PMGT2822 Chap.4 Option Ranking, Bias and Decision Quality

Option Ranking, Bias and Decision Quality

Option ranking begins after the problem and objective are sufficiently explicit. Criteria should follow stakeholder value, feasibility, risk, adoption and project purpose. If criteria are inherited from the sponsor's first frame, the matrix can reproduce the original bias while appearing analytical. Trace each criterion to the reframed objective and affected party.

Measures need direction, scale and evidence.

Cost in dollars and trust on a five-point judgement cannot be treated as equally precise. Normalisation and scoring rules can change order. Use qualitative categories where data do not support fine numbers and explain the consequence of uncertainty rather than hiding it in decimal places.

Weights express trade-offs and values; they are not empirical facts.

Elicit them transparently, test alternatives and identify threshold criteria that should not be traded. A safety requirement may eliminate an option before weighted scoring. If the ranking changes under modest defensible weights, the decision is sensitive and needs deliberation or more evidence.

Sensitivity analysis changes assumptions, scores, weights or scenarios and observes the recommendation.

It can identify a robust option, a crossover and a valuable information need. Producing many tables without linking them to a decision is not sensitivity analysis.

State what action follows if a parameter crosses its threshold.

Behavioural biases can affect projects: anchoring on an early estimate, availability after a vivid failure, confirmation in evidence search, optimism about duration, status quo preference and planning fallacy. Labels should not be used to dismiss opponents.

Redesign the process through independent estimates, reference classes, premortems and staged review.

Escalation of commitment occurs when prior investment and identity support continuation despite deteriorating future value. Sunk cost is irrelevant to the forward choice except where it creates future capability or switching consequence.

Compare expected future benefits, costs and risk of continuing, changing or stopping.

Worked matrix: Option A wins because the sponsor assigns 40% weight to visible delivery speed. Users assign higher weight to accessibility and operators to transition reliability. Report the base ranking, test stakeholder weight sets, identify non-negotiable criteria and seek evidence where scores drive reversal.

Do not average values without explaining legitimacy.

Group decision methods need process fairness. Participants should understand criteria, contribute before anchoring and see how input changes the result. A decision owner remains accountable and may depart from the matrix, but should explain why.

The tool is a record of judgement, not a machine that owns the choice.

For quizzes and exam preparation, translate common biases into process changes. If anchoring is likely, obtain independent estimates; if overconfidence is likely, use calibrated ranges and base rates; if escalation is likely, separate review from the original champion and use stop criteria.

Application earns more than a list.

Ranking should end with a decision note, not merely a score. State the selected option, threshold checks, decisive criteria, sensitivity range, evidence gaps, dissent and review trigger.

If the owner chooses a lower-ranked option, explain the legitimate consideration omitted or poorly represented by the model rather than quietly changing weights afterward.

Value-of-information reasoning can focus the next test. Ask whether plausible new evidence could reverse the choice and whether that evidence is cheaper and faster to obtain than the cost of delay.

Testing every uncertain input wastes time; ignoring a single decision-driving uncertainty creates false confidence. Prioritise variables near a crossover.

Distribution matters alongside totals. An option with the best aggregate score may impose a concentrated burden on a low-power group.

Display who receives each benefit and cost, apply legitimate non-tradeable constraints and show whether compensation or redesign changes the choice. Do not bury distribution inside an average stakeholder score.

Keep the unweighted raw evidence beside every aggregate result so reviewers can reconstruct the recommendation.

What this chapter covers

• 01

  decision criterion

• 02

  sensitivity analysis

• 03

  escalation of commitment

• 04

  compare project options transparently while detecting framing effects, overconfidence, escalation and false numerical precision

• 05

  A ranking method structures judgement; it cannot manufacture comparable data, neutral weights or a legitimate objective.

Audit a weighted ranking

Key terms

decision criterion

A dimension and measurement rule used to compare how well options serve the reframed objective.

sensitivity analysis

A test of how conclusions change when assumptions, weights or uncertain values change.

escalation of commitment

Continued investment in a course of action because of prior commitment despite evidence favouring change.

Option Ranking, Bias and Decision Quality FAQ

Exam move

Retrieve the frame, map assumptions and stakeholders, generate a changed option, test one uncertainty and state the adoption boundary.

Working through Option Ranking, Bias and Decision Quality in PMGT2822? Sia is AskSia’s AI Project Management tutor — ask any PMGT2822 Option Ranking, Bias and Decision Quality question and get a clear, step-by-step explanation grounded in how PMGT2822 is taught and assessed. Read this chapter free, then take your hardest questions to Sia.