Skip to main content
AQRApplied Quantitative Reasoning

Quarter 3

Build a Decision Tool / App

Vocabulary for choices, criteria, costs, risks, decision rules, testing, and revision.

Back to all vocabulary sections

Published vocabulary

21 terms in the standard AQR format.

Start with the plain-language meaning, then use the fuller definition and sentence stem as the term becomes familiar.

  1. Quarter-essential

    Decision rule

    Full definition
    A decision rule is an explicit procedure that uses inputs, criteria, conditions, scores, or thresholds to produce a result or recommendation.
    Plain language
    The rule the tool follows to turn information into an answer.
    How we use it in AQR
    Your tool must show its logic so another person can understand and test how the recommendation was produced.
    Example question or sentence stem
    The decision rule says that if ______, then ______.

  2. Quarter-essential

    Failure case

    Full definition
    A failure case is a situation in which a model or tool produces an unreasonable, harmful, incorrect, confusing, or unusable result.
    Plain language
    A situation where the tool breaks down or gives bad advice.
    How we use it in AQR
    You will deliberately look for cases that expose weak logic so you can revise before final use.
    Example question or sentence stem
    One failure case occurs when ______.

  3. Quarter-essential

    Opportunity cost

    Full definition
    Opportunity cost is the value of the best alternative not chosen when a decision is made. It focuses on what could have been gained from the next-best option.
    Plain language
    What you miss out on by choosing something else.
    How we use it in AQR
    You will identify the most important alternative being given up, not list every possible loss.
    Example question or sentence stem
    The opportunity cost of this choice is ______.

  4. Quarter-essential

    Option

    Full definition
    An option is one possible choice, action, product, plan, route, or response being considered in a decision. Options should be clear enough to compare using the same criteria.
    Plain language
    One of the choices you could make.
    How we use it in AQR
    Your decision tool needs a defined set of realistic options rather than a vague question with no comparable choices.
    Example question or sentence stem
    The options being compared are ______.

  5. Quarter-essential

    Priority

    Full definition
    A priority is a goal, need, value, or criterion judged to be more important than competing considerations.
    Plain language
    Something that matters more than other things in the decision.
    How we use it in AQR
    A user's priorities help determine criteria, weights, and acceptable trade-offs.
    Example question or sentence stem
    My highest priority is ______.

  6. Quarter-essential

    Sensitivity

    Full definition
    Sensitivity describes how much a model's output or recommendation changes when an input, weight, criterion, or assumption changes.
    Plain language
    How easily the answer changes when one part changes.
    How we use it in AQR
    You will change one reasonable input at a time to see whether the recommendation is stable or fragile.
    Example question or sentence stem
    The recommendation is sensitive to ______.

  7. Quarter-essential

    Validation

    Full definition
    Validation is the process of checking whether a model, tool, method, or result works reasonably for its intended purpose and users. Validation may involve examples, known cases, user tests, or comparison with evidence.
    Plain language
    Checking whether the tool works the way it is supposed to.
    How we use it in AQR
    AQR validation means more than clicking buttons; it means testing whether outputs make sense for real situations.
    Example question or sentence stem
    We validated the tool by ______.

  8. Quarter-essential

    Weight

    Full definition
    A weight is a value assigned to show how much influence a criterion has in a score, model, or decision rule. Larger weights give that criterion more effect on the result.
    Plain language
    A number showing how important a criterion is in the model.
    How we use it in AQR
    You will explain weights as value choices, not pretend they are automatically objective.
    Example question or sentence stem
    This criterion receives more weight because ______.

  9. Recognize and begin using

    Conditional

    Full definition
    A statement or recommendation is conditional when it depends on another condition, input, or event being true.
    Plain language
    It is true or recommended only if something else is true.
    How we use it in AQR
    Many good recommendations should be conditional because the best choice changes with the user or situation.
    Example question or sentence stem
    This recommendation is conditional on ______.

  10. Recognize and begin using

    Expected value

    Full definition
    Expected value is the probability-weighted average outcome of a repeated uncertain situation. It combines possible outcomes with their chances, but it does not guarantee what will happen once.
    Plain language
    The long-run average outcome when chances and results are combined.
    How we use it in AQR
    Expected value may help compare risky options when probability information is available and meaningful.
    Example question or sentence stem
    The expected value helps compare ______.

  11. Recognize and begin using

    Prototype

    Full definition
    A prototype is an early working version of a product, model, or tool created for testing, feedback, and revision before final use.
    Plain language
    A first working version built to learn from.
    How we use it in AQR
    Your prototype is expected to be incomplete; its job is to reveal what needs to be improved.
    Example question or sentence stem
    The prototype revealed that ______.

  12. Recognize and begin using

    Reliability

    Full definition
    Reliability is the degree to which a tool, measure, process, or source produces consistent and dependable results under similar conditions.
    Plain language
    How consistently something works or gives similar results.
    How we use it in AQR
    You will examine whether the same inputs produce stable outputs and whether the tool behaves dependably across tests.
    Example question or sentence stem
    The tool's reliability is limited by ______.

  13. Recognize and begin using

    Scenario

    Full definition
    A scenario is a possible set of conditions used to imagine, compare, or test what may happen in a situation.
    Plain language
    A possible version of the situation.
    How we use it in AQR
    You will test tools under ordinary, difficult, and unusual scenarios instead of checking only one easy example.
    Example question or sentence stem
    Under this scenario, ______.

  14. Recognize and begin using

    Score

    Full definition
    A score is a numerical result used to summarize performance, fit, or value across one or more criteria. A score depends on the rules used to create it.
    Plain language
    A number the model uses to summarize how an option did.
    How we use it in AQR
    You will explain what a score represents and avoid treating it as truth without examining the formula behind it.
    Example question or sentence stem
    This option received a score of ______ because ______.

  15. Recognize and begin using

    Threshold

    Full definition
    A threshold is a cutoff value or condition at which a category, output, decision, or action changes.
    Plain language
    A cutoff point that changes what happens next.
    How we use it in AQR
    Decision tools may use thresholds, but you should test whether the cutoff is reasonable and what happens near it.
    Example question or sentence stem
    If the value crosses the threshold of ______, then ______.

  16. Extension vocabulary

    Calibration

    Full definition
    Calibration is the process of adjusting a model, measure, or tool so its outputs better match known evidence, expected behavior, or real results.
    Plain language
    Tuning a tool so its results better match reality.
    How we use it in AQR
    You may calibrate weights, thresholds, or scoring rules after testing reveals that the tool consistently overstates or understates something.
    Example question or sentence stem
    We calibrated the model by changing ______ after noticing ______.

  17. Extension vocabulary

    Multi-criteria analysis

    Full definition
    Multi-criteria analysis is a structured method for comparing options using several criteria that may differ in importance or measurement.
    Plain language
    Comparing choices across several things that matter.
    How we use it in AQR
    A stronger decision tool may combine criteria, weights, scores, and trade-offs while keeping the logic visible.
    Example question or sentence stem
    The multi-criteria analysis favors ______ because ______.

  18. Extension vocabulary

    Normalization

    Full definition
    Normalization is the process of converting values measured on different scales into a common form so they can be compared or combined more fairly.
    Plain language
    Putting different kinds of scores onto a shared scale.
    How we use it in AQR
    Extension learners may normalize cost, distance, ratings, or other variables before combining them in a model.
    Example question or sentence stem
    We normalized ______ so it could be compared with ______.

  19. Extension vocabulary

    Optimization

    Full definition
    Optimization is the process of finding the best available solution under stated goals, criteria, and constraints.
    Plain language
    Finding the best choice that still fits the limits.
    How we use it in AQR
    Extension learners may improve a decision rule by identifying what is being maximized or minimized and which constraints must remain satisfied.
    Example question or sentence stem
    This model is trying to optimize ______ while respecting ______.

  20. Extension vocabulary

    Robustness

    Full definition
    Robustness is the degree to which a model, method, or conclusion remains useful and stable when reasonable assumptions, inputs, or conditions change.
    Plain language
    How well the result holds up when things change a little.
    How we use it in AQR
    A robust recommendation survives several reasonable sensitivity tests instead of collapsing after one small change.
    Example question or sentence stem
    The model is robust because ______.

  21. Extension vocabulary

    Utility

    Full definition
    Utility is the usefulness, satisfaction, or value that an outcome or option has for a particular person or purpose. Utility may differ across users even when measurable costs are the same.
    Plain language
    How useful or valuable an outcome is to the person making the decision.
    How we use it in AQR
    Decision tools can represent different users' values while making those value choices visible.
    Example question or sentence stem
    This option has greater utility for this user because ______.