Prepare by practicing one skill repeatedly: name the decision in a scenario before choosing an action. Structure answers around stakeholders, missing evidence, options, criteria, and risk. Drill written case answers and ethics scenarios against a rubric so your first sentence frames the decision and every recommendation is tied to an explicit criterion.
What a competency-based selection assessment actually measures
It evaluates observable behaviors — analysis, collaboration, communication, results orientation, and integrity — through scenario and experience-based tasks, so preparation means building a repeatable way to structure decisions, not memorizing development terminology.
A competency is a describable behavior; knowledge is a fact you can state. Assessors typically score responses against behavioral anchors, such as whether you identified the affected stakeholders or whether your recommendation named a tradeoff. This distinction changes how you study: instead of collecting definitions of project management, collect examples where you diagnosed a problem, weighed options, and acted — and practice describing the behavior, not the job title.
The institutional context matters too. The IDB is a multilateral development bank owned by member countries and working in Latin America and the Caribbean, so scenario answers need to respect country ownership, evidence-based recommendations, and coordination with counterparts — not a generic corporate instinct to move fast and decide alone. Translate each of your own work or academic stories into: situation, my specific action, the reasoning behind it, and what changed as a result.
Situational judgment scenarios: name the decision before the action
Scenario items present a messy situation with several defensible-looking options. The skill is naming the real decision — who must be consulted, what evidence is missing, what risk each option creates — before selecting an action.
Scenario options are built to pull at two instincts: the action-oriented option that skips consultation and the passive option that defers everything upward or waits for perfect information. Both are usually defensible-sounding and both are incomplete. Before ranking options, write one sentence per option describing its most likely consequence. This forces you to compare outcomes rather than compare how decisive each option sounds.
A practical ranking rule: weight options by risk to people, risk to data or record integrity, and risk to working relationships, in that order. Then check consultation: has the person with the relevant authority been informed, and has the counterpart who owns the information been asked? An option that is fast but bypasses a counterpart or hides a risk is rarely the strongest answer, even when the scenario's time pressure makes it tempting. Use the table below when reviewing any practice scenario.
| Scenario type | Weak impulse | Structured response | Risk the response manages |
|---|---|---|---|
| Task conflict (deadline vs quality) | Cut corners silently to deliver on time | Flag the tradeoff, propose a scoped interim deliverable, let the owner decide | Record integrity and team trust |
| Stakeholder disagreement | Pick a side early to look decisive | Map each party's interest, find shared criteria, escalate with options if unresolved | Country ownership and working relationships |
| Ethical pressure | Stay silent to protect a relationship | Pause the step, decline clearly, use the designated channel, document | Accountability and personal integrity |
| Data ambiguity | Fill gaps with plausible assumptions | Request the data with a deadline, present partial evidence labeled as partial, state assumptions explicitly | Evidence base and counterpart ownership |
| Unclear instruction from above | Guess and act, or stall entirely | Confirm scope in one clarifying message, proceed on the unambiguous parts, note the assumption | Speed of results without misdirection |
Worked scenario: the missing dataset three days before a report
A junior analyst waits on a country counterpart's data for a due report while the team lead travels. The tempting fix is filling gaps with assumptions; the stronger response protects data integrity, consults, and communicates the timeline risk.
Scenario: you need sector statistics from a ministry counterpart for a progress report due in three days; your lead is unreachable until the deadline. The plausible mistake is drafting the report with placeholder numbers dressed up as estimates and submitting on time to look reliable. Why it fails: it corrupts the record the team and the counterpart will rely on, it replaces a counterpart's ownership of their own data with your guess, and the error surfaces later at a much higher cost than a delayed draft.
The better decision has three moves. First, send the counterpart a specific, narrowed request — exactly which figures, by when, with an offer to accept partial coverage if complete data is impossible. Second, draft the report now with the sections you can support, marking missing figures visibly rather than invisibly. Third, message your lead a short options note: submit with gaps labeled, request a short extension, or both. This matters because it treats the deadline as a constraint to manage openly, not a reason to compromise the evidence base — which is the judgment the scenario is really probing.
Country and sector context: apply concepts, do not recite them
Development questions reward applying concepts — targeting, sustainability, coordination with country systems — to the specific situation. Reciting a definition or naming a strategy document without connecting it to the scenario adds little.
Learn the difference between knowing a term and using it. Knowing 'sustainability' means you can define it; applying it means that, given a scenario about a rural water project, you ask who maintains the system after implementation closes, whether a tariff or budget line funds that maintenance, and which institution is accountable. In written answers, one applied sentence — 'the recommendation depends on the municipality having a funded maintenance unit, which the case does not establish' — demonstrates more than a paragraph of terminology.
Build a small context file rather than trying to master the whole region. Pick one or two sectors and one or two countries you can honestly claim familiarity with, and for each record: the main sector challenge, one institutional actor, one type of evidence used to measure progress, and one recurring implementation constraint. Before the assessment, rehearse attaching each item to a decision: 'given this constraint, I would prioritize X because Y.' Applied context becomes an argument; memorized context becomes decoration.
Structuring written case answers with a results chain
A results chain — inputs, activities, outputs, outcomes, impact, with assumptions — gives written responses a spine: problem, evidence, options, criteria, recommendation, risks, and how results would be monitored.
For a one-page case answer, work the chain in order. State the problem and the evidence you actually have. Generate two or three realistic options, not one obvious option plus a strawman. Choose explicit criteria — cost, speed of results, risk to beneficiaries, fit with country systems, measurability — and score the options against them. Only then recommend, and follow the recommendation with the two biggest assumptions it rests on and how you would monitor whether it is working. This structure is fast to execute because each section has a fixed job.
The most common structural mistake is recommending first and reverse-engineering reasons afterward; the reader cannot tell whether you compared anything. The fix is mechanical: write your criteria sentence before your recommendation sentence, every time. A self-test: cover your recommendation and read only the criteria and options — could a stranger reconstruct why you chose it? If not, the reasoning is decoration rather than analysis, and the answer reads as opinion. Keep results-chain vocabulary light in the prose; the logic should be visible even to a reader who has never seen a logframe.
Ethics and integrity questions in a multilateral setting
Integrity items ask what you do when speed, loyalty, or hierarchy pressures conflict with rules. The reliable pattern: pause the questionable step, decline clearly, use the designated reporting channel, document, and protect people and records.
Distinguish ethics questions from compliance questions. A compliance question applies a known rule — procurement thresholds, data restrictions — and mainly tests whether you know a designated channel exists. An ethics question involves competing legitimate values, such as a colleague's career versus the accuracy of a report. For both, the pattern holds: do not perform the questionable step yourself, do not quietly fix it alone, and do not treat staying silent as neutrality. Silence in these scenarios is a decision, and usually the weakest one.
Worked micro-scenario: a colleague asks you to backdate a training certificate so a team milestone looks complete. The plausible mistake is signing it to preserve the relationship, rationalizing that the training really happened. Why it matters: backdating makes a record false even when the underlying fact is true, and the institution's accountability depends on records being trustworthy. The better decision: decline plainly, state the reason (the date must reflect reality), offer the legitimate path — report the actual completion date and flag the milestone delay to the lead — and raise it through the appropriate channel if the colleague proceeds. Notice the answer still solves the colleague's real problem; it just does not solve it dishonestly.
A two-week practice sequence with readiness checks
Build the sequence around written scenario decisions scored against a rubric: alternate case practice with context reading, then verify you can structure any answer quickly and defend every choice against explicit criteria.
Suggested adaptable sequence: days one to four, build your decision framework and convert three of your own experiences into behavior-led narratives; days five to eight, write timed case answers using the results-chain structure, one per day; days nine to twelve, drill ethics and judgment scenarios using the consequence-ranking rule from earlier; days thirteen and fourteen, run a full mock set, then re-read your answers after a break and score them. Adjust the weighting toward whichever task type feels weakest — the sequence is a template, not a prescription.
Core exercise: choose three scenarios (one task conflict, one stakeholder disagreement, one ethics item). Write each answer cold, then score it with this rubric and rewrite once. Expected observations on the first pass: your drafts overweight immediate action, mention consultation late or not at all, and state recommendations without criteria. After the rewrite, a strong answer names the decision in its first sentence, identifies the missing information, and ties the recommendation to at least one explicit criterion. You are ready when that pattern appears on the first pass, not the rewrite.
- Readiness check 1: first sentence of any scenario answer names the decision, not the situation.
- Readiness check 2: every recommendation is preceded by at least one stated criterion.
- Readiness check 3: every answer identifies what information is missing and who owns it.
- Readiness check 4: ethics answers name a channel and a documentation step, not just a refusal.
- Readiness check 5: re-reading a draft after 48 hours, you can still reconstruct the reasoning from the criteria alone.
References and further reading
Use these references to explore the concepts and check the latest information from the relevant organizations.