Does Interview Anxiety Predict Job Performance and Does it Influence the Predictive Validity of Interviews?

Interview anxiety is negatively related to interview performance; however, its relation to job performance is unknown. It could correlate negatively with job performance, and it could moderate the interview performance&#8211;job performance relation. In a sample of applicants for Residence Assistant positions, interview anxiety had near&#8208;zero correlations with job performance, rated by supervisors and supervisees. It moderated the relation between interview performance and supervisor&#8208;rated facilitating peer and team performance, such that interview performance did not predict this job performance component for anxious applicants. The moderation was not found for supervisor&#8208;rated task proficiency, or for supervisee ratings of either job performance component, suggesting that the impact of interview anxiety depends on rater source and which job performance component is rated.


| INTRODUC TI ON
The employment interview is one of the most commonly used employee selection methods (McCarthy & Cheng, 2014;Pulakos, 2005).
Although job applicants tend to prefer interviews to other selection methods (Anderson, Salgado, & Hülsheger, 2010;Hausknecht, Day, & Thomas, 2004), many applicants also report high levels of interview anxiety (McCarthy & Goffin, 2004). Anxiety in a selection context might be problematic. Indeed, interviewers rate anxious applicants more harshly than their less anxious counterparts (Powell, Stanley, & Brown, 2018). Similarly, general anxiety has been found to be negatively related to invitations to second interviews (Cook, Vance, & Spector, 2000), and communication apprehension is negatively related to the number of job offers candidates receive (Ayres & Crosby, 1995). As a result, anxious candidates have a lower likelihood of being hired, given the typical high stakes of a selection context.
Despite work showing that anxious candidates pay a penalty in terms of interview performance, it remains unclear if interview anxiety is actually predictive of future job performance. That is, are candidates who are anxious at the time of the interview lower performers, if hired? Ideally, the factors that influence interview performance also influence job performance (Huffcutt, Van Iddekinge, & Roth, 2011; see also McCarthy et al., 2013). For example, candidates with the required knowledge, skills, and abilities for the job should obtain high interview scores and subsequently perform well once hired because of that knowledge, skills, and abilities. Yet, factors that have little to no impact on job performance may influence interview performance. This could be the case with candidates who experience interview anxiety.
In this paper, we investigate whether and how interview anxiety is related to overall job performance, as well as performance on specific components of the job. Second, we investigate whether interview anxiety influences the relation between interview performance and job performance, such that the interview is less predictive of job performance for anxious candidates.
The dimensions of interview anxiety as conceptualized by McCarthy and Goffin (2004) underscore the interpersonal nature that is characteristic of interviews. As a selection tool, interviews contain a social element requiring an interaction between a job candidate and at least one employer representative. In a short amount of time, the interviewer and job candidate must build a rapport, the candidate must convey competence, interest in, and fit with, the organization, and ensure that the interviewer judges him or her accordingly. As a result, both verbal and nonverbal communication takes place during the interview (Bonaccio, O'Reilly, O'Sullivan, & Chiocchio, 2016). Smiles, eye contact, and handshakes, are examples of nonverbal communication that occur in interviews and that highlight their interpersonal nature. In an interview context, candidates' trembling hands or quivering voices can be evaluated by interviewers as signs of anxiety. Indeed, interviewers' perceptions of both verbal and nonverbal communication influence how they assess a given candidate (DeGroot & Gooty, 2009).
Anxious interviewees typically receive lower interview scores than their less anxious counterparts. A recent meta-analysis found a corrected correlation of −0.19 between overall interview anxiety and interview performance (Powell et al., 2018). Powell et al. proposed a number of reasons why anxiety could decrease interview performance.
First, because interviews are cognitively demanding, anxiety could have its effect through attentional control (Eysenck et al., 2007); anxious individuals may preferentially allocate their attention to threat-related stimuli (e.g., a distracting thought such as worrying what the interviewer is thinking), to the detriment of the task at hand (i.e., formulating interview answers). Additionally, social anxiety can cause individuals to disproportionally focus on themselves, making them appear less warm, likable, and assertive. Social anxiety is also related to lower self-disclosure (Clark & Wells, 1995); in interview contexts, this could lead to less detailed responses. Thus, there are a number of possible mechanisms through which anxiety may interfere with interview performance, and the relation appears to be a moderate negative one (Powell et al., 2018).

| Is Interview anxiety related to job performance?
Although interview anxiety is moderately negatively correlated with performance during the interview, it is unknown whether anxiety experienced during the interview has any implications for subsequent job performance. One way in which interview anxiety could be related to job performance is via an underlying personality trait that affects both feelings of anxiety during interviews, and feelings of anxiety on the job. That is, there may be a dispositional component to interview anxiety that also affects job performance. There is some evidence to support the argument that interview anxiety is indicative of a general behavioral tendency. Dispositions such as general anxiety have been found to be positively related (r =.40) to interview anxiety (McCarthy & Goffin, 2004). Because of the influence of stable dispositions on interview anxiety, anxiety during the interview may capture a behavioral pattern that corresponds to how anxious an individual would be during stressful interpersonal situations on the job. Situations like the interview may capture the level of anxiety that individuals experience when having to perform tasks like building relationships and influencing others. Thus, interview anxiety may be negatively related to job performance, particularly when it requires frequent interactions with others. McCarthy, Trougakos, and Cheng (2016) proposed two mechanisms through which anxiety could influence job performance. For maximal performance, anxiety may lead to cognitive interference, which draws employees' attention away from the task at hand, leading to lower levels of task performance. However, for typical, or sustained, dayto-day job performance, anxiety causes emotional exhaustion but not interference. In some cases, anxiety could even be beneficial. Indeed, Cheng and McCarthy (2018) proposed that, under some circumstances, anxiety may have a facilitative effect on job performance through selfregulatory processing-anxiety can signal that there is a discrepancy between desired and actual goal progress, which can increase motivation to engage in a task. Thus, while there is evidence that interview anxiety is moderately negatively related to interview performance, and there is some evidence workplace anxiety may, in some cases, be negatively related to job performance, it is currently unknown whether anxiety at the stage of the interview is related to subsequent job performance. McCarthy et al. (2013) conducted the only study to date that has investigated the relation between selection test anxiety and job performance. They found that selection test anxiety about a work sample test, situational judgment test, and personality test was unrelated to measures of task and interpersonal job performance for customer service representatives (r = −.03 to −.06) measured concurrently. In contrast, test anxiety about all three assessments was significantly negatively related, although weakly (r = −.09 to −.12) to task and interpersonal performance for product technicians. Therefore, the limited empirical evidence to date has not shown a strong relation between selection test anxiety and job performance, with some evidence of a weak negative correlation. However, McCarthy et al.'s study did not examine interview anxiety and its relation with job performance. Our study fills this gap, and we do so in a predictive rather than concurrent research design.
In addition to investigating the relation between interview anxiety and overall job performance, it is also worth looking at its relation with specific components of performance. If interview anxiety does have an effect on later job performance, it is likely to have a stronger effect on components of job performance when there is conceptual overlap between interview anxiety and job performance. As discussed above, the employment interview is an interpersonal situation, which requires social interaction (McCarthy & Goffin, 2004). The interview is distinct from other types of employment selection tools, such as personality or situational judgement tests, because employment interviews not only assesses knowledge, skills, and abilities, but the candidate has to manage the interpersonal situation of the interview at the same time as demonstrating competence. Of note, interviewers routinely assess social skills in interviews (Huffcutt, Conway, Roth, & Stone, 2001). Similarly, some aspects of job performance contain interpersonal requirements, whether these interactions take place between colleagues at the same level, managers and employees, or employees and customers. When thinking about the relation between interview anxiety and job performance, interview anxiety may be more likely to be related to job performance if the job performance ratings measure managing interpersonal situations at work. After all, both the interview and the job context would call upon similar social and interpersonal skills on the part of an employee. For this reason, anxiety about the interview situation could be related to later job performance components that are also interpersonal in nature.
In their model of job performance, Campbell, McCloy, Oppler, and Sager (1993) proposed eight latent performance components that summarize the performance requirements of all jobs. They argued that not every job has all eight components as performance requirements, but that for any single job, a subset of these components is sufficient for describing its performance requirements.
Based on the Campbell model, we focused on two components of job performance: (a) job-specific task proficiency and (b) facilitating peer and team performance. Job specific task proficiency refers to the core technical tasks that are central to a job. Facilitating peer and team performance is defined as the degree to which an employee supports their peers, provides help to solve problems, and facilitates group functioning by being a good model. Following McCarthy et al. (2013), using a task and an interpersonal component of job performance allows us to test the idea that interview anxiety might be more predictive of components with which it shares a conceptual match. Based on the conceptual match between the interview situation and the more interpersonal aspect of facilitating peer and team performance, we hypothesized that: Hypothesis 1 There will be a negative relation between interview anxiety and job performance.

Hypothesis 2 The relation between interview anxiety and job per-
formance will be stronger for the "facilitating peer and team performance" compared to the "job specific task proficiency" component of job performance.

| Interview anxiety as a moderator between interview performance and job performance
The preceding section positioned interview anxiety as a predictor of job performance. That is, our hypothesis was based on the assumption that there might be some shared variance between anxiety regarding the interview (which is an interpersonal situation), and subsequent interpersonal performance on the job. In contrast, because interview anxiety is a situation-specific trait (McCarthy & Goffin, 2004), it may only be relevant in the interview setting, and may not be indicative of job performance at all. For example, anxious candidates may receive lower ratings if they have difficulty articulating a coherent answer in the interview but may communicate clearly once hired. If anxiety is indeed situation specific, interview anxiety could be source of error, which would attenuate the predictive validity of interview performance. Thus, interview performance will under-predict job performance ratings for anxious candidates. Indeed, in an education setting, Bonaccio, Reeve, and Winford (2012) found that the predictive validity of a test (i.e., the correlation between a measure of cognitive ability and final examination grades, which took place 2 months later) decreased slightly for candidates higher in the "worry" component of test anxiety. It could be the case that, in a similar way, interview anxiety may moderate the relation between interview performance and job performance. That is, if interview anxiety reduces interview performance, but is unrelated to job performance, then it could reduce the predictive validity of the interview. McCarthy et al. (2013) proposed that this type of moderation was more likely if the test reaction is situational (e.g., if anxiety is about the interview specifically). Therefore, we propose the following hypothesis: Hypothesis 3 Interview anxiety will moderate the relation between interview performance and job performance, such that the interview-job performance relation will be weaker for candidates with high anxiety than for those with low anxiety.
We test this hypothesis with both components of job performance: job specific task proficiency and facilitating peer and team performance.

| The current study
We aim to understand the relation between interview anxiety, interview performance, and job performance, and we do so in real, high-stakes interviews, for university residence assistants.
Residence assistants oversee students who live in university residence. Importantly, this job allows us to look at both the task and the interpersonal components of job performance discussed above. Resident Assistants must exhibit job-specific task proficiency through job duties such as knowledge of rules, written communication, and enforcing community standards. As well, a critical component of their job involves facilitating peer and team performance through job duties such as advising students, community building, and conflict resolution. Furthermore, in typical selection contexts, the ratio of interviewed to hired job candidates is low. This ratio implies that candidates who receive even moderate scores on the interview are not hired; consequently, no job performance data is available for them. Our study takes place in a selection context where the ratio of interviewed to hired candidates is high, which allows for a stronger test of the relation between interview anxiety and job performance.

| Participants
Candidates for Residence Assistant positions at a Canadian univer-

| Interview Anxiety
Interview anxiety was assessed following the interview using

| Interview performance
Interviewers consisted of one professional staff member and an experienced student staff member (typically 2-3 years on staff).
The professional staff member received interview training from the University's Human Resources department. The student staff members received a minimum of 1 hr of interview training. Training included a review of the questions and rating scales, followed by a role-play and practice marking, and a discussion on rating consistency. The structured interviews consisted of six behavioral descriptive interview questions with standardized probes. The questions assessed conflict resolution, flexibility, self-management, critical thinking and problem solving, leadership, and crisis management. The interviewers rated each of the six structured interview questions out of a 3-point scale, and also gave the candidate an overall suitability score. Interview questions were the same across the three hiring cycles but the numerical response scale used varied. Consequently, to combine data from the three hiring cycles, we converted all scores on a common 25-point rating scale, which had been used by the organization in the first hiring cycle.
Because interviews were conducted by two interviewers, we calculated the intraclass correlations (ICC; McGraw & Wong, 1996).
We calculated the ICC for each of the hiring cycles separately. ICC (two-way random effects, absolute agreement) values ranged from 0.90 to 0.93 for the three hiring cycles. Based on the high level of agreement between the interviewers, their scores were averaged to form the overall interview score.

| Job Performance
Job performance was assessed by two sources. These supervisors are all professional staff members, with 1-5 years of experience in the position. Prior to completing the assessments, these supervisors spent 2-3 hr as a group reviewing the performance rating tool and discussing feedback best practices. Second, the students who were supervised by the residence assistant (an average of 17 students per Residence Assistant) also provided ratings. Supervisors of, and students supervised by, the Residence Assistant used standardized job performance appraisal forms developed by university administrators for this position. All feedback was shared with the residence assistants as part of their performance appraisal.
The performance appraisal forms were designed by the organization, with no input from the research team. The language used in the job performance appraisal forms differed between those filled out by supervisors and by students, reflecting the different perspectives of the evaluation. Like for the interviews, the forms also changed slightly across data collection cycles, as the organization updated its performance evaluations practices. Importantly, while the specific questions were worded differently, the constructs assessed by the performance evaluation questions were similar across the years and the rater sources. Specifically, to test our hypotheses, we had to recover the performance dimensions of "job specific task proficiency" and "facilitating peer and team performance" from the forms employed by the organization. The second and third authors independently sorted each job performance item into either of these categories based on a conceptual match between the Campbell et al. (1993) definitions of these job components and the detailed definitions and behavioral examples provided for each job performance dimensions in the rating forms.
We began with the performance questions rated by supervisors.
Sample items are "Administrative: job responsibilities (particularly documentation) are completed on time" and "Communication and interpersonal skills: Has effectively navigated potentially confrontational situations with students, both in their community and/or oncall, attending to relationships" for "job specific task proficiency," and "facilitating peer and team performance," respectively. We repeated this process for each of the three data collection cycles. A research assistant blind to the hypotheses and the purpose of the sorting calculate the agreement on the sorting. Initial agreement ranged from 96% to 100% per dimension. Disagreement was resolved through discussion until 100% agreement was reached. We combined the data across the data collection years to generate one variable reflecting "job specific task proficiency" and one reflecting "facilitating peer and team performance" as evaluated by supervisors.
We repeated the process with the student ratings. Sample items are "My RA is knowledgeable about resources and activities on campus and shares at our meetings" and "My RA provides alternative words and/or phrases when challenging inappropriate and non-inclusive language" for "job specific task proficiency," and "facilitating peer and team performance," respectively. Unlike the supervisor evaluation forms, the student forms also contained items unrelated to RA job performance, such as "I have studied with other resident students on at least one occasion." Items unrelated to RA job performance were coded as "neither." Initial agreement on sorting the items ranged from 89% to 100% per dimension and disagreements were resolved through discussion.

| RE SULTS
Interview ratings were positively correlated with supervisor ratings (r = .20, p = .01) of overall job performance, providing evidence for the predictive validity of the interview, as shown in Table 1.
Consistent with past research, interview anxiety and interview performance were negatively related (r = −.11, p = .046), indicating that more anxious interviewees received lower scores on the interview. Two components of anxiety that were most predictive of interview performance were communication anxiety (r = −.15, p = .01) and social anxiety (r = −.16, p = .01).
Turning to the relation between interview anxiety and job performance, interview anxiety was not significantly related to overall job performance as rated by the supervisor (r = .07, p = .41); thus, Hypothesis 1 was not supported. Anxiety had near-zero correlations with both supervisor-rated job-specific task proficiency (r =.04, p = .48), and with supervisor-rated facilitating peer and team performance (r =.05, p = .50). Hypothesis 2, that interview anxiety would be more strongly related to the social aspect of job performance, was not supported; anxiety had near-zero correlations with both components of performance, as rated by supervisors. When looking at the separate dimensions of interview anxiety, the correlations between anxiety and job performance ranged from −.04 (social anxiety with both types of job performance) to .12 (behavioral anxiety with facilitating peer/team performance). Thus, the correlations between the dimensions of anxiety and the components of job performance ranged from small to moderate in magnitude (c.f., Bosco, Aguinis, Singh, Field, & Pierce, 2015).

TA B L E 1 Descriptive statistics and correlations among study variables
Next, we considered student ratings of job performance.
Interview anxiety was not related to overall student-rated job performance (r = .05, p = .58); thus, Hypothesis 1 was not supported with student ratings. Anxiety had near-zero correlations with both student-rated job-specific task proficiency (r =.06, p = .42), and with student-rated facilitating peer and team performance (r =.02, p = .76). Hypothesis 2, that interview anxiety would be more strongly related to the social aspect of job performance was not supported, as anxiety had near-zero correlations with both components of performance, as rated by students. When looking at the separate dimensions of interview anxiety, the correlations between anxiety and job performance ranged from −.04 (behavioral anxiety with facilitating peer/team performance) to .07 (communication and performance anxiety with task proficiency). Overall, the correlations between the dimensions of anxiety and the components of student-rated job performance ranged from small negative to small positive in magnitude (c.f., Bosco et al., 2015).
Our third hypothesis required assessing whether the predictive validity of the interview is moderated by interview anxiety. For the moderation hypothesis, we focused on overall interview anxiety and whether it moderates the relation between interview performance and job performance. We used Fast Interaction (Stanley, 2014) to test our hypothesis. Fast Interaction is a free software program that "conducts regression analyses to determine if two continuous variables interact to predict a criterion variable; it automatically mean-centers the predictor variables and creates the corresponding product term predictors" (http://www.stats canbe fun.com/fasti ntera ction ).
First, we tested the interaction with overall supervisor-rated job performance as the dependent variable. The interaction between interview score and interview anxiety was significant, (β = −.20, t = −2.60, p = .01). We did a simple slopes analysis to investigate the interaction. At one SD above the mean in interview anxiety (i.e., high anxiety), interview performance did not significantly predict supervisor-rated job performance (t = −0.14, p = .89). In contrast, at one SD below the mean in interview anxiety (i.e., low anxiety), interview score was a significant predictor of supervisor-rated job performance (t = 3.64, p < .001). The interaction is plotted in Figure 1.
We now turn to the moderation analyses considering the conceptual match between job performance components with interview anxiety. The interaction between interview score and interview anxiety was significant, β = −0.23, t = −2.97, p = .003, when we looked at facilitating peer and team performance, a performance dimension which has a conceptual match with interview anxiety. We did a simple slopes analysis to investigate the interaction. At one SD above the mean, interview performance did not significantly predict supervisor-rated job performance (t = 0.137, p = .89). In contrast, at one SD below the mean on anxiety, interview score was a significant predictor of supervisor-rated job performance (t = 4.48, p < .001).
The interaction is plotted in Figure 2.
The pattern of results was different when we considered the aspect of job performance that was not a conceptual match with interview anxiety. The interaction between interview score and interview anxiety was not significant, β = −0.12, t = −1.48, p = .14 when we looked at the job-specific task proficiency job performance rating.
We did a simple slopes analysis to investigate the interaction. At one SD above the mean in anxiety, interview performance did not significantly predict supervisor-rated job performance (t = −0.33, p = .74). As well, at one SD below the mean on anxiety, interview score was not a significant predictor of supervisor-rated job performance (t = 1.81, p = .07). The interaction is plotted in Figure 3.
When we considered student ratings of job performance as the dependent variable, none of the interaction terms were significant.
Specifically, for overall student-rated performance, β = −.03, t = −0.35, p = .73, for student-rated facilitating peer and team performance, β = −.009 t = −0.12, p = .90, and for student-rated job-specific task proficiency, β = −.05, t = −0.58, p = .56, the interaction terms were not significant. The interview scores were not a significant predictor of either dimension of student-rated job performance, for either high or low anxious candidates.

| Exploratory analyses with the dimensions of interview anxiety
We conducted exploratory analyses looking at the different dimensions of interview anxiety as a moderator of the interview performance-job performance relation. We focused our analyses on job-specific task proficiency and facilitating peer and team performance to gain a more granular understanding of the relations.
When considering supervisor-rated task proficiency, none of the dimensions of anxiety were a significant moderator (β values ranged F I G U R E 1 Supervisor-rated overall job performance as a function of interview performance and interview anxiety

| D ISCUSS I ON
Our study investigated whether anxiety in a high-stakes employment interview might predict subsequent job performance for university residence assistants. Consistent with previous research (McCarthy & Goffin, 2004), interview anxiety was rated around the midpoint of the 5-point rating scale, and was negatively related to interview performance. In particular, communication anxiety (which reflects apprehension about verbal and nonverbal communication skills, and listening skills) and social anxiety (which reflects apprehension about one's social behavior) were more strongly negatively related to interview performance relative to the other dimensions. In contrast, concerns about one's physical appearance (e.g., worry about dressing appropriately), and about the physiological symptoms (e.g., shaky hands) appeared to have less of an effect on interview performance.
Although overall interview anxiety was negatively related to interview ratings, neither overall interview anxiety, nor its dimensions were related to subsequent job performance, either as rated by supervisors, or by the residence assistants' "clients," that is, the students they serve. Furthermore, interview anxiety did not predict narrower performance components, of job-specific task proficiency, or facilitating peer and team performance across job performance rater types. These findings suggest that interviewees who are anxious are not necessarily those who will display low job performance once they are on the job. These findings are consistent with McCarthy et al.'s (2016) contention that anxiety may not be a damaging factor for day-to-day typical job performance. The results further indicate that for jobs like those of residence assistants, which involve facilitation of young adults in individual and group settings in addition to administrative duties (such as community center staff, or recreational service workers), it would appear that anxiety during the job interview should not factor into a decision to hire or not hire a candidate.
Not only was anxiety not significantly correlated with later job performance, but for highly anxious candidates (above 1 SD), interview performance did not predict job performance. These results suggest that interview anxiety may decrease the effectiveness of the interview as a predictive tool. A similar finding was reported in the context of test anxiety such that cognitive ability tests were less useful predictors of exam performance for high anxiety test takers (Bonaccio et al., 2012).
It is interesting to note that when looking at anxiety as a moderator of the interview performance-job performance relation for supervisor ratings, it became apparent that the validity of interview ratings seems to be entirely due to the facilitating peer/team component of job performance. The interview was actually not a valid predictor of the task-specific job proficiency component. This finding may be explained by the very nature of interviews: interviews are very interpersonal, and so may be particularly effective for predicting the interactive aspects of jobs. Indeed, Huffcutt et al. (2001) note that over 60% of characteristics rated during interviews are personality traits and social skills. Perhaps other tools, such as work sample tests, would be better suited for assessing potential at less social tasks like completing documentation on time.
The interview also was not a significant predictor of job performance as rated by the residence assistants' "clients," that is, the students living in residence. This finding is consistent with Borman's (1997)  about what effective performance is, and also supervisors would have access to different samples of the incumbents' job performance. In the case of the Residence Assistants, the interviewers were full-time residence staff members, who would be quite similar to the staff members evaluating the Resident Assistants' performance. However, the students making the performance ratings would be less like the original interviewers. This is the case for both what the students might think of as ideal job performance (e.g., residents might prefer leniency on, rather than strict adherence to, rules) and also in terms of the kinds of work performance they might have access to (e.g., students might see the Residence Assistant at their worst, because they live together).

| Strengths, limitations, and future directions
We conducted our study in a group of residence assistants at one university. Our sample therefore was composed of relatively young workers in one occupation and employer. In order to further understand the implications of interview anxiety for future job performance, it will be important to investigate this question with a more diverse sample both with respect to demographics and job type. For example, because older adults experience more difficulty in finding a new position following job loss (Wanberg, Kanfer, Hamann, & Zhang, 2016), it would be interesting to investigate the issue of interview anxiety in mid-to-late career job candidates. Although these candidates have a wealth of work experience to draw upon in an interview, the interview itself may feel unsettling to them given the lack of recent "practice." Our sample had notable strengths. First, our study was conducted in a real employee selection context that still had a high selection ratio. This ensured that we had job performance data for candidates who may have otherwise been rejected in other selection contexts. Second, applicants who were hired did not have, on average, lower anxiety scores compared to people who were not hired, had similar standard deviations, and comparable ranges. From a study design perspective, this similarity between hired and not-hired candidates is a strength, especially as it ensures similar distributions of scores on the main variable of interest, interview anxiety.
Future research should continue to investigate whether interview anxiety is indeed a situation-specific disposition, as originally described by McCarthy and Goffin (2004), or whether there could be aspects of interview anxiety that vary across interview settings.
For example, if an interviewer does things to put the candidate at ease, could this action reduce the situational interview anxiety, and subsequently improve the predictive validity of the interview? In this study, we adapted the wording of the questionnaire items on the original MASI, which referred to interviews in general (e.g., "I become so apprehensive in job interviews that I am unable to express my thoughts clearly"), so that in our study, they refer specifically to the interview the person just had (e.g., "I became so apprehensive in the interview that I was unable to express my thoughts clearly").
Thus, in this study we do assess anxiety that is specific to a particular interview, rather than a general tendency. This may be an important distinction, such that we may have found different relations with performance if we had asked about interview anxiety as a disposition, rather than specific to this interview. Still, the relation between interview anxiety and interview performance is weaker when considering anxiety on interviews in general (vs. anxiety on that specific interview; c.f. Powell et al., 2018). As such, it is unlikely that our conclusions would have changed.
In summary, we encourage hiring managers to take steps to reduce job candidates' interview anxiety. For example, interviewers can put candidates at ease with friendly conversations prior to the beginning of the structured interview, and give candidates time to answer questions rather than adding time pressure to the interview.
If organizations reject suitable candidates based on anxiety, they risk incurring both missed opportunity costs, and costs related to hiring the wrong candidate.