Applied Psychology Hrm Cascio Aguinis

1. 1Anchor appraisal in a behavioral basis: define performance in terms of observable job behaviors and results, not personality traits or self-concept. Build criteria that pass the evaluation tests before you rate anyone against them.
2. 2Test each criterion for relevance, sensitivity/discriminability, and practicality; then check it against deficiency (missing important dimensions) and contamination (measuring things outside actual performance) to keep the measure procedurally defensible.
3. 3Choose your rating source deliberately by matching source to what they can actually observe: immediate supervisor, peers, subordinates, self, or clients served. For team tasks, evaluate at the team level rather than forcing individual scores onto interdependent work.
4. 4Select a rating system on purpose. Absolute systems (rate against a standard) support development and interpersonal fairness; relative systems (employee comparisons) force distinctions but strain interpersonal justice. Decide before rating, not after.
5. 5Train raters to counter the named judgmental biases: leniency/severity, central tendency, and halo. Rater training is a fundamental requirement, not optional polish.
6. 6Manage the social and interpersonal context: judge your own performance first, encourage subordinate preparation, use priming information, warm up and encourage participation, be an active listener, and be specific.
7. 7In the feedback interview, judge performance not personality, avoid destructive criticism and threats to the employee's ego, set mutually agreeable and formal goals, then continue to communicate and assess progress toward goals regularly.
8. 8Close the loop on distributive fairness: make organizational rewards contingent on performance, and communicate frequently so the process is continuous rather than an annual event.

Consider a self-managing work team of the kind described in the source (autonomous work groups whose members manage their own task and interpersonal processes). Because the work is interdependent, you evaluate performance at the team-task level rather than assigning individual ranks that would misattribute shared outcomes. You define criteria behaviorally and check them: a criterion capturing only individual output would be deficient (it omits the collaborative problem-solving the team is authorized to do) and potentially contaminated. You pull ratings from multiple sources — peers who see day-to-day contribution, clients served who see outputs, and self-assessment — then check agreement and equivalence across sources before combining. You train raters against halo and central tendency so strong output on one dimension doesn't inflate every score. In the feedback session you judge performance not personality, set mutually agreeable formal goals with the team, and tie rewards to those goals so the process is felt as fair both procedurally and interpersonally.

• If the task is interdependent or team-based, then evaluate at the team level; if the task is individual, then rate individuals against an absolute standard.
• If your purpose is development and interpersonal fairness, then use an absolute rating system; if you must force distinctions among employees, then use a relative system but expect fairness strain.
• If a criterion omits important performance dimensions, then it is deficient; if it captures factors outside the person's actual performance, then it is contaminated — fix both before rating.
• If a rater cannot directly observe the relevant behavior, then do not use that source for that dimension; match source to observability (supervisor, peer, subordinate, self, client).
• If ratings across sources disagree beyond expected equivalence, then investigate bias and observation gaps before averaging or acting on the scores.

• Rating personality or self-concept instead of behavior. Tell: feedback language describes who the person 'is' rather than specific observable actions.
• Uncorrected judgmental biases — leniency/severity, central tendency, halo. Tell: score distributions cluster at the top or the middle, or one dimension predicts all others.
• Skipping rater training. Tell: raters interpret the same scale differently and scores are not equivalent across raters.
• Forcing individual scores onto interdependent team work. Tell: team members dispute attribution and cite shared responsibility for outcomes.
• Destructive criticism and ego threats in the interview, with rewards disconnected from performance. Tell: employees leave defensive, goals are vague, and reward decisions feel arbitrary.

• ▸Strategic alignment: state how this role's performance links to the organization's value-creation story.
• ▸Behavioral criteria: list observable behaviors/results to be judged (no traits).
• ▸Deficiency/contamination check: note any missing dimension and any outside factor to remove.
• ▸Rating source(s): name who rates each dimension and confirm they can observe it.
• ▸Task type: mark individual vs. team; set evaluation level accordingly.
• ▸Rating system: choose absolute or relative and state the purpose driving the choice.
• ▸Rater training/bias controls: list steps taken against leniency, central tendency, halo.
• ▸Feedback plan: preparation prompts, priming info, and specific/active-listening commitments.
• ▸Goals & rewards: record mutually agreed formal goals, review cadence, and reward contingency.

1. 1Map every employment decision (recruitment, initial screening, selection, performance management, organizational exit) against the specific statutes that govern it: Title VII of the Civil Rights Act of 1964 (race, color, religion, sex, national origin), the Age Discrimination in Employment Act of 1967, the ADA of 1990, the Civil Rights Act of 1991, the Equal Pay Act of 1963, FMLA of 1993, and USERRA of 1994.
2. 2For each selection procedure, run it against both the Scientific Guidelines checklist (sources of validity evidence, generalizing validity, fairness and bias, operational considerations) and the Legal Guidelines checklist (adverse impact, validation, criterion-related validity, content validity, construct validity, validity generalization, application).
3. 3Screen pre-employment inquiries and testing practices for hidden adverse impact — the unfair advantage or disadvantage given to members of one group relative to others that produces a denial or restriction of employment.
4. 4Build a sound internal complaint system so problems are resolved before they escalate to formal legal proceedings through state fair employment practice commissions, the EEOC, the OFCCP, or federal and state district courts.
5. 5Where you use supervisory ratings for performance decisions, ensure the instruments are well designed, free of bias/error, focused on job-relevant elements, and that raters are trained and cooperative — because merit-based rewards are ethically and morally dependent on each individual getting a fair shake.
6. 6Protect employee privacy through fair information practices, disciplined handling of employee searches and other workplace investigations, and honoring obligations to those who are evaluated.
7. 7Establish a corporate ethics program that codifies obligations to one's profession, to those evaluated, and to employers, and account for individual-difference antecedents of ethical behavior.
8. 8When adverse impact appears in selection, apply documented mitigation — including forms of test-score banding evaluated from legal, technical, and societal points of view — and re-run differential prediction tests to check accuracy.

A firm using supervisory ratings to allocate merit pay finds that its rating instrument focuses on elements irrelevant to effective job performance and that raters are untrained. Under the text's standard, this means the ideal of fairness will never be realized, because a merit-based reward system is ethically and morally dependent on each individual getting a fair shake. The firm redesigns the instrument to target job-relevant strengths and weaknesses and trains its raters, minimizing bias and error. It then treats performance management as a daily feedback process, not a once- or twice-a-year appraisal, closing the feedback loop back to job analysis, recruitment, selection, and training. When a rated employee alleges unfair treatment, the firm routes the complaint through its internal complaint system — the simplest, least costly alternative — rather than forcing the employee toward the EEOC or district court, where solutions become time-consuming and expensive. This preventive posture reflects the text's advice to deal with problems before they escalate to formal legal proceedings.

• If a selection procedure produces adverse impact, then it must satisfy the Legal Guidelines validation requirements (criterion-related, content, or construct validity) or be modified/replaced.
• If a complaint arises and no internal mechanism exists (or employees are unaware of it), then the complainant is pushed toward costly formal legal routes — so build and publicize an internal complaint system first.
• If supervisory ratings drive merit or reward decisions, then the instrument must be validly designed and the raters trained, or fairness cannot be achieved.
• If a job requirement is challenged, then justify it as a bona fide occupational qualification or defensible business necessity rather than assumption.
• If you must reduce adverse impact, then evaluate test-score banding from legal, technical, and societal points of view before adopting it.

• Treating performance appraisal (done once or twice a year) as equivalent to performance management (daily feedback) — the tell is a rating cycle with no timely feedback loop back into selection and training.
• Relying on poorly designed rating instruments or untrained, uncooperative raters — the tell is ratings focused on elements irrelevant or unimportant to effective job performance.
• Assuming no adverse impact because no discrimination is intended — the tell is unexamined pre-employment inquiries and tests that quietly restrict one group's employment.
• Having no internal complaint mechanism, or one employees don't know about — the tell is disputes going straight to fair employment practice commissions, the EEOC/OFCCP, or the courts.
• Ignoring privacy and fair information practices in workplace investigations and searches — the tell is data handling that breaches obligations to those who are evaluated.

• ▸Decision point: which employment-process stage (recruitment/screening/selection/performance management/exit)
• ▸Governing statutes: list applicable laws (Title VII, ADEA, ADA, CRA 1991, Equal Pay Act, FMLA, USERRA)
• ▸Adverse impact check: does the procedure disadvantage a protected group? (yes/no + evidence)
• ▸Validity basis: criterion-related / content / construct evidence on file (cite)
• ▸Fairness and bias review: differential prediction test run and result
• ▸Rater/instrument quality: instrument job-relevant and raters trained? (yes/no)
• ▸Privacy safeguards: fair information practices applied to any searches/data (yes/no)
• ▸Internal complaint route: mechanism exists and is communicated to employees (yes/no)
• ▸Mitigation taken: banding or other adverse-impact reduction, with legal/technical/societal note

1. 1Reverse the conventional sequence: fix values and people practices first, let strategy follow. Per O'Reilly & Pfeffer, treat strategic decisions as secondary to living a set of values and creating alignment between values and people — build distinguishing capabilities, then choose how and where to compete.
2. 2Make strategic planning democratic rather than staff-siloed. Instead of relegating planning to a separate staff, include line managers, customers, and suppliers so top managers can hear conditions and shift plans midstream.
3. 3Establish the HR-to-business-strategy linkage explicitly for workforce planning: map how strategic business plans drive strategic workforce plans, then forecast supply and demand against them, integrate the two forecasts, and match forecast results to action plans.
4. 4Run the three-step training alignment routine (Tannenbaum): (1) for any important change or initiative, identify the new capabilities needed, compare them to current capabilities, and specify steps to bridge the gap; (2) periodically identify key strategic capabilities the organization will need going forward; (3) compare current programs and services against those strategic needs.
5. 5Embed selection utility in strategic context. Recognize validity and SDy are not constant over time — as strategic objectives (survival, market-share growth) vary, so must the alignment of labor, capital, and equipment; re-evaluate a selection system's contribution when strategic goals change.
6. 6Answer the Russell et al. questions when justifying a system: what performance level a selectee generates (delta-U per selectee), the gain from a new system for a single cohort, and whether expected performance with/without the system is adequate to meet the firm's strategic needs across existing and new cohorts.
7. 7Set control and evaluation loops: monitor performance against plan, take corrective action, and treat the whole thing as an ongoing evaluation process (as done in workforce planning at IBM) rather than a one-time forecast.

Cisco Systems shows alignment where strategy comes last. Cisco is not wedded to any particular technology because it recognizes customers are the arbiters of choice; it listens carefully to customers, then offers the solutions they want. When satisfying customer demand requires capabilities it lacks, it acquires companies to get the technology — more than 80 acquisitions in the last decade. Its productive employees let it save on capital and labor costs while delivering outstanding service, and it retains intellectual capital as the industry shifts around it. The competitive strategy exists and is well-developed, but it is secondary to the values-people alignment that produces distinguishing capabilities. That mindset lets Cisco move in whatever direction the market takes rather than locking into one vision of the future.

• If you are choosing between building capabilities and locking in a strategy, build the values-people alignment first — strategy is the secondary, downstream decision.
• If the future changes quickly and rigid plans risk locking you into a vision that may not come to pass, then make planning democratic (line managers, customers, suppliers) and prepare to shift midstream.
• If strategic goals change over time, then re-assess the relative contribution of your selection system rather than assuming its validity and SDy are constant.
• If a new selection system produces high dollar utility but adverse impact, then use multiattribute analysis and consult constituents (top management, HR, in-house counsel) before deciding — dollar value alone is not the strategic verdict.
• If training does not translate into effective behaviors or organizational results, then check for lack of alignment with strategic direction before blaming program content.

• Treating strategy as the starting point instead of the outcome — the tell is that HR practices are designed to serve a fixed strategic vision that the market has already outrun.
• Relegating strategic planning to a separate staff — the tell is line managers, customers, and suppliers were never consulted and top management cannot explain shifting conditions.
• Judging a selection system by its utility dollar value alone — the tell is a large SDy-based number quoted with no acknowledgment that validity and economic value vary over time or that adverse impact affects constituents differently.
• Running training disconnected from broader organizational systems — the tell is programs continue unchanged while strategic capability needs shift, and no gap analysis between needed and current capabilities exists.
• Confusing performance appraisal (once or twice a year) with performance management (daily improvement) — the tell is alignment is 'checked' annually and never reinforced against organizational objectives day to day.

• ▸Distinguishing capability: what people/practices set us apart before strategy is chosen
• ▸Competitive strategy: how and where we compete (secondary decision)
• ▸Planning participants: line managers, customers, suppliers included (Y/N + names)
• ▸Needed capabilities vs. current capabilities: the gap and steps to bridge it
• ▸Future strategic capabilities: what the organization will need going forward
• ▸HR system under review (selection/training/workforce plan) and its link to the business strategy
• ▸Utility in strategic context: delta-U per selectee / per cohort / adequacy vs. strategic needs
• ▸Control loop: performance monitored against plan + corrective action owner
• ▸Adverse-impact / constituent check: top management, HR, counsel sign-off

1. 1Conduct an organization analysis before any training rollout: the passages treat 'Organizational Characteristics Related to Effective Training' as a distinct determinant, so assess whether leadership, climate, and reward systems will actually support skill application, not just whether the content is sound.
2. 2Build transfer of training into the design phase, not the follow-up. Pair it with self-management to maintain behavior change and adaptive guidance so learners know how to sustain new skills once back in the daily work context.
3. 3Install reinforcement as a standing mechanism. The passages list reinforcement and knowledge of results (feedback) as principles that enhance learning — schedule concrete feedback loops and rewards tied to using the new behavior on the job.
4. 4Make organizational rewards contingent on performance. This is stated directly in the appraisal interview guidance: without the reward-behavior link, the environment does not encourage application.
5. 5Use the appraisal/goal-setting interview as the reinforcement channel: communicate frequently, set mutually agreeable and formal goals, and continue to communicate and assess progress toward goals regularly.
6. 6Route peer support through team training and supportive-relationship behaviors. Teamwork in the passages means sharing information, developing supportive relationships with colleagues, and creating team spirit — design work so these are structurally possible.
7. 7Measure whether the environment worked. Use the outcome-measurement criteria (behavior-level and results-level) and, where possible, a classical experimental or quasi-experimental design to separate real behavior change from noise.
8. 8Guard against destructive feedback: in every review, judge performance not personality or self-concept, be specific, be an active listener, and avoid threats to the employee's ego — hostile feedback suppresses application of learned skills.

A firm runs a cross-cultural training program and wants the skills used on assignment, not forgotten. Person analysis shows openness predicts instructor ratings of cross-cultural performance (corrected correlation .33) while cognitive ability does not, so the environment must actively reinforce openness-driven behaviors rather than assume smart hires will transfer skills. Managers are trained in appraisal (get training in appraisal) and hold frequent goal-setting interviews where they set mutually agreeable, formal goals about applying cross-cultural skills. Because agreeableness correlated negatively (-.26) — too-agreeable people are seen as pushovers in some cultures — supervisors are coached to reinforce assertive-but-culturally-aware behavior rather than penalize it. Rewards are made contingent on demonstrated on-the-job cross-cultural competence, and progress is reviewed regularly. Outcomes are then measured at the behavior level, not just satisfaction, using a quasi-experimental design to confirm the supportive environment produced transfer.

• If skills fade after training, then check the organization analysis first — the failure is usually environmental (no reinforcement, no reward contingency), not a training-content defect.
• If a behavior you want applied is not tied to a reward or feedback loop, then it will not persist — make organizational rewards contingent on performance.
• If feedback risks becoming personal, then judge performance not personality, be specific, and avoid destructive criticism, or you suppress the very behavior you trained.
• If a valued behavior is culturally bound (e.g., assertiveness vs. agreeableness), then define what supportive means locally before rewarding it, because a universally 'positive' trait can read as weakness in some cultures.
• If you cannot show behavior-level change with a controlled or quasi-experimental design, then do not claim the environment supported transfer — you have only measured reactions.

• Treating training as complete at delivery. Tell: no self-management, adaptive guidance, or reinforcement plan exists, and no one owns post-training follow-up.
• Rewarding attendance or personality instead of applied performance. Tell: reviews praise 'attitude' but rewards are not contingent on the specific learned behavior.
• Infrequent, one-shot appraisals. Tell: feedback happens annually, violating 'communicate frequently' and 'assess progress toward goals regularly.'
• Destructive feedback that threatens the ego. Tell: reviews cite personality flaws, are vague, and the ratee stops trying the new skill.
• Assuming a trait is universally desirable. Tell: agreeableness or another 'nice' behavior is rewarded even where, per the data, it undermines perceived professional competence (-.26).

• ▸Target behavior — the specific learned skill to be applied on the job
• ▸Organization analysis note — is climate/leadership ready to support it? yes/no + gap
• ▸Reward contingency — which organizational reward is tied to this behavior
• ▸Feedback cadence — how often knowledge of results is delivered (frequent, dated)
• ▸Peer/team support mechanism — how colleagues share info and reinforce it
• ▸Goal statement — mutually agreed, formal goal set with the employee
• ▸Cultural check — is this behavior culturally bound? adjust definition if yes
• ▸Progress review dates — scheduled points to assess movement toward goal
• ▸Outcome measure — behavior-level criterion + design used to verify transfer

1. 1Start with job analysis — the fundamental building block on which all later employment decisions rest. Produce a detailed specification of the work to be performed, the skills needed, and the training required for satisfactory performance.
2. 2Translate the job analysis into a KSAO profile: itemize the knowledge, skills, abilities, and other characteristics the job demands, distinguishing minimum qualifications from preferred ones using established methods for setting minimum qualifications.
3. 3Assess each candidate's aptitudes, abilities, personality, and interests, then profile those characteristics against the job's KSAO requirements — the Utopian aim being to place individuals in jobs suited to them and to society.
4. 4Separate ability from motivation when evaluating fit: performance differences on a job may stem from ability, from motivation, or from both, so assess the direction and strength of forces impelling effort, not just capability.
5. 5Select measures deliberately — follow the steps for selecting and creating tests and classify each measure, then confirm reliability (consistency) before relying on it for a fit judgment.
6. 6Validate that predictors actually forecast job performance using criterion-related evidence (predictive or concurrent studies) or content-related evidence, and cross-validate to guard against capitalizing on chance.
7. 7Re-run the job analysis whenever the work changes — mechanization or automation alters requirements, which reverberates through wages, workforce planning, recruiting, selection, and training.

The passages give the mismatch case directly: history Ph.D.s driving taxicabs for lack of professional work, and young people full of enthusiasm, drive, and intelligence placed in monotonous, routine, dead-end jobs. In fit terms, the taxi-driving Ph.D. holds far more knowledge and ability than the job's KSAO requirements demand — a mismatch of overqualification driven by external supply, not by a bad selection decision. The enthusiastic young worker in a dead-end job shows the opposite failure: high motivation and ability squandered against requirements that neither use nor develop them. In both, the individual's profiled characteristics and the job's requirements are misaligned, so society fails to make maximal use of its most precious resource. The corrective is the idealized sequence — assess aptitudes, abilities, personality, and interests; profile them; then match to the job whose requirements they fit. Doing this well requires an accurate up-front job analysis so the requirements side of the equation is real, not assumed.

• If performance is deficient, then diagnose whether the gap is ability (a true fit problem) or motivation (an effort problem) before concluding the person-job match is wrong.
• If the job's work content changes through mechanization or automation, then update the job analysis first and expect downstream effects on wages, workforce planning, recruiting, selection, and training.
• If a measure used to judge fit is not demonstrably reliable, then do not use it — validity is bounded by reliability.
• If local validation of a fit predictor is not feasible, then gather evidence through synthetic validity, test transportability, or validity generalization rather than assuming the predictor works.
• If a candidate is qualified for a job, then that individual has a fundamental right to compete for it regardless of race, age, gender, disability, religion, or national origin.

• Skipping or under-specifying the job analysis — the tell is a KSAO list that reads generic and cannot be traced to the work actually performed.
• Confounding ability with motivation, so an unmotivated but capable person is judged a poor fit (or vice versa) — the tell is a fit conclusion drawn without separating effort from capability.
• Treating requirements as fixed after mechanization or automation changes the job — the tell is stale job descriptions still driving recruiting and pay while the work has moved on.
• Relying on a predictor that is consistent but unvalidated, or validated only in one setting without cross-validation — the tell is fit decisions defended by 'it's always worked here' with no criterion evidence.
• Optimizing only the individual's talents and ignoring aggregate outcomes — the tell is visible mismatches (overqualified people in dead-end roles) accepted as normal rather than corrected.

• ▸Job requirements source — cite the job analysis producing the work specification, skills, and training needed
• ▸Required KSAOs — list knowledge, skills, abilities, and other characteristics the job demands
• ▸Minimum vs. preferred — mark which requirements are minimum qualifications and which are preferred
• ▸Candidate profile — record assessed aptitudes, abilities, personality, and interests
• ▸Ability match — note where candidate capability meets, exceeds, or falls short of each requirement
• ▸Motivation check — estimate strength of forces impelling effort on this job's tasks/goals
• ▸Predictor & evidence — name the measure used and its reliability plus validity evidence (predictive, concurrent, or content)
• ▸Change flag — note any mechanization/automation or content change requiring job-analysis update
• ▸Fit verdict & reverberations — state match/mismatch and downstream effects on pay, recruiting, selection, training

1. 1Separate what you measure from why: use achievement tests to assess present competence (effects of standardized learning like an apprenticeship or a programming course) and aptitude tests to forecast whether an employee will profit from training or succeed in a new situation. Neither distinction applies rigidly since both reflect prior learning.
2. 2Distinguish tests from inventories before you interpret scores. Tests (cognitive: aptitude, achievement) measure what the person knows or can do; inventories (affective: interests, values, motives, attitudes, temperament) measure what the person says he or she feels. Do not treat a self-report of interpersonal competence as demonstrated competence.
3. 3For any competence rated by multiple sources (self, peers, supervisors, subordinates, clients), establish measurement equivalence FIRST, before assuming ratings are directly comparable. Use confirmatory factor analysis or item response theory to test invariance across raters.
4. 4When equivalence holds, assess agreement across raters using a hybrid multitrait-multirater analysis; when it fails, treat the sources as measuring different things rather than averaging them.
5. 5Match the source to the construct: expect equivalence for behaviorally based, developmental ratings (e.g., team-building skills, job knowledge, productivity, quality, leadership) but expect NON-equivalence for broader dimensions collected for research (interpersonal competence, administrative competence, compliance/acceptance of authority).
6. 6For competencies only observable in face-to-face interaction (appearance, speech, poise, interpersonal competence, empathy, personal initiative), use a well-designed interview, which can add prediction of job performance over and above cognitive ability, Conscientiousness, and experience.
7. 7When using recommendations/reference checks to confirm competence, verify the four preconditions (adequate observation opportunity, rater competence, candor, interpretable expression) before weighting them, given their average validity of only .14.
8. 8Train raters and require frequent, specific feedback: judge performance not personality, use priming information, set mutually agreeable formal goals, and make organizational rewards contingent on demonstrated performance.

You are validating a 360 measure of managerial competence. Following Maurer, Raju, & Collins (1998), a team-building skills measure showed measurement equivalence across peers and subordinates, and Facteau & Craig (2001) found equivalence across self, peers, supervisors, and subordinates for 48 behaviorally oriented items covering 10 dimensions of managerial performance. So you can compare and aggregate those ratings. But Viswesvaran, Schmidt, & Ones (2002) found lack of invariance for interpersonal competence, administrative competence, and compliance/acceptance of authority between supervisors and peers, even though overall job performance, productivity, effort, job knowledge, quality, and leadership WERE equivalent. The lesson: you aggregate the behaviorally specific dimensions across sources, but you report interpersonal and administrative competence separately by source rather than blending them. One plausible reason is that behaviorally based, developmental ratings are more likely equivalent than broad dimensions collected for research.

• If you need present competence at end of training, then use an achievement test; if you need to forecast future success or trainability, then use an aptitude test.
• If the instrument asks 'I think / I feel', then it is an inventory reporting felt states, not demonstrated competence—do not score it as ability.
• If ratings come from multiple sources, then test measurement equivalence before comparing; if invariance fails, then report by source rather than averaging.
• If the construct is broad (interpersonal or administrative competence, compliance) and collected for research, then expect and check for non-equivalence; if it is behaviorally specific and developmental, then equivalence is more likely.
• If a competency can only be observed in face-to-face interaction (poise, empathy, initiative), then assess it via a structured interview; if it is documented ability or history, then use records/references and confirm the four preconditions.

• Averaging peer and supervisor ratings of interpersonal competence—the tell: you assumed comparability without testing invariance, and Viswesvaran et al. found exactly this measure lacks equivalence.
• Treating an inventory self-report as proven competence—the tell: the item reads 'I feel...' yet you file it as an ability score, feeding the popular distrust of testing that stems from confusing tests and inventories.
• Relying on a reference check to confirm ability without checking preconditions—the tell: validity near .14 and the recommender either lacked observation opportunity or was unwilling to be candid (the most serious deficiency).
• Applying the aptitude/achievement split rigidly—the tell: you ignore that both measure current behavior shaped by prior learning and mis-select a test for the wrong purpose.
• Letting rating judgments drift into leniency, central tendency, or halo—the tell: distributions are compressed or one impression colors every dimension, so demonstrated competence looks uniform when it is not.

• ▸Competency name: state the specific KSAO being measured
• ▸Measure type: test (aptitude/achievement) or inventory (self-report)
• ▸Purpose: present competence vs. forecast/development vs. research
• ▸Rating sources: list who rates (self/peers/supervisor/subordinates/clients)
• ▸Behavioral specificity: rate item wording from broad-dimension to concrete-behavioral
• ▸Equivalence method: CFA or IRT to test invariance across sources
• ▸Equivalence result: equivalent / non-equivalent (aggregate vs. report-by-source)
• ▸Corroborating evidence: reference-check precondition check and interview-observed traits
• ▸Bias watch: note any leniency, central tendency, or halo flagged in this measure

1. 1Separate the two drivers of performance before you intervene: observe whether a performance gap traces to ability or to motivation, because energizing effort is useless if the person cannot do the task. Watch a group doing identical work and note who works faster, makes fewer errors, and appears to enjoy it more.
2. 2Set mutually agreeable, formal goals during the appraisal interview rather than imposing them. Documented goals direct effort toward specific organizational objectives and give both parties a shared reference for later progress checks.
3. 3Make organizational rewards contingent on performance. Where effort and results are not tied to reward, the force impelling effort toward the goal weakens; wire the reward system so that measured performance changes what the employee receives.
4. 4Deliver feedback (knowledge of results) frequently, not once a year. Continuous communication about progress toward goals maintains effort between formal reviews and keeps direction from drifting.
5. 5Judge performance, not personality or self-concept, and avoid destructive criticism and threats to the employee's ego. Ego threat redirects energy into defense rather than into the task; be specific about behaviors and outcomes instead.
6. 6In the feedback session, warm up and encourage participation, use priming information, and be an active listener. Subordinate involvement in defining goals raises commitment to those goals.
7. 7In training contexts, build in reinforcement, practice, goal-setting, and adaptive guidance so that motivation to learn transfers into maintained behavior change on the job.
8. 8Account for the new psychological contract: with lifelong employment gone and job mobility now the norm, expect lower default commitment and intentions to stay, and design motivation levers that do not assume paternalistic loyalty.

Take the passage's group of carpenters building cabinets. You observe that some work faster, make fewer mistakes, and enjoy the work more than others, and you ask the diagnostic question: why? You first rule out or confirm ability differences — some carpenters may be stronger, have keener eyesight, or better motor coordination. For those whose output lags despite adequate ability, the cause is motivation: the strength of forces impelling them to put effort into the task varies. For a capable but under-performing carpenter, you hold a goal-setting interview, judge the work not the person, set a mutually agreed cabinet-quality-and-pace goal, and tie a reward to hitting it. You then communicate progress regularly rather than waiting for an annual review, sustaining the effort you energized. The observed 'work differences' now reflect corrected 'people differences' in motivation, not misdiagnosed ability.

• If a capable worker underperforms, then treat it as a motivation problem and use goals plus contingent rewards — not more training.
• If the worker lacks the ability the task requires, then no amount of motivation fixes it; place or train for capability first (the Utopian-ideal matching of talents to jobs).
• If you must give corrective feedback, then judge performance and be specific; if criticism would threaten the ego, then reframe it, because ego threat drains effort from the task.
• If a goal is set, then make it mutually agreeable and formal and attach a contingent reward, or effort will not reliably follow.
• If the employee operates under the new psychological contract (expects temporary tenure), then do not rely on loyalty-based commitment to sustain effort — build explicit, near-term reward contingencies.

• Misattributing a motivation gap to ability (or vice versa). Tell: you keep sending a capable person to remedial training and performance still doesn't move.
• Rewarding tenure or presence instead of results. Tell: high and low performers receive the same rewards, and effort flattens toward the middle.
• Annual-only feedback. Tell: goals set in the review are forgotten by mid-year and effort drifts off-target with no correction.
• Attacking personality or self-concept in feedback. Tell: the employee becomes defensive, disengages, and satisfaction and commitment drop.
• Imposing goals unilaterally. Tell: employees comply minimally, show no ownership, and effort collapses once supervision relaxes.

• ▸Employee & task: name the person and the specific job/task observed.
• ▸Ability check: list evidence the person can do the task (skills, prior output).
• ▸Motivation check: describe observed effort intensity, direction, and persistence.
• ▸Diagnosis: mark whether the gap is ability, motivation, or both.
• ▸Mutually agreed goal: state the specific, formal goal both parties accepted.
• ▸Reward contingency: name the organizational reward tied to reaching the goal.
• ▸Feedback cadence: set how often progress will be communicated.
• ▸Feedback discipline: note how you will judge performance (not personality) and stay specific.
• ▸Review date & outcome: log the check-in date and progress toward the goal.

1. 1Separate the two fairness channels before designing any HR decision: procedural fairness (are the process and criteria just?) and distributive fairness (is the outcome equitable?). Employees judge both independently, so track them separately.
2. 2Give ratees a documented voice in the performance management system. The Cawley, Keeping & Levy (1998) meta-analysis of 27 studies (32 samples) found employee participation correlated .61 (corrected) with positive reactions — build in ratee self-evaluations, an active role in the appraisal interview, and input into system design.
3. 3Specify performance criteria explicitly using critical incidents before any assessment is collected. This is the direct lever for cutting the perception of friendship bias in peer assessment, which is negatively related to user acceptance (McEvoy & Buller, 1987).
4. 4Sequence feedback ahead of administrative decisions: use peer or performance results in a joint employee-supervisor review of progress first, then make administrative decisions later — so the process is seen as informing, not ambushing.
5. 5Frame assessments as developmental where possible; developmental use is positively related to user acceptance (McEvoy & Buller, 1987), which protects perceived fairness even when scores are unfavorable.
6. 6Audit HR practices against the legal fairness framework (Chapter 8): adverse impact, differential validity, differential prediction. Where adverse impact appears, apply mitigations such as test-score banding before decisions harden into litigation.
7. 7Anticipate the downstream consequences of unfairness perceptions (per the Chapter 5 update): reduced acceptance of feedback, and — under the broken psychological contract — lower satisfaction, commitment, intent to stay, and trust in the organization (Lester et al., 2003).

Consider negative peer-rating feedback. In a controlled study (DeNisi, Randolph & Blencoe, 1983), telling a group they had been rated poorly by peers produced significantly lower perceived group performance, plus lower cohesiveness, satisfaction, and lower peer ratings on the next task; positive feedback produced only nonsignificant gains. The fairness problem is that recipients attribute the poor rating to friendship bias rather than to real performance. The operational fix is to define the criteria with critical incidents up front so raters and ratees see the same behavioral basis, then route the results into a joint employee-supervisor progress review before any administrative decision. This converts a raw, disputable score into a procedurally grounded, developmental conversation — raising both perceived utility and perceived fairness (Cawley et al., 1998).

• If ratees have no voice in evaluation, interview, or design, then expect low acceptance and add participation before rolling out — the .61 correlation is the reason.
• If peer or subordinate assessments will be used, then anchor them in explicit critical-incident criteria; otherwise friendship-bias perceptions will suppress acceptance.
• If an assessment result is unfavorable, then deliver it developmentally and via a joint review before any administrative action, not as a standalone verdict.
• If a selection or appraisal practice shows adverse impact, then test for differential prediction/validity and apply mitigation (e.g., test-score banding) rather than defending the raw practice.
• If employees perceive procedures as unjust, then treat downstream drops in commitment, trust, and intent to stay as expected costs, not surprises (Lester et al., 2003).

• Treating a fair outcome as sufficient — the tell is employees accepting the result but resenting how it was reached; you optimized distributive fairness and ignored procedural fairness.
• Collecting peer ratings without published criteria — the tell is recipients blaming 'who likes whom' rather than their behavior; friendship-bias perception erodes acceptance.
• Using assessment scores straight into administrative decisions with no prior developmental review — the tell is defensiveness and drops in group cohesiveness and satisfaction after feedback.
• Assuming widespread agreement on fairness ends means agreement on means — polls show majorities support EEO yet disagree sharply on methods (Von Drehle, 2003); the tell is a technically defensible practice that still triggers grievances.
• Ignoring adverse impact until it becomes litigation — the tell is a legally vulnerable practice that no one flagged because outcomes 'looked fine' in aggregate.

• ▸Decision under review: name the HR action (rating, selection, promotion, layoff).
• ▸Procedural fairness check: list the criteria used and whether they are documented via critical incidents.
• ▸Ratee voice: state where ratees provide self-evaluation, interview input, and design input.
• ▸Distributive fairness check: describe the outcome and the equity standard it's judged against.
• ▸Feedback sequencing: confirm developmental/joint review occurs before administrative action (yes/no + when).
• ▸Bias exposure: note friendship-bias risk (peer sources) and any adverse-impact/differential-prediction findings.
• ▸Mitigation applied: record fixes used (criteria specification, banding, developmental framing).
• ▸Expected consequences: predict effects on acceptance, commitment, trust, intent to stay if unfair.
• ▸Owner and review date: assign accountability and next audit date.

1. 1Define the construct as behavior, not outcome: list the observable things the person does that are relevant to organizational goals, then explicitly separate those from results (effectiveness). Sales volume is an outcome; making calls, controlling expenses, and handling reports are performance behaviors.
2. 2Map the three behavioral families for the role: task performance (core job-defined behaviors), contextual performance (helping, cooperation, supporting morale), and counterproductive behaviors (disciplinary transgressions, property damage, personal aggression, political deviance, substance abuse).
3. 3Distinguish the ultimate criterion (the full conceptual domain of success over the person's entire tenure) from the measurable domains you can actually sample. Accept that any measure is a segment of the universe, not the whole.
4. 4Scale each behavior by the level of performance it represents, so raters can place observed actions on a continuum rather than a present/absent judgment.
5. 5Locate each behavior on the observation-to-abstraction continuum. Directly observable behaviors (typing, sales calls) support content-based judgment by incumbents and supervisors; abstract qualities (reasoning) require construct-based inference.
6. 6Address the three criterion challenges head-on: unreliability of performance over time, difficulty of observation, and multidimensionality. Decide whether you will report a composite criterion (single index) or multiple criteria (separate dimensions).
7. 7Screen your criterion set for deficiency (behaviors you failed to capture) and contamination (variance from things other than the person's behavior, e.g., group membership, rater bias).

Take the salesperson example from Campbell. The ultimate criterion of performance includes total sales volume over the entire tenure, total new accounts brought in during the career, customer loyalty built up, influence on the morale and sales records of other salespeople, and effectiveness in planning calls, controlling expenses, and handling reports and records. Notice that sales volume and new-account counts are outcomes (effectiveness), while planning calls, controlling expenses, and completing reports are the observable behaviors that constitute performance. Because this ultimate criterion is strictly conceptual and cannot be measured, you operationally define a job performance domain that is only a segment of the universe. A typing test for a secretary is the parallel case: it captures one domain of a multi-domain job. If you scored only sales volume, your criterion would be deficient (missing the morale-influence and planning behaviors) and contaminated (sales volume depends on territory and economy, not just the person).

• If you are counting a result (volume, accounts, revenue), then classify it as effectiveness, not performance, and keep it separate from behavioral criteria.
• If the behavior sits at the observation end of the continuum (work products like typing), then use content-related evidence and incumbent/supervisor judgment; if it is highly abstract (reasoning), then use construct-related evidence.
• If a role has several performance domains but you can only measure one, then flag the criterion as deficient and document what is left uncaptured.
• If a behavior falls in the counterproductive family (disciplinary transgressions, aggression, substance abuse, political deviance), then measure it as a distinct dimension, not by subtracting from task scores.
• If observed score variance comes from anything other than the person's behavior (group membership, predictor knowledge, rater lenience), then treat it as contamination and control for it before interpreting.

• Confusing outcomes with behaviors: rating a salesperson only on volume. Tell — the criterion moves with the economy or territory, not the person's actions.
• Collapsing a multidimensional job into one score. Tell — high performers on the measured domain still fail on unmeasured domains like helping or planning.
• Ignoring the counterproductive dimension. Tell — the appraisal has no way to record property damage, aggression, or substance abuse, so they surface only in exit interviews.
• Applying content-related evidence to an abstract construct. Tell — you claim a reasoning measure is 'obviously job-relevant' but cannot show the inferential link.
• Treating one measurement as the whole construct. Tell — you speak of the measure as 'the performance,' forgetting it is a sample of behavior standing in for an unobservable universe.

• ▸Role and goals: which organizational goals this job's behaviors serve.
• ▸Task performance behaviors: core job-defined actions, one line each.
• ▸Contextual behaviors: helping/cooperation/morale actions relevant to goals.
• ▸Counterproductive behaviors: transgressions, aggression, deviance, substance abuse to be tracked.
• ▸Outcome vs behavior check: list any results included and confirm they are labeled effectiveness, not performance.
• ▸Observation-to-abstraction placement: mark each behavior observable / reportable / abstract.
• ▸Evidence strategy: content-related for observable, construct-related for abstract.
• ▸Deficiency note: performance domains in the universe you are NOT measuring.
• ▸Contamination note: non-behavioral variance sources (group, rater, territory) to control.

1. 1Frame the selection or HR decision in decision-theory terms first: treat every hire as a bet whose payoff depends on how well the predictor forecasts job performance and how much variability in performance exists to be captured.
2. 2Establish the three parameters that drive utility: the selection ratio (proportion of applicants you can afford to hire), the base rate (proportion of current employees judged successful without the new procedure), and the validity coefficient of your predictor against a job-performance criterion.
3. 3Choose a utility model appropriate to your question — use the Naylor-Shine model when you want to show expected improvement in criterion score per selectee, and the Brogden-Cronbach-Gleser (BCG) model when you must express payoff in dollars.
4. 4For the BCG dollar estimate, gather the standard deviation of job performance in dollars (SDy), the validity coefficient, the average predictor score of those selected (in standard-score units), the number selected, and the cost of the selection procedure per applicant.
5. 5Compute net utility as the gain from improved selection minus the total cost of testing all applicants, then scrutinize the estimate — question every input, especially SDy and validity, before presenting it to decision-makers.
6. 6Translate the result into the language of business: present net dollar payoff, and connect it to the strategic context (workforce planning targets, the productivity ratios and demand forecasts the organization already uses).

A firm screening applicants with a valid test wants to know whether it pays off. It sets the selection ratio low (hiring the top applicants), notes its current base rate of successful performers, and applies a predictor with a known validity coefficient against the job-performance criterion. Using the Brogden-Cronbach-Gleser model, it multiplies the number selected by the validity, by the standard deviation of performance in dollars (SDy), and by the average standardized test score of those hired, then subtracts the total cost of testing all applicants. The result is a net dollar payoff figure — the workforce utility of adopting the procedure. Because the estimate hinges on SDy and validity, the firm scrutinizes both inputs before reporting. This reframes selection from a compliance activity into a quantified economic contribution, exactly the shift Chapter 13 emphasizes when it states that the utility of a selection system speaks the language of business.

• If you need to persuade executives, then express utility in dollars via the BCG model rather than in criterion-score units, because dollar payoff speaks the language of business.
• If the base rate is already very high or very low, then the marginal utility of a new predictor shrinks — check base rate before investing in a costly procedure.
• If the selection ratio is high (you must hire nearly everyone), then even a valid predictor yields little utility, so improving recruitment supply comes first.
• If an SDy or validity input cannot be defended, then treat the utility estimate as a range and scrutinize it rather than reporting a single point figure.
• If the decision is one link in a sequence (screening, selection, training), then evaluate utility across the interdependent decision network, not for one stage in isolation.

• Reporting a single large dollar figure without scrutinizing SDy — the tell is a utility number with no stated assumptions or sensitivity range.
• Ignoring the base rate and selection ratio and attributing all payoff to validity — the tell is a claim that a valid test alone guarantees value regardless of applicant flow.
• Treating the criterion as one-dimensional when job performance is multidimensional and dynamic — the tell is a utility model built on a deficient or contaminated criterion.
• Computing utility for selection in isolation from workforce planning and recruitment — the tell is a payoff estimate disconnected from supply/demand forecasts and action plans.
• Documenting past practice instead of challenging it — the tell is a utility analysis used only to justify the existing procedure rather than to compare alternatives.

• ▸Predictor and criterion: name the test and the job-performance measure it predicts
• ▸Validity coefficient: enter the correlation of predictor with criterion
• ▸Selection ratio: proportion of applicants you will hire
• ▸Base rate: proportion currently successful without the procedure
• ▸SDy: standard deviation of job performance expressed in dollars
• ▸Mean standard score of selectees: average predictor score of those hired
• ▸Number selected: how many people the procedure will hire
• ▸Total procedure cost: cost per applicant times all applicants tested
• ▸Net utility and scrutiny note: dollar payoff minus cost, plus which input is least defensible

1. 1Audit your workforce against Barney's three tests: does it add positive economic benefit to producing goods or delivering services; are its skills distinguishable from competitors' (through education and workplace learning); and are those skills not easily duplicated. Rate each honestly — a workforce that fails any one test is not a source of sustained advantage.
2. 2Map your human resource system — the set of interrelated processes designed to attract, develop, and maintain human resources — and check whether it enhances or destroys the advantage (Lado & Wilson, 1994). A strong workforce inside a fragmented HR system leaks value.
3. 3Install the most valid selection process possible so you begin with higher-ability workers; validity here compounds every downstream investment because higher-ability hires yield larger productivity gains from training and require less training time.
4. 4Layer continuous development onto that selection base. Training viewed as a strategy for building sustained advantage means committing resources to develop high-caliber employees over time, producing a deep reservoir of firm-specific human capital no competitor can match.
5. 5Reinforce learned skills on the job through the immediate supervisor. If the supervisor does not reward what was learned in training, the training becomes 'encapsulated' (Haire, 1964) and transfer is minimal or negative — destroying the development investment.
6. 6Set quantitative control-and-evaluation yardsticks so you can measure deviations objectively and guide workforce-planning activities through time, identifying causes of drift from the plan across functional boundaries.
7. 7Have managers articulate and communicate the value-creation story — what the organization stands for, how it competes — and translate it into implications for employee knowledge and behavior, using it as a benchmark to assess progress.

Set an HR objective as a quantitative yardstick: reduce the first-year attrition rate of clerks from the present 50 percent to 20 percent within three years. Because the objective is numeric, the control-and-evaluation process is simplified — at the end of the third year you compare actual attrition against the 20 percent target and measure deviation precisely. Suppose you pair this with a highly valid selection process that yields higher-ability clerks; the productivity gain from a follow-on training program (say, financial analysis) is greater with these employees than it would be with poorly selected ones, and their training time is shorter, so net training effectiveness rises. Over three years the firm accumulates firm-specific human capital that competitors cannot quickly replicate. The advantage is sustained not by the training alone but by the combination of valid selection plus continual development. This is the deep reservoir of firm-specific human capital the passages describe.

• If a workforce capability can be bought or copied by any competitor in the global economy (capital, equipment, technology, information), then it cannot be your source of advantage — only the non-mobile workforce qualifies.
• If a candidate can be excluded on demonstrable job-related criteria, then the exclusion is proper; if the basis is a prohibited, non-job-related factor (age, race, gender, disability), then it is unlawful discrimination and destroys workforce quality claims.
• If you can only afford one improvement, then strengthen selection validity first, because higher-ability hires amplify the return on every subsequent training dollar.
• If the immediate supervisor will not reinforce trained behavior, then delay or redesign the training — otherwise transfer is minimal or negative.
• If a workforce skill fails any one of Barney's three tests (economic benefit, distinguishable, not duplicable), then reclassify it as table-stakes, not competitive advantage.

• Treating training as an isolated cost rather than a strategy layered on valid selection — the tell is training rolled out uniformly to high- and low-ability employees with no measured productivity differential.
• Letting an inconsistent HR system undermine good hires; the tell is strong recruiting but weak development and retention, so the human resource system destroys rather than enhances the advantage.
• 'Encapsulated' training — the tell is high satisfaction scores in the classroom but no on-the-job behavior change because supervisors never reinforce it.
• Setting only qualitative goals for workforce planning; the tell is vague year-end reviews where deviations from plan cannot be measured precisely and causes go unidentified.
• Assuming skills that everyone in the industry has still confer advantage; the tell is claiming competitive advantage from capabilities that are easily duplicated or freely available across the global economy.

• ▸Capability under review: name the specific workforce skill or process
• ▸Economic benefit test: what positive benefit does it add to producing goods/delivering services?
• ▸Distinguishability test: how do these skills differ from competitors' (education/workplace learning)?
• ▸Duplication test: what makes these skills hard for competitors to copy?
• ▸HR system link: which attract/develop/maintain processes support this — and do any destroy it?
• ▸Selection validity: how valid is the process that supplies people for this capability?
• ▸Development commitment: what ongoing resources build the firm-specific human capital?
• ▸Reinforcement check: will immediate supervisors reward the trained behavior on the job?
• ▸Quantitative yardstick: numeric target and timeframe to measure deviation (e.g., attrition 50%→20% in 3 yrs)

1. 1Define the job performance universe first, then partition it into identifiable job performance domains — the specific segments about which you will make inferences (Lawshe, 1975). Most jobs contain several domains; name each one operationally.
2. 2Align your analysis method with your purpose before choosing a technique. If the objective is establishing minimum qualifications, defining job specifications, or building selection predictors, the method differs — choose observation, interview, SME panels, questionnaires (PAQ), F-JAS, or critical incidents accordingly.
3. 3Locate each job behavior on the observation-to-abstract continuum. Directly observable and reportable behaviors support content-related evidence (proficiency, job knowledge, work sample tests); highly abstract qualities (e.g., inductive reasoning) require construct-related evidence.
4. 4Justify Inference 10 by producing a valid description of the performance domain — the job description — that reflects actual job demands adequately analyzed.
5. 5Justify Inference 11 by verifying the links between job behaviors and job outcomes; job analysis is the process used to discover and specify these behavior-outcome links.
6. 6Derive KSAOs and criterion dimensions from this analysis so that criterion measures are as psychometrically rigorous as predictor measures, and are embedded in the same theoretical networks as constructs on the predictor side.
7. 7Assess reliability and validity of the job analysis information itself, using sound judgments from job incumbents, supervisors, or other job experts at the observation end of the continuum.
8. 8Feed the results forward: use the documented tasks, KSAOs, and behavior-outcome links as the shared foundation for selection, workforce planning, training, and performance management.

Consider a secretary whose job performance universe consists of several job performance domains, only one of which is typing (Lawshe, 1975). You partition the universe and isolate the typing domain because it is directly observable and constitutes a significant aspect of the job. Because these behaviors sit at the observation end of the content-construct continuum, you develop content-related evidence: a typing test drawn as a purposeful, meaningful sample of the typing universe. You define the sampling process with enough precision that a user can judge how well the sample typifies performance on that domain. You do not treat the typing test as covering the whole job — it measures one domain, so other domains still require their own analysis and measures. This keeps the criterion tied to actual job demands (Inference 10) rather than drifting into deficiency or contamination.

• If the quality measured is not unduly abstract and constitutes a significant aspect of the job, then treat it functionally as content and use content-related evidence (proficiency, job knowledge, work sample tests).
• If the behavior sits at the abstract end of the continuum (e.g., inductive reasoning), then the inferential leap is larger and construct-related evidence is appropriate.
• If the objective is managerial decision making, then weight criterion elements into a single composite of overall economic worth; if the objective is understanding predictor-criterion relationships, then keep the criterion elements separate.
• If the sample and sampling process cannot be defined precisely enough for a user to judge representativeness, then you have not yet met the assumptions for content-related evidence — do not proceed.
• If you cannot verify the links between job behaviors and job outcomes (Inference 11), then the criterion is unsupported no matter how strong the predictor side looks.

• Neglecting evidence for Inferences 7, 8, and 10 — the tell is criterion measures that are less rigorous psychometrically than the predictor measures.
• Treating one job performance domain as the entire universe (e.g., scoring a typist only on typing) — the tell is criterion deficiency, a criterion that misses significant aspects of the job.
• Relying solely on direct predictor-criterion correlation (Inference 5) while ignoring Inference 8 — the tell is confident claims about Inference 9 that rest on an unvalidated operational criterion.
• Assuming a costlier selection procedure is automatically more accurate — the tell is spending on instruments without assessing base rate, selection ratio, and actual predictive ability.
• Building criteria with no prior theorizing about the performance domain's dimensions — the tell is criteria poorly embedded in theoretical networks, limiting construct validation.

• ▸Job title & performance universe: name the overall job and its full scope
• ▸Performance domains: list each identifiable segment (mark which are being sampled)
• ▸Task/duty inventory per domain: enumerate observable and reportable behaviors
• ▸KSAOs required: map each behavior to its knowledge, skill, ability, or other characteristic
• ▸Abstraction rating: place each behavior on observation-to-abstract continuum (content vs. construct evidence)
• ▸Data source & method: name source (incumbent/SME/PAQ/F-JAS/critical incidents) matched to purpose
• ▸Behavior-outcome links (Inference 11): state how each behavior connects to a job outcome
• ▸Criterion measure & rigor check: define the operational measure and note deficiency/contamination risks
• ▸Downstream use: specify which HR practice (selection, training, planning) this feeds

1. 1Anchor the workforce plan to the business strategy first: read the strategic business plan and translate each objective into the roles, skills, and headcount it implies, so HR strategy derives from business strategy rather than running parallel to it.
2. 2Build and maintain a talent inventory. Decide the information type to capture (skills, experience, performance, promotability), establish how it is obtained and updated, and specify the uses before collecting anything.
3. 3Forecast workforce supply from two sources: external supply (labor market availability for the roles you need) and internal supply (current employees who can move via promotion, transfer, or development), including formal management succession planning and CEO succession for critical leadership roles.
4. 4Forecast workforce demand: select a predictor (a business variable that historically tracks headcount), examine its historical and projected relationship to staffing, apply productivity ratios, and project workforce requirements from those relationships.
5. 5Integrate the supply and demand forecasts to expose the gap (shortage or surplus) for each role and time horizon.
6. 6Match forecast results to action plans — recruitment, training and development, promotion, redeployment, or reduction — so each identified gap has a named remedy.
7. 7Install control and evaluation: monitor actual performance against the plan, take corrective action when they diverge, and set a defined time horizon and assigned responsibility for the planning cycle (as done in the IBM workforce planning example).

IBM's workforce planning shows the control-and-evaluation loop in operation. The plan begins from business strategy, then forecasts internal and external supply against projected demand over a set time horizon. Management succession plans, including CEO succession, cover the critical leadership pipeline. Actual staffing outcomes are monitored against the forecast, and where performance diverges from the plan, corrective action is taken. Responsibility for the planning process is explicitly assigned rather than left ambient. The result is a closed loop: strategy sets the target, forecasts set the expectation, and monitoring keeps action plans honest against reality.

• If the HR strategy cannot be traced back to a specific business-strategy objective, then stop and re-derive it — workforce plans that run parallel to business plans instead of from them produce the wrong headcount.
• If internal supply plus feasible development covers projected demand for a role, then favor promotion/transfer and succession planning; if not, then activate external recruitment.
• If a role is a critical leadership position, then require a formal succession plan (and CEO succession for the top seat) rather than relying on general internal supply forecasts.
• If a demand predictor's historical relationship to headcount is unstable or weak, then question the projection before committing action plans — accuracy of demand forecasts is a deliberate judgment, not an assumption.
• If monitoring shows actual performance diverging from the plan, then take corrective action within the assigned responsibility rather than waiting for the next planning cycle.

• Treating workforce planning as an HR-only exercise disconnected from business strategy; the tell is a plan that lists headcount targets with no linked strategic objective.
• Forecasting demand without selecting and validating a predictor; the tell is projected requirements that rest on last year's numbers plus a guess rather than a historical/projected relationship.
• Ignoring internal supply and succession planning, then scrambling externally for leadership roles; the tell is repeated open critical positions with no ready internal candidate and no CEO succession plan.
• Producing forecasts but never closing the loop; the tell is no monitoring of actual versus planned and no named owner for corrective action.
• Building a talent inventory that is never updated or whose intended uses were never defined; the tell is stale skill data nobody consults when gaps appear.

• ▸Linked business objective — the strategic goal this staffing need serves
• ▸Role / occupational group — the job being planned
• ▸Time horizon — the period this forecast covers
• ▸Demand forecast — projected requirement, with the predictor and productivity ratio used
• ▸Internal supply — current qualified employees plus feasible promotions/transfers
• ▸External supply — labor-market availability for this role
• ▸Gap — shortage or surplus (demand minus total supply)
• ▸Action plan — recruit, develop, promote, redeploy, or reduce to close the gap
• ▸Owner & monitoring metric — who is responsible and what actual-vs-plan measure triggers corrective action

1. 1Start with recruitment planning: translate the workforce demand forecast into concrete staffing requirements — how many hires, in which roles, by when. This anchors recruitment to the strategic workforce plan rather than treating openings as isolated events.
2. 2Run staffing requirements and cost analyses: quantify the number of applicants you must generate at each stage to yield the required hires, and attach a cost to each. Work backward from hires needed through expected yield rates at screening, selection, and offer-acceptance.
3. 3Conduct source analysis: identify external and internal sources (web, referrals, agencies, internal supply from the talent inventory), and evaluate each on quality of applicants produced, not just volume.
4. 4Design operations: decide how applications flow through hiring-management systems and resume-processing software, and how organizational image and recruiter characteristics will be managed, since research shows these affect applicant quantity and quality.
5. 5Deploy realistic job previews (RJPs) so applicants self-select on accurate information about the job, improving downstream fit and reducing early turnover.
6. 6Build measurement, evaluation, and control into the process: track yield ratios, cost-per-applicant, and source quality against the plan, then take corrective action when a source underdelivers on qualified candidates.
7. 7Consider the applicant's perspective throughout — treat job search as a two-sided matching process, because recruitment influences both who applies and who accepts.

An organization forecasts, from its workforce demand analysis, that it needs 10 qualified hires for a role. Working backward through expected yield rates — say roughly half of candidates survive each stage from application through screening, selection, and offer acceptance — the recruiter computes that a much larger applicant pool must be generated at the top of the funnel to net those 10 hires. Source analysis compares the web, employee referrals, and agencies on the quality of applicants each has historically produced, not merely how many. The recruiter chooses sources projected to yield enough qualified applicants, prices each option in the cost analysis, and uses a realistic job preview to let applicants self-select. During operations, intelligent software processes incoming resumes and a hiring-management system tracks flow. In the measurement and control step, the recruiter monitors yield ratios and cost-per-hire per source; when one source produces high volume but few qualified survivors, it is dropped and budget shifts to the higher-quality source.

• If a source produces high applicant volume but low survival through screening and selection, then reallocate budget to sources with better quality yield rather than chasing raw numbers.
• If the job carries a high risk of early turnover from mismatched expectations, then use a realistic job preview to let applicants self-select before applying.
• If yield ratios at any funnel stage fall below what the staffing analysis assumed, then widen the top-of-funnel applicant pool or take corrective action on that stage.
• If organizational image or recruiter behavior is depressing applications, then treat these as controllable variables and address them, since research links them to recruitment outcomes.
• If demand forecasts drive the plan, then size the applicant pool by working backward through expected stage-by-stage yield rates, not by guessing.

• Optimizing for applicant quantity while ignoring quality — the tell is a full funnel that collapses at screening and selection with few qualified survivors.
• Treating recruitment as reactive event-filling rather than an extension of strategic workforce planning — the tell is openings handled ad hoc with no link to demand forecasts.
• Skipping realistic job previews and overselling the job — the tell is high offer acceptance followed by early turnover and poor fit.
• Never measuring source effectiveness — the tell is the same sources funded year after year with no yield or cost data to justify them.
• Ignoring recruiter characteristics and organizational image as controllable inputs — the tell is unexplained drop-off in application rates despite adequate outreach spend.

• ▸Hires needed by role and date — pull directly from the workforce demand forecast.
• ▸Stage yield rates — record expected survival percentage at screening, selection, and offer acceptance.
• ▸Required applicant pool size — compute by working backward from hires needed through the yield rates.
• ▸Sources selected — list each external/internal source you will use.
• ▸Quality yield per source — track qualified survivors, not raw applicants, per source.
• ▸Cost per source — enter cost-per-applicant and cost-per-hire for each source.
• ▸RJP content — note what accurate job information applicants receive to enable self-selection.
• ▸Operations/system — specify hiring-management system and resume-processing tools handling flow.
• ▸Control triggers — define the yield threshold below which you reallocate budget or widen the pool.

1. 1Anchor selection in a job analysis before choosing any instrument. Use direct observation, interviews, SME panels, or structured questionnaires (PAQ, F-JAS, critical incidents) to produce job specifications and minimum qualifications, since validity is an inference about a specific job.
2. 2Define and evaluate the criterion first. Specify job-performance dimensions (static, dynamic, individual) and screen the criterion measure for relevance, sensitivity/discriminability, and practicality while guarding against criterion deficiency and contamination.
3. 3Select predictors matched to the criterion: cognitive ability tests, personality inventories, biodata/WABs/BIBs, structured interviews, work samples, situational judgment tests, or assessment center exercises. Confirm reliability first (test-retest, parallel forms, internal consistency, interrater) because reliability caps attainable validity.
4. 4Gather validity evidence appropriate to your situation: content-related (predictor samples the job domain), criterion-related via predictive or concurrent studies, or construct-related. For criterion-related work, watch factors that shrink observed validity: range restriction, range enhancement, position in the employment process, and nonlinear predictor-criterion form.
5. 5Cross-validate to check that obtained validity holds on a fresh sample. Where local validation is not feasible, use synthetic validity, test transportability, or validity generalization.
6. 6Test for fairness before use: assess differential validity, adverse impact, and differential prediction (including moderator variables). Where adverse impact appears, consider test-score banding to reduce it while preserving job-relatedness.
7. 7Choose a data-combination and decision model: multiple regression, multiple cutoff, or multiple hurdle, and set the selection ratio against the base rate. Evaluate the full system with utility analysis so the payoff of the selection procedure is expressed in business terms.
8. 8Establish ongoing measurement, evaluation, and control so validity and adverse-impact evidence is refreshed as jobs, applicant pools, and criteria change.

A firm hiring managers builds the selection system from the criteria of managerial success and the importance of context, then assembles instruments of prediction: cognitive ability tests, objective personality inventories, motivation-to-manage measures, and personal history data, plus work samples of managerial performance run inside an assessment center (leaderless group discussion, in-basket test, business game, situational judgment tests). Assessors are trained, exercises are set at a defined level and purpose, and performance feedback is planned. The team documents reliability of the assessment process and then validity, and separately examines fairness and adverse impact before rollout. Because assessment centers carry cost, they run assessment-center utility figures to justify the investment. They combine predictors deliberately rather than adding exercises that overlap, and note potential problems such as assessors reacting to the same behaviors. The result is a defensible inference from performance in the center to future managerial performance.

• If the applicant pool is pre-screened or restricted in range, then expect obtained validity to be understated and correct for range restriction before judging the predictor.
• If local criterion-related validation is not feasible (small sample, no criterion data), then use synthetic validity, test transportability, or validity generalization rather than skipping validation.
• If a predictor shows adverse impact, then test for differential validity and differential prediction, and consider test-score banding to reduce impact while keeping the procedure job-related.
• If the criterion measure is unreliable or contaminated, then fix the criterion first, because validity cannot exceed what the criterion allows.
• If you must apply several predictors and some are non-compensatory, then use a multiple-hurdle or multiple-cutoff model; if scores can trade off, use multiple regression.

• Choosing a test before analyzing the job. Tell: you cannot state which job specifications or performance dimensions the instrument predicts.
• Treating reliability as sufficient. Tell: the report cites test-retest or interrater consistency but never presents content-, criterion-, or construct-related validity evidence.
• Reading raw validity coefficients at face value. Tell: no mention of range restriction, range enhancement, or position in the employment process, so a small coefficient is misjudged as a weak predictor.
• Reporting validity from the development sample only. Tell: no cross-validation, so shrinkage on new applicants is unknown.
• Ignoring fairness until challenged. Tell: adverse impact, differential validity, and differential prediction are absent from the record, leaving no defense under EEOC/OFCCP scrutiny.

• ▸Job specification source: name the job-analysis method used (observation, interview, SME panel, PAQ, F-JAS, critical incidents).
• ▸Criterion measure: state the performance dimension and check relevance, sensitivity, practicality; flag deficiency/contamination.
• ▸Predictor(s): list each instrument and its intended inference to the criterion.
• ▸Reliability estimate: record type (test-retest, parallel, internal, interrater) and value.
• ▸Validity evidence type: content / criterion-related (predictive or concurrent) / construct-related.
• ▸Validity-attenuation factors: note range restriction, range enhancement, employment-process position, relationship form.
• ▸Cross-validation / alternative strategy: cross-validated? or synthetic validity / transportability / validity generalization used.
• ▸Fairness check: differential validity, adverse impact, differential prediction results; banding applied? yes/no.
• ▸Decision model & utility: model (regression/multiple-cutoff/multiple-hurdle), selection ratio, base rate, utility estimate.

1. 1Run needs assessment before you design anything, using the three levels named in the source: organization analysis (does the organizational context support training?), operations analysis (what tasks and behaviors the job requires), and person analysis (which employees need what). Skip none — each answers a different question.
2. 2Convert the operations and person findings into explicit training objectives that define what is to be learned. State them as observable behaviors, not vague topics, so you can later measure whether the change occurred.
3. 3Design the learning environment around the principles named in the source that enhance learning: knowledge of results (feedback), transfer of training, reinforcement, practice, motivation, goal-setting, and behavior modeling. Build each into the program deliberately rather than assuming it happens.
4. 4Account for trainability and individual differences up front — screen or segment learners so the design fits their starting capability, and add adaptive guidance and self-management components so trainees can maintain behavior change after the program ends.
5. 5Engineer transfer of training explicitly: make the practice conditions resemble the job, and use self-management so the relatively permanent behavior change survives return to the workplace instead of decaying.
6. 6Before implementing, decide why you are measuring outcomes and choose criteria that are relevant, sensitive, and free of deficiency and contamination — the same criterion standards applied to job performance measurement.
7. 7Evaluate with a real design (e.g., the classical experimental design referenced in Chapter 16), not just post-training satisfaction, so you can attribute behavior change to the program and translate results into financial impact for managers.
8. 8Feed program-evaluation data back to decision makers to influence future training investment decisions, closing the loop between measurement and the next cycle of design.

A firm wants to reduce a recurring performance gap. Organization analysis confirms managers and the incentive system support the change. Operations analysis specifies the exact job behaviors required; person analysis identifies which employees fall short. From these, the designer writes objectives as observable behaviors. The program then builds in feedback (knowledge of results), spaced practice, goal-setting, and behavior modeling, plus a self-management module so the change holds after trainees return to their jobs. Outcomes are measured against relevant, sensitive criteria using a classical experimental design, and the results are expressed in financial-impact terms and reported back to managers to guide the next training investment.

• If organization analysis shows the context or reward system will not support the new behavior, then fix the environment first — training alone will not produce permanent change.
• If you cannot state what is to be learned as an observable behavior, then you cannot yet design or evaluate the program; return to operations and person analysis.
• If trainees vary widely in starting capability, then segment by trainability and add adaptive guidance rather than delivering one uniform program.
• If behavior improves in training but not on the job, then treat it as a transfer-of-training failure and add self-management and job-similar practice, not more content.
• If you can only report reaction or satisfaction data, then you have not measured whether relatively permanent behavior change occurred — use a controlled design and job-relevant criteria.

• Skipping one of the three needs analyses — the tell is a program built around available content or a popular topic rather than a documented job gap.
• Confusing cost with value — assuming a more expensive program is a better one; the source warns there is no one-to-one relationship between cost and utility.
• Deficient or contaminated criteria — measuring what is easy (attendance, quiz scores) instead of the targeted behavior; the tell is glowing evaluations with no on-the-job change.
• No transfer mechanism — the tell is skills demonstrated in the classroom that vanish within weeks back on the job.
• Measuring only satisfaction and skipping a controlled design — the tell is you cannot rule out that changes came from something other than the training.

• ▸Organization analysis: does context/rewards support the change? (yes/no + evidence)
• ▸Operations analysis: exact job tasks/behaviors required
• ▸Person analysis: who needs training and on what
• ▸Training objectives: target behavior stated observably
• ▸Learning principles built in: feedback, practice, reinforcement, goal-setting, behavior modeling (check each)
• ▸Trainability/individual differences: segmentation and adaptive guidance plan
• ▸Transfer plan: self-management + job-similar practice steps
• ▸Outcome criteria: relevance / sensitivity / deficiency / contamination check
• ▸Evaluation design + financial-impact metric reported to managers