Business•Chapter 3•7 min read•Updated October 4, 2026

Human Resource Management — Recruitment and Selection

O
OIYO EditorialContributor
3/10

1. Forecast capabilities and capacity, not just headcount

Recruitment creates an applicant pool; selection decides who receives an offer. Both depend on a prior question: which work must be performed, to what standard, and with which capabilities? A vacancy count alone cannot answer it.

Start with job analysis: identify tasks, the competencies they require, and which requirements must be present on entry rather than developed later. OPM describes this task–competency connection as the basis for assessment and selection design. OPM: job analysis.

Demand forecasts can use workload assumptions, managerial judgment, or statistical projections. Supply forecasts consider existing skills, departures, transfers, promotion, and the time needed to become productive. A historical trend is not a commitment that demand or employee availability will continue unchanged.

A. A workload calculation

The following monthly plan is fictional. Assume 9,600 service cases, an average of 15 minutes of direct work per case, and 120 productive case-work hours per full-time equivalent (FTE) per month. The productive-hours figure already excludes training, meetings, leave, and other non-case work.

QuantityCalculationResult
Required direct-work hours9,600 × 15 ÷ 602,400 hours
Required capacity2,400 ÷ 12020 FTE
Expected internal supply16 − 1.6 + 216.4 FTE
Expected capacity gap20 − 16.43.6 FTE

The supply calculation assumes 16 current FTE, expected departures equivalent to 10% of that capacity, and two incoming transfers available for the whole month. 1.6 is an expected capacity loss, not a claim that a fraction of a person departs. Four additional fully productive full-time hires would cover the forecast gap under these assumptions.

That is not an instruction to hire four people immediately. Recheck demand uncertainty, ramp-up time, service peaks, scheduling, skill mix, and alternative process improvements. Overtime, outsourcing, redeployment, and separation each introduce contractual, legal, cost, and operational constraints; they are not interchangeable buttons for closing a spreadsheet gap.

2. Recruitment quality concerns the pool as well as its size

Internal recruitment may provide evidence of existing performance and a shorter learning path, but an internal move can create a vacancy elsewhere. External recruitment can bring capabilities absent internally, while adding search and onboarding work. Neither source is always cheaper or better.

A realistic job preview clarifies the work and conditions rather than presenting only an attractive employer brand. Distinguish recruitment reach from the number of applicants meeting job-relevant entry requirements. A large pool obtained through vague criteria may mainly increase screening costs.

Suppose a separate fictional campaign produces 200 applications, 40 interviews, 25 offers, and 20 accepted offers. The final selection ratio is 20/200 = 10%; the interview-stage progression is 25/40 = 62.5%; offer acceptance is 20/25 = 80%. These are different denominators. None establishes that the 20 hires will perform well, and rejected candidates cannot automatically be counted as “unqualified.”

3. Reliability, validity, and utility answer different questions

Reliability concerns score consistency under specified measurement conditions. Validity concerns the evidence supporting a score interpretation for an intended use. Utility concerns the value of the resulting decision after costs and constraints.

QuestionEvidence to examineWhat not to infer
Are ratings consistent?Agreement between trained raters, or stability where the construct should remain stableConsistency proves job relevance
Is the assessment job-related?Tasks and competencies represented in the assessmentA realistic-looking exercise represents every important task
Does a score predict a criterion?Relevant performance data, sample, measurement quality, and uncertaintyA correlation is the probability that a hire succeeds
Is the process worth its cost?Decision benefits, applicant burden, administration cost, and alternativesA higher validity coefficient guarantees higher net value

Content-related evidence examines whether the assessment represents relevant work. Criterion-related evidence examines its relationship with a defined outcome; predictive evidence uses subsequently observed outcomes, while concurrent evidence uses measurements gathered in a contemporaneous setting. Construct-related evidence examines whether the intended competency interpretation is supported. These are not interchangeable labels that can be attached without data.

For example, an assumed predictor–performance correlation of 0.45 does not mean “45% hiring accuracy.” In a simple linear regression with one predictor and an intercept, the squared correlation is 0.2025: it describes the fitted proportion of sample outcome variance accounted for under that model, not a success probability or causal effect. Generalization to a new job or applicant pool still needs evidence.

4. Structure the evidence, not just the conversation

OPM’s structured-interview guidance uses predetermined questions and common rating standards, and includes both past-experience and hypothetical-situation questions. OPM: structured interviews. Structure supports comparability; it does not make irrelevant questions useful or establish that an interview is the best assessment for every role.

A behavioral question asks for evidence of a past action. A situational question asks how a candidate would respond to a proposed scenario. The STAR format can organize an answer, but a fluent story is not independent verification that the event occurred or that the same performance will recur.

Here is a fictional rubric for resolving a service escalation:

Score anchorEvidence sought
1 — IncompleteNo clear issue diagnosis or responsibility for follow-up
3 — AdequateIdentifies the issue, proposes an authorized response, and explains follow-up
5 — StrongAlso examines relevant constraints, communicates trade-offs, and checks the outcome

A real rubric must be derived from actual work and piloted. Specify the evidence required rather than rewarding confidence, shared background, or conversational warmth. Train raters and record reasons for ratings. A work sample can complement an interview, but its task coverage, available tools, time limit, and accessibility also affect interpretation.

5. Fairness is not guaranteed by an assessment label

The following legal reference is U.S.-specific, not a worldwide hiring rule. EEOC guidance discusses discriminatory treatment, disparate impact, job-related justification, and alternatives with less adverse impact. It does not make the phrase “structured interview” a universal compliance certificate. EEOC: employment tests and selection procedures.

The guidance also distinguishes ADA hiring stages: before a conditional offer, disability-related inquiries and medical examinations are restricted; post-offer examinations have same-job-category conditions; examinations of current employees have a different job-related/business-necessity standard. Therefore “post-offer, job-related only” is not a complete statement of all stages. Reasonable accommodation and other applicable laws require separate attention.

For an actual decision, confirm coverage, local rules, the intended assessment use, and current requirements. Privacy-sensitive checks and employment-health data are not ordinary evidence to collect casually.

Practice: critique a selection proposal

Problem

A manager proposes interviewing twice as many applicants using the same unstructured questions. The manager says this will double predictive validity, make selection legally safe, and solve the fictional 3.6-FTE capacity gap. What can be established?

Worked answer

Interviewing more people changes the volume of assessment, not automatically the relationship between assessment scores and subsequent performance. The questions, rating standards, job relevance, applicant pool, and observed criterion determine what evidence can support the intended interpretation.

The legal conclusion does not follow from interview volume or format. The capacity forecast also assumes fully productive hires; offers, acceptance, start dates, and ramp-up can prevent four new employees from supplying four FTE of immediate case-work capacity.

A better proposal specifies entry requirements, a job-related assessment and rubric, expected applicant progression, available start dates, training capacity, and the evidence needed to evaluate performance after hiring. It states assumptions and uncertainty rather than turning a hiring target into a promise.

O

OIYO Editorial

Editorial Desk

The OIYO editorial desk researches money, law, lifestyle, and self-understanding topics against primary sources and public statistics. Every piece carries source notes and is reviewed on a regular cycle for accuracy and usefulness.