghassemi, S., & de Vries, R. E. (2024). Applicant reactions to algorithm- versus recruiter-based evaluations of an asynchronous video interview and a personality inventory. Journal of Occupational and Organizational Psychology, 97(1), 160-189. https://doi.org/10.1111/joop.12465

Patel, R. D., Powell, D. M., Roulin, N., & Spence, J. R. (2025). Tell me more! Examining the benefits of adding structured probing in asynchronous video interviews. International Journal of Selection and Assessment, 33(1), Article e12514. https://doi.org/10.1111/ijsa.12514

Roulin, N., Lukacik, E., Bourdage, J. S., Clow, L., Bakour, H., & Diaz, P. (2023). Bias in the background? The role of background information in asynchronous video interviews. Journal of Organizational Behavior, 44(3), 458-475. https://doi.org/10.1002/job.2680

Springle, M., & Bourdage, J. S. (2025). Scene, not heard: Exploring the influence of socioeconomic status background cues in asynchronous video interviews. International Journal of Selection and Assessment, 33(1), Article e12504. https://doi.org/10.1111/ijsa.12504

Tilston, O., Krings, F., Roulin, N., Bourdage, J. S., & Fetzer, M. (2024). Reactions to asynchronous video interviews: The role of design decisions and applicant age and gender. Human Resource Management, 63(2), 313-332. https://doi.org/10.1002/hrm.22202