PSYCH 639 Week 3 Ethics in Testing and Selection Example

An AI Video Interview and a Look at Applicants' Social Media: Ethical Questions Before a Grocery Chain Changes How It Hires Managers

[Student Name]

University of Phoenix

PSYCH/639: Ethics and Professional Skills in I-O Psychology

Week 3 Assignment

[Instructor Name]

[Date]

The grocery chain, the vendor offer and the pilot are composites written for a model paper; research and professional guidance come from the sources listed.

New technology has reached hiring faster than evidence about it. I-O psychologists are often asked to approve tools whose claims are impressive and whose evidence is thin. This paper evaluates two practices proposed at my company against ethical standards for assessment and research on validity, fairness and applicant reactions.

The Proposal

Desert Fresh Markets hires about sixty assistant store managers a year. Our recruiting team is overwhelmed with applications and proposed two changes. First, a vendor offers an asynchronous video interview: applicants record answers to six questions, and software analyzes their word choice, vocal tone and facial expressions to produce a "leadership potential" score, ranking applicants for recruiters. Second, recruiters want to look at applicants' public social media profiles to "check for red flags and culture fit." The recruiting director asked me whether we should do either.

Standards for Assessment

Section 9 of the APA code tells psychologists to use a test only for purposes research backs, to pick tools shown to work for the kind of people being tested and to get consent where it is required. The code also asks psychologists who offer automated scoring to retain responsibility for the appropriate application and interpretation of the instruments. The joint testing standards make the same point from another angle: a score is only valid for some stated purpose, and the employer using it, not just the seller, has to check the evidence for that purpose.

The AI Video Interview

I asked the vendor for its technical manual. It provided a white paper describing a correlation between its scores and supervisor ratings in a sample of retail managers at an unnamed company, with no information on how ratings were collected, no breakdown by demographic group and no explanation of which features drive scores, citing trade secrecy.

Tippins et al. (2021) issued a call to action on AI-based personnel selection tools, arguing that many tools have been adopted with little published validity evidence and that their complexity makes it hard to know what they measure. They raised concerns about training data that reflect historical bias, about features such as facial expressions or speech patterns that may relate to race, disability, age or accent rather than job performance, about lack of transparency to applicants and employers and about privacy. They called for vendors and users to provide the same kinds of validity, fairness and documentation evidence expected of traditional assessments.

Facial expression analysis raises particular concerns. Expressions vary across cultures and are affected by disabilities, such as facial paralysis or autism, and by camera quality and lighting. A store manager's job requires coaching, scheduling and handling problems, not producing particular facial movements on camera.

A score built from smiles and vocal tone, with no published evidence, cannot be defended to an applicant who asks why she was screened out.

Social Media Screening

One study (Van Iddekinge et al., 2016) tested whether recruiters' ratings of job applicants' Facebook profiles predicted later job performance and turnover. Recruiters rated profiles on dimensions such as personality traits and overall hireability. The ratings were generally unrelated to supervisors' later performance ratings and to turnover, and they showed some evidence of favoring white and female applicants. The authors cautioned organizations against using social media in hiring without evidence of validity.

Informal social media review also exposes recruiters to information they should not use, such as religion, pregnancy, disability, age and political views, creating risk of discrimination whether intended or not. It is inconsistent, since some applicants have public profiles and others do not, and "culture fit" judgments can mask bias.

How Applicants See It

Hausknecht et al. (2004) meta-analyzed research on applicant reactions to selection procedures. Applicants rated interviews, work samples, résumés and references more favorably than personality tests, honesty tests and graphology. Perceptions of fairness and job relatedness were associated with intentions to pursue the job, accept offers and recommend the employer. For a grocery chain whose applicants are also customers, a process that feels opaque or invasive can damage the brand.

Why the Pressure to Adopt Is Strong

It helps to understand why such tools are attractive. Recruiters at our company screen hundreds of applications for each opening, and a ranked list promises relief. Vendors present dashboards and polished case studies, and competitors' use of similar tools creates a fear of falling behind. None of these pressures is unethical in itself; they are reasons the recruiting team deserves a workable alternative, not just a refusal. An ethical review that ends only with "no" leaves the real problem in place and invites someone else to approve the tool without review.

What Evidence Would Be Enough

If the vendor returns, I will ask for specific evidence before any pilot: a description of the training data and how it was collected, validity studies in jobs similar to assistant store manager with defined performance criteria, results broken down by sex, race and ethnicity, age and disability status, an explanation of which features drive scores and an independent audit. I will also ask whether the facial analysis component can be switched off. These requests mirror what we would expect from any traditional test publisher.

Competence and Responsibility

If I approved the video tool, I would be responsible for its use without being able to explain what it measures. The code does not allow psychologists to hand that responsibility to a vendor. I would also need to explain to applicants how their recordings are stored and used, which the vendor's privacy terms did not make clear.

Recommendations

I recommended against the facial and vocal analysis and against informal social media screening. To address the real problem, too many applications, I proposed a structured, transparent process. Applicants complete a short online questionnaire on job-related experience, scored with a key developed from our job analysis. Those who pass record answers to structured video questions, which trained recruiters score with anchored rating scales, not software. Applicants are told in advance how each step is used. If the vendor can later provide independent validity and fairness evidence for a version of its tool that excludes facial analysis, we can consider a local pilot comparing its scores with recruiter ratings and later performance.

Fairness to Applicants With Disabilities

Any video-based step must also allow accommodations. An applicant who is deaf, who stutters or who has limited mobility should be able to request an alternative, such as a written response or a live interview, without penalty, and the invitation should say so plainly.

Applicant Feedback and Data Retention

Applicants will receive a clear explanation of each step and a contact for questions. Video recordings will be deleted after the hiring decision plus the legally required retention period.

Conclusion

New selection tools must meet the same ethical and scientific standards as old ones. The AI video interview lacks evidence and risks bias; social media screening has shown little validity and exposes recruiters to protected information. A structured, transparent process meets the business need while respecting applicants and the standards of the profession.

References

Hausknecht, J. P., Day, D. V., & Thomas, S. C. (2004). Applicant reactions to selection procedures: An updated model and meta-analysis. Personnel Psychology, 57(3), 639-683. https://doi.org/10.1111/j.1744-6570.2004.00003.x

Tippins, N. T., Oswald, F. L., & McPhail, S. M. (2021). Scientific, legal, and ethical concerns about AI-based personnel selection tools: A call to action. Personnel Assessment and Decisions, 7(2), Article 1. https://doi.org/10.25035/pad.2021.02.001

Van Iddekinge, C. H., Lanivich, S. E., Roth, P. L., & Junco, E. (2016). Social media for selection? Validity and adverse impact potential of a Facebook-based assessment. Journal of Management, 42(7), 1811-1835. https://doi.org/10.1177/0149206313515524