Selection interviews are one of the most popular methods for personnel selection (Wilk & Cappelli, 2003). However, applicants commonly use faking to present themselves favorably in selection interviews (Levashina & Campion, 2007). In line with this, Weiss and Feldman (2006), for example, found that 81% of the interviewees told at least one lie in a short 15-min interview, with an average of 2.19 lies per interview. Furthermore, there is widespread fear that faking can jeopardize the criterion-related validity of selection procedures (e.g., Salgado, 2016). In addition, it has been found that faking behavior can hardly be detected in interviews (see Melchers et al., 2020; for a review). Because of this, knowledge on suitable countermeasures that avoid or at least reduce faking in selection interviews is of relevance for high-stakes selection settings. However, hardly any research on such countermeasures is available (Melchers et al., 2020). Therefore, the first aim of our research was to identify possible countermeasures and to provide an initial evaluation of their effects on faking intentions or faking behavior in interviews.

In addition to the unclear effectiveness of possible countermeasures to reduce faking in interviews, another question concerns potential side effects of these countermeasures. Specifically, some countermeasures such as the use of warnings not to fake might lead to negative applicant reactions (Burns et al., 2015; Lammers, 2017). However, given that interviews are often not only used to select but also to recruit applicants (Dipboye et al., 2012), countermeasures that lead to negative applicant reactions may impair the recruitment function of interviews. Therefore, the second aim of this paper was to examine the effects of potential countermeasures against faking in interviews on applicant reactions.

1 THEORETICAL BACKGROUND OF INTERVIEW FAKING

Levashina and Campion (2007) define faking in interviews as “conscious distortions of answers to the interview questions to obtain a better score on the interview and/or otherwise create favorable perceptions” (p. 1639). According to them, four subfacets of faking can be distinguished: ingratiation, image protection, slight image creation, and extensive image creation. Ingratiation represents a behavior in which applicants insincerely compliment the interviewer or organization, or try to express the same values, beliefs, or opinions as them. Image protection means that applicants deliberately omit or disguise disadvantageous information or distance themselves from it to maintain the image of a good applicant. In the case of slight image creation, they exaggerate their competencies, or their fit with the organization, or the job. Finally, applicants show extensive image creation when they invent qualifications or stories, or present accomplishments of others as their own (Levashina & Campion, 2007).

There are different theoretical models to explain the occurrence and amount of faking (e.g., Levashina & Campion, 2006; McFarland & Ryan, 2000, 2006; Roulin, Krings & Binggeli, 2016). Although these models differ in detail—and not all of them are specifically directed at interview faking—they all have in common that situational, motivational, and ability antecedents are relevant for the occurrence of faking. This is also the case with the faking model by Levashina and Campion (2006), which is the only model that explicitly refers to faking in interviews.

The model by Levashina and Campion (2006) assumes that faking depends on applicants' capacity, willingness, and opportunity to fake. Capacity to fake is defined by capabilities that enable applicants to fake effectively. This includes, for example, oral skills, cognitive ability, and knowledge of the constructs being measured. Willingness to fake refers to dispositional and situational antecedents that influence the applicants' tendency to distort their answers. Dispositional antecedents include personality traits such as Agreeableness, Honesty-humility, or Integrity, and situational antecedents include aspects such as the probability of getting caught or unfair treatment. Finally, opportunity to fake is defined by interview characteristics that cannot be influenced by applicants and that determine to what extent faking is possible. This involves, for example, the purpose of the interview (selection vs. recruitment), whether an interview is structured or not, and what the interview questions measure.

2 PREVIOUS RESEARCH ON FAKING IN INTERVIEWS

Research has shown that faking can have a positive effect on interview outcomes. Some studies found relationships between self-reported faking behavior and interview outcomes in real selection interviews (e.g., Levashina & Campion, 2007), or between self-reported faking behavior in general and ratings of interviewees' performance in simulated interview settings (Buehl & Melchers, 2017; Ingold et al., 2015; but see Ho et al., 2021). Furthermore, experimental studies found that interviewees in mock interviews were able to obtain higher interview ratings in a condition in which they were instructed to answer as they would in a real selection setting than in a condition in which they were instructed to provide honest answers (Buehl et al., 2019; Van Iddekinge et al., 2005).

In addition to evidence that faking can improve interview performance, previous research also revealed that the use of faking or the intention to fake in interviews is related to personality variables that are also associated with important work outcomes such as task performance, organizational citizenship behavior, or counterproductive work behavior. Specifically, previous research found positive correlations between interview faking or faking intentions and all facets of the Dark Triad (Narcissism, Machiavellianism, and Psychopathy) and negative correlations with Honesty-humility (e.g., Bill et al., 2020; Bourdage et al., 2018; Law et al., 2016; Roulin & Bourdage, 2017; Roulin & Krings, 2016). For the former, meta-analytic research by O'Boyle et al. (2012) revealed negative correlations between traits from the Dark Triad and task performance and positive correlations with counterproductive work behavior. In addition, other primary studies also revealed negative relationships between the Dark Triad and organizational citizenship behavior (e.g., Szabó et al., 2018; Webster & Smith, 2019). Similarly, the meta-analysis by Lee et al. (2019) found that Honesty-humility is strongly negatively related to counterproductive work behavior and slightly positively to task performance and organizational citizenship behavior. Given the evidence that faking can lead to better evaluations in selection interviews and also given that it is related to personality traits that are related to negative work outcomes, it seems possible that faking might indeed impair the criterion-related validity of selection interviews.

In addition to a potentially higher saturation of interview scores with undesirable personality attributes, Salgado (2016) also assumed that faking not only leads to increased mean scores in a selection procedure but also to decreased variation of applicants' scores. He suggested that this leads to a homogenization of scores, which in turn impairs the criterion-related validity.

So far, there are only a few studies that have dealt with the potential influence of faking in selection interviews on criterion-related validity. However, at least partial support for the predicted negative impact of interview faking on interview validity was found. Specifically, Henle et al. (2019) found medium-sized negative correlations between three different types of self-reported interview faking behaviors (inventing, embellishing, and omitting) and self-reported job performance. Additionally, they found large positive correlations between the different faking behaviors and organizational as well as interpersonal deviance. In contrast to this, Ingold et al. (2015) found no significant relationship between self-reported faking behavior in a simulated interview and supervisor ratings of interviewees' job performance. Finally, Buehl et al. (2019) used an interview to predict academic performance and interviewed students in an honest and an applicant condition. They found that interview scores in the applicant condition were a better predictor for students' grade point average than in the honest condition. In contrast, interview scores in the honest condition were a better predictor for university citizenship behavior than in the applicant condition.

Although results concerning the effects of faking on interview validity are mixed, they suggest that faking in interviews can affect criterion-related validity in at least some situations. Therefore, it is of practical relevance to extend knowledge on suitable ways to prevent faking behavior in selection interviews. In addition to face-to-face interviews, technology-mediated interviews are becoming increasingly important due to technological progress and the COVID-19 pandemic (Melchers et al., 2021). These can either be videoconference interviews, in which the interviewer and interviewee communicate via microphone and camera, or asynchronous video interviews, in which the questions are presented on the screen and the interviewees' answers are recorded using a camera and microphone (Lukacik et al., 2022). Previous research revealed differences between face-to-face and technology-mediated interviews concerning interviewees' performance, interviewees' use of impression management (IM), and interviewees' perceptions of and reactions to these interviews.

Concerning interview performance, previous research found that interviewees' performance in videoconference interviews is lower compared to face-to-face interviews (Blacksmith et al., 2016; Melchers et al., 2021). Furthermore, Basch et al. (2021) found that the difference in interview performance ratings was mediated by IM, meaning the lower level of IM in videoconference interviews led to a lower performance. This is also in line with a result by Basch et al. (2020) that the perceived possibility to show IM is lower for videoconference interviews and asynchronous video interviews compared to face-to-face interviews. However, it should be noted that even if the perceived opportunity to use IM (Basch et al., 2020) and the reported IM behavior were lower for videoconference interviews (Basch et al., 2021), the respective mean values turned out to be around the midpoint of the corresponding scale. Therefore, although this previous research has not distinguished between faking and honest IM, we suspect that faking also matters in technology-mediated interviews and thus may be associated with a similar possible threat to criterion-related validity. Furthermore, with regard to applicant reactions, it has been found that technology-mediated interviews are perceived more negatively than face-to-face interviews (Basch et al., 2020; Blacksmith et al., 2016).

3 COUNTERMEASURES AGAINST FAKING

Research has shown that interviewers are hardly able to detect faking in interviews. Roulin et al. (2015), for example, found that interviewers' detection rate of faking tactics was only between 11.8% and 18.5%. Additionally, they only found nonsignificant correlations between interviewees' self-reported and interviewers' perceived use of faking tactics. In line with this, the review by Melchers et al. (2020) pointed out that faking in interviews can currently not be detected. However, there is at least one promising recent approach by Roulin and Powell (2018) that might help to detect faking. Specifically, Roulin and Powell used a criterion-based content analysis (CBCA) adapted from the legal context to detect faking. To do so, the interviewees' answers were transcribed and coded according to 14 CBCA indicators. These indicators were, for example, the quantity of details, logical structure, and contextual embedding. With this approach, it was possible to achieve a faking/honest assessment with an accuracy of up to 63.4%. However, Roulin and Powell (2018) emphasized that this approach is not yet ready for use in organizations. Thus, it is important to consider measures to reduce faking, while it is not yet possible to detect faking when it occurs.

So far, different approaches have been investigated to reduce faking in job interviews but only some of them seem to be promising (see Melchers et al., 2020). These approaches were the use of follow-up questions, different types of interview questions, the number and type of interviewers, and warnings against faking. With regard to follow-up questions, Levashina and Campion (2007, Study 5) investigated whether these questions are suitable to reduce faking in a structured selection interview. However, it turned out that follow-up questions had exactly the opposite effect as expected and led to more faking. Concerning the effect of different question types on faking, Levashina and Campion (2007, Study 5) found that situational questions were more prone to faking than past behavior questions and this effect seemed to be mainly driven by slight image creation and ingratiation. However, in a survey among recent applicants, Bourdage et al. (2018) found that applicants reported less image protection, and less slight and extensive image creation in interviews when situational questions were asked compared to when no such questions were asked. No difference was found for past behavior questions. In addition, Bourdage et al. (2018) found that applicants reported more deceptive ingratiation when resume-based questions were asked compared to when no such questions were asked. Finally, concerning the effect of the number and type of interviewers on faking, Bourdage et al. (2018) also hypothesized that panel interviews or interviews with a future supervisor lead to less faking. However, these hypotheses could not be confirmed. Interviews with a future supervisor even led to more deceptive ingratiation.

Another approach to reduce faking in selection interviews is to use warnings. This approach has repeatedly been used in personality tests and biodata inventories (e.g., Dwight & Donovan, 2003; Kluger & Colella, 1993; McFarland & Ryan, 2006; Schrader & Osburn, 1977). Dwight and Donovan (2003) reviewed 15 studies investigating the impact of warnings on noncognitive selection measures (including personality inventories, biodata inventories, and self-assessments). In 12 of these studies, the samples consisted of real job applicants or military recruits. They found that identification warnings (i.e., warnings about the inclusion of items that are capable of identifying faking) had almost no effect on test scores. However, warnings about consequences (i.e., disciplinary actions) or combination warnings consisting of an identification warning and a warning about consequences had a small negative effect on test scores. In their own experimental study, Dwight and Donovan (2003) examined a student sample and also found that the only warning that led to significantly lower personality scores compared to an unwarned condition was a combination warning. Additionally, they found that an identification warning, a warning about consequences, and a combined warning reduced the ratio of potential fakers within the top ten students. Therefore, Dwight and Donovan (2003) assumed that all three warnings were suitable to reduce faking. In line with this, McFarland and Ryan (2006) told their participants in a laboratory experiment that a social desirability scale was included to identify dishonest individuals in a personality test. They found that faking intentions were less likely and that personality scores were indeed lower on four of the Big Five personality dimensions among participants who received this warning.

In contrast to personality testing, there are only two studies that have examined the impact of warnings on faking or faking intentions in selection interviews so far. In the first of these, Law et al. (2016) compared an identification warning, a moral warning, and a combination warning with an unwarned condition. In the identification warning condition, participants were told that the interview would contain questions suitable for measuring honesty. In the moral warning condition, it was pointed out that honesty was the right thing to do and fair to all applicants. Finally, in the combined warning condition, participants were told that the interview contained questions suitable for measuring honesty and that they would be rewarded (i.e., considered for a $50 prize) if they were honest. Law et al. (2016) found that only the identification warning led to less faking compared to the unwarned condition. However, in a recent experimental vignette study, Bill et al. (2020) found a verification warning had hardly any effect on participants' faking intentions. In this study, participants were told that their answers would be verified via self-named references (e.g., instructors, coworkers, and previous supervisors) and that lying would lead to exclusion from the application process. Interestingly though, Bill et al. (2020) found a small but significant positive effect of a verification warning on intentions to use honest IM. Thus, it might be that participants tried to compensate for the lower possibility to fake by using alternative strategies such as emphasizing their actual qualities more strongly.

Taken together, the available knowledge on potential countermeasures in selection interviews is limited and results are mixed even for the few countermeasures that have already been investigated. Therefore, more information on potential countermeasures to reduce faking in interviews is needed.

4 EFFECTS OF COUNTERMEASURES ON APPLICANT REACTIONS

Interviewers often not only want to select but also recruit applicants (Dipboye et al., 2012). Therefore, it is also important to know whether potential countermeasures against faking provoke negative applicant reactions that impair the recruiting function of interviews. According to the model by Gilliland (1993), there are ten different rules that determine whether a selection system is perceived as procedurally fair or unfair. This perception in turn leads to different applicant reactions. The ten procedural rules are related to three general aspects of selection systems: formal characteristics, explanations, and interpersonal treatment. They concern aspects such as the opportunity to perform (i.e., the possibility for applicants to demonstrate their qualifications), information about the selection process, honesty towards applicants, the interpersonal effectiveness of the administrators of the selection process, and the propriety of the questions asked during the selection process.

To ensure that countermeasures do not lead to negative applicant reactions, they should not violate the different justice rules. However, a possible problem with warnings and other countermeasures is that they might violate some of the procedural rules from Gilliland's (1993) fairness model such as applicants' perceived opportunity to perform. In line with this, McFarland (2003) argued that if organizations take precautions against faking, applicants may feel restricted in the way they present themselves. Furthermore, it might also be possible that warnings violate the interpersonal effectiveness rule. The violation of this rule depends on the extent to which applicants perceive their treatment as respectful and warm (Gilliland, 1993). When applicants are warned about faking countermeasures, they might take this as a suspicion that they would otherwise fake, which could be perceived as disrespectful.

Besides these theoretical assumptions, there is at least some empirical evidence concerning effects of warnings on procedural fairness and applicant reactions. The majority of this research again stems from the field of personality testing. Some of these studies found no relationships between identification warnings and overall fairness perceptions, procedural fairness perceptions, recommendation intentions, or intentions to take a job offer (Lammers, 2017; McFarland, 2003). However, some studies found that negatively framed identification warnings (i.e., dishonest answering will invalidate the test results) impaired perceived interpersonal justice, reapplication intentions, perceptions of the organization's benevolence, and other applicant reaction variables and these warnings also led to higher test anxiety in personality tests and other assessment procedures (Burns et al., 2015; Lammers, 2017). In contrast, however, if they were positively framed (e.g., honesty qualifies for a $50 prize) these warnings could also lead to a higher motivation in assessment procedures (Burns et al., 2015). Finally, in the only study that examined the effect of warnings in selection interviews on applicant reactions, Law et al. (2016) found that none of their three warnings (an identification warning, a moral warning, and a combination warning) affected performance anxiety and procedural justice. Thus, taken together, the results are mixed and partially depend on the framing of the warnings.

5 THE PRESENT RESEARCH

As described above, there has been relatively little research on measures that can be applied to reduce faking or faking intentions in interviews. The few promising findings to date suggest that certain question types may be more or less susceptible to faking (Bourdage et al., 2018; Levashina & Campion, 2007) and that warnings may reduce faking in interviews (Bill et al., 2020; Law et al., 2016). With regard to the question type, however, the findings are mixed and, depending on what one wants to know from an applicant, it is not always possible to choose a specific question type. With regard to warnings, the findings are also mixed (cf. Bill et al., 2020) and seem to be impacted by how warnings are designed in terms of content.

To compare different countermeasures with respect to their effectiveness and to identify potentially effective warnings, we conducted an exploratory study as a first step. In addition, we also wanted to look for additional countermeasures, since there are only relatively few different approaches to reduce faking in job interviews to date. Furthermore, given that the use of technology-mediated interviews also allows the potential use of technology-based countermeasures such as the use of algorithms that analyze video or audio recordings (see Lukacik et al., 2022), we decided to cover countermeasures that seem more suitable for technology-mediated interviews, as well as countermeasures that can also be used in face-to-face interviews. We then conducted three experimental studies. The goal of Study 2 was to investigate how effective the most promising countermeasures for technology-mediated interviews are in reducing faking intentions in videoconference interviews in a vignette study. In Study 3, simulated videoconference interviews were then conducted to determine whether these countermeasures have a comparable effect on faking behavior. In Study 4, we then examined the effect of the most promising countermeasures for face-to-face interviews on faking intentions in another vignette study. By doing so, we want to test the following hypotheses and seek an answer to the subsequent research question:

H1: Countermeasures against faking reduce faking intentions and faking behavior in selection interviews.

RQ1: Which countermeasures are most suitable to reduce faking intentions and faking behavior in selection interviews?

In addition to faking (i.e., deceptive IM), we investigated whether the countermeasures identified in the exploratory study also affect honest IM intentions and honest IM. As described above, a verification warning against faking was able to increase honest IM intentions in a previous study (Bill et al., 2020). Even though the effect was very small and this result refers specifically to a verification warning, we suspect that applicants might try to compensate for the limited possibility to fake in general by using more honest IM such as emphasizing their actual qualities more strongly. Therefore, we pose the following hypothesis:

H2: Countermeasures against faking increase honest IM intentions and honest IM behavior in selection interviews.

As noted above, we assume that countermeasures may violate Gilliland's (1993) procedural fairness rules by limiting applicants' perceived opportunity to perform and/or impairing the interpersonal effectiveness rule. Results from the field of personality testing and other assessment procedures also provided partial evidence that countermeasures can impair applicant reactions (Burns et al., 2015; Lammers, 2017). Therefore, the final aim of the present research was to investigate whether the countermeasures from the exploratory study impair applicant reactions. Thus, we investigated potential effects in the three main studies to test the following hypothesis:

H3: Countermeasures against faking negatively affect applicant reactions to selection interviews.

6 STUDY 1

7 STUDY 2

8 STUDY 3

9 STUDY 4

10 GENERAL DISCUSSION

In the present studies, we addressed the limited knowledge on potential countermeasures against faking in selection interviews. We identified different countermeasures against faking in an exploratory study and evaluated the most promising of these countermeasures in three subsequent experimental studies. We also investigated their influence on applicant reactions. Taken together, our studies make several contributions and also suggest valuable next steps for future research.

As a first contribution, based on empirical evidence from Study 1, we provided information about the most suitable countermeasures against faking in interviews. The significant differences between the various possible countermeasures suggest that these countermeasures might differ in their actual effectiveness. Furthermore, the qualitative answers from the open-ended questions in Study 1 also revealed a surprising result that was not predicted by any of the current faking models (e.g., Levashina & Campion, 2006; McFarland & Ryan, 2000, 2006; Roulin, Krings & Binggeli, 2016). However, Ho et al. (2021) found similar results in a recent qualitative study. In our study, nearly 20% of the answers indicated that a personable interviewer would prevent participants from faking in an interview. In the past, the role of the interviewer was mainly investigated with regard to their role for recruitment purposes (e.g., Chapman et al., 2005; Wilhelmy et al., 2016).

Second, in contrast to our expectations, the three subsequent studies revealed that the different countermeasures did not affect faking intentions or self-reported faking behavior at all, despite our focus on the most promising countermeasures from Study 1. For videoconference interviews, neither a warning referring to a criterion-based content analysis nor a lie detecting algorithm focusing on nonverbal and paraverbal cues had any effect on faking intentions in a hypothetical scenario nor on self-reported faking in simulated videoconference interviews. These results are in contrast to the theoretical assumption (Levashina & Campion, 2006) that faking should be reduced by a higher probability of getting caught. However, it is in line with previous studies that examined different countermeasures and situational variables such as a verification warning (Bill et al., 2020), the future supervisor as the interviewer, the use of panel interviews (Bourdage et al., 2018), or of follow-up questions (Levashina & Campion, 2007). These studies also found little evidence that faking can be reduced (Bill et al., 2020; Bourdage et al., 2018; Levashina & Campion, 2007), or detected (Roulin et al., 2015). So far, only Law et al. (2016) found that an identification warning can reduce faking in interviews. In contrast to our Studies 2 and 3, Law et al. (2016) tested this countermeasure in a face-to-face interview and used a different type of warning. Reasons for the different results could be that the perceived possibility to use IM (Basch et al., 2020) and the actual use of IM is lower for videoconference interviews (Basch et al., 2021). Especially in the simulated videoconference interview study (Study 3), the mean values for faking self-reports were very low even in the control condition, which actually left little room to reduce faking any further. It could also be that participants believed more strongly that certain questions could detect faking than a CBCA or a nonverbal and paraverbal algorithm could.

Additionally, we found no impact of either objective questions or a personable interviewer on faking intentions in face-to-face interviews. The result concerning objective questions is contrary to the theoretical assumption by Levashina and Campion (2006) that verifiable questions provide less opportunity to fake. It is also contrary to the finding of less faking in biodata items that are more objective and verifiable (Becker & Colquitt, 1992). However, in a survey, Bourdage et al. (2018) also found that resume-based questions, which are in fact objective and verifiable, did not reduce faking compared to other questions. Possibly, objective questions have to be combined with a verification measure to credibly discern faking. However, even for such a combination, Bill et al. (2020) found virtually no effect on faking intentions in a vignette study when participants could name their own referees. Therefore, objective questions may need to be combined not only with a credible verification source, but also with a source that is not controllable by the applicant, to be effective. Furthermore, the potential effects could be tested in a controlled experimental study with actual interviews given that all the available evidence so far stems from studies using vignettes or retrospective surveys. However, the convergence concerning the results from our Studies 2 and 3 would at least indirectly suggest, that a similar null effect might also be found for objective questions in an actual interview.

Concerning the possible effects of a personable interviewer on faking intentions, an empirical test revealed a null effect. This is in contrast to Study 1 and to the qualitative study by Ho et al. (2021) that found that a friendly interviewer reduced interview anxiety and therefore the probability of faking. A reason why we did not find any effect could be that participants could not perceive any interview anxiety due to the hypothetical scenario and thus an interviewer described as personable had no impact on faking intentions. Another reason could be that personality variables just had a stronger effect on faking intentions in contrast to potential countermeasures. Specifically, participants who were high on the Dark Triad were also more likely to intend to fake. However, the Dark Triad in turn is associated with duplicity and emotional coldness (Paulhus & Williams, 2002), so that individuals high on the Dark Triad may be unimpressed by personableness and may not reward this kind of interviewer behavior with honesty.

As a third contribution, we extended evidence concerning honest IM. However, we found that the different countermeasures did not affect honest IM intentions and self-reported honest IM. For videoconference interviews, we found no impact of warnings referring to a criterion-based content analysis or to a lie detecting algorithm on honest IM intentions and self-reported honest IM behavior. For face-to-face interviews, we also found that objective questions and a personable interviewer had no impact on honest IM intentions. Even though there is evidence that honest IM can be affected by situational factors (e.g., Basch & Brenner, Melchers, et al., 2021; Bourdage et al., 2018), the results concerning the null effects of warnings are nevertheless in line with previous research that also found only slightly higher honest IM intentions in a verification warning condition compared to a no warning condition (Bill et al., 2020).

Fourth, even though the different warnings did not reduce faking intentions, they had negative effects on justice perceptions and applicant reactions in Study 2. This result is in line with the assumption that warnings could violate procedural rules by Gilliland (1993) and with previous research on personality tests and other assessment procedures that also found that warnings can impair applicant reactions (Burns et al., 2015; Lammers, 2017). However, this negative influence was not evident in Study 3, in which we conducted actual interviews. This is consistent with the only other study that examined the impact of warnings in a simulated interview context and also found no impact on procedural justice and, furthermore, on performance anxiety (Law et al., 2016). One possible reason why warnings in our simulated interview did not have an effect on applicant reactions could be that the warnings were perceived as less salient here. Even though the manipulation check in both of our studies clearly showed that the perceived verification was higher in the warning conditions than in the control condition, the effect size in the simulated interview study was much smaller than in the vignette study. This could also be due to the fact that the warning takes up a much smaller proportion of the simulated interview situation than when participants are only supposed to imagine the selection interview. Furthermore, there is also evidence from other studies that differences in applicant perceptions of different interview situations are smaller after participants had completed the actual interview than before (Melchers et al., 2021). Nevertheless, possible negative effects should be taken into account when applying warnings in high-stakes settings because applicant reactions can be considerably more pronounced in field settings than in simulated lab settings (e.g., Truxillo et al., 2009). For face-to-face interviews, we found that objective questions had little impact on applicant reactions and a personable interviewer—obviously—even had a positive effect on applicant reactions. However, the large effect sizes could also be due to the fact that we described the interviewer as very unsympathetic in the low personableness condition. In real interviews, it may well be that interviewers actually behave in this way, but (with the potential exception of stress interviews, e.g., Chen et al., 2019) this is probably relatively uncommon.

CONFLICTS OF INTEREST

The authors declare no conflicts of interest.