Results-Based Accountability

RBA Glossary

Plain-language explanations of the framework, measurement, equity, and evidence terms used in this prototype.

Framework Concepts and Prototype Conventions

Use this glossary to understand RBA language and interpret the scorecards. Core framework terms are distinguished from the measurement, privacy, and display conventions used by this site. For individual metric definitions and formulas, use the Indicator/KPI Glossary.

A Key Distinction: In formal RBA, an indicator describes a population condition, while a performance measure describes a program's services or customer outcomes. This site's Indicator/KPI navigation label is broader; it does not make every metric a population indicator.

96 of 96 Terms Shown

RBA Foundations

Core framework concepts distinguish community conditions from the work and outcomes of a particular program.

Results-Based Accountability

Also Known as: RBA

A framework developed by Mark Friedman that starts with the conditions people should experience, then uses evidence and shared action to work toward those conditions. It separates population accountability from performance accountability.

Result / Outcome

Also Known as: Population Result / Condition of Well-Being

In population-level RBA, a desired condition for a defined community, such as residents being able to afford essential transportation. It describes people's lives, not a service, funding amount, or numerical target.

Population

The complete group whose well-being is being considered, defined by characteristics, geography, and time. It is not necessarily the same as the people enrolled in a program.

Population Accountability

Shared responsibility among partners for improving conditions across a whole population. No single program controls all the factors affecting community-wide results.

Performance Accountability

Responsibility for a program's services, their quality, and outcomes for its customers or participants. Program outcomes can contribute to community results without proving that the program caused a population-wide change.

Indicator

Also Known as: Population Indicator / Community Indicator

In formal RBA usage, a measure of a population condition, such as the regional share of households facing a high cost burden. The prototype uses Indicator/KPI as a broader navigation label; check each measure's actual population and purpose.

Performance Measure

Also Known as: Program Performance Measure

A measure of a program or service that addresses how much it does, how well it works, or whether its customers are better off. Its scope is the defined program, service, or customer group rather than automatically the entire region.

Key Performance Indicator

Also Known as: KPI / Metric

A metric selected to support a decision or monitor progress. In this prototype, KPI is a general label for a catalog measure, not proof that its definition, source, target, or importance has been approved.

Participant / Customer

Also Known as: Client / Service Recipient

A person, household, organization, or other defined unit receiving a program's service. Identify the unit explicitly: a household, application, dwelling, project, and trip are not interchangeable.

Beneficiary

A person or community intended to benefit from an activity or investment, whether or not they directly receive program administration or funding. For a grant program, the funded jurisdiction and the residents benefiting from a project are different groups.

Result Statement

Also Known as: Proposed Result

A plain-language description of the condition the work is intended to support. The prototype's proposed program results guide discussion; they do not establish that the desired condition has been achieved.

The Three RBA Questions

Read quantity, quality, and participant outcomes together; service activity alone does not demonstrate a benefit.

How Much Did We Do?

Also Known as: How Much / Quantity of Effort / Output / Outputs

The amount of service or activity delivered, such as applications processed, trips provided, or projects completed. A higher count shows more activity, not necessarily better quality, equity, or participant outcomes.

How Well Did We Do It?

Also Known as: How Well / Quality of Effort / Service Quality

How services perform against relevant standards, such as timeliness, reliability, accessibility, or participant experience. A percentage can describe service quality without being a better-off measure.

Is Anyone Better Off?

Also Known as: Better Off / Participant Outcome / Customer Outcome

Whether people experience a meaningful improvement, such as improved comfort, reduced hardship, or more useful mobility options. These outcomes can be measured as counts or percentages and require an appropriate follow-up and comparison method.

Quantity of Effect

The number of participants experiencing a specified better-off outcome. It is different from the number served; the outcome must actually be measured for the relevant participants.

Quality of Effect

The proportion of the relevant participants experiencing a specified better-off outcome. The denominator and follow-up coverage are essential to interpreting that proportion.

Input

A resource used to deliver work, such as staff time, funding, or equipment. Spending and resources are not themselves proof of a participant benefit.

Measures and Comparisons

These measurement terms explain the periods, calculations, and comparisons used in the scorecards.

Baseline

The reference for assessing change. Turn-the-curve work considers the historical pattern and an explicit forecast of what could happen without a change in strategy. Here, baseline metadata identifies the comparison period; the prototype does not supply an unlabeled forecast.

Baseline Period

The time window used for the reference observation. Compare it with a suitably matched follow-up period using compatible definitions, populations, and units.

Baseline Pending

A suitable comparison period or value has not yet been established for the measure. It is not a zero baseline and should not be replaced with an invented historical observation.

Forecast

An explicitly estimated future path under stated assumptions. It is different from an observed value, a target, or a synthetic demonstration series.

Reporting Period

Also Known as: Observation Period

The month, quarter, year, or other window described by a value. A period-end snapshot, activity during a period, and a cohort follow-up can have different meanings even when their dates match.

Target

Also Known as: Approved Target

A desired level for a specific measure, population, unit, and time window. An approved target needs a recorded decision and authority; displaying a number does not make it approved.

Target Period

The time window to which a target applies, which may be a reporting period or a defined cohort's follow-up. It must not be inferred automatically for an approved source target.

Benchmark

A reference used for comparison, such as an approved service standard, a historical level, or a comparable program. A benchmark is not necessarily a target and must be comparable before it is used.

Threshold

A boundary used to classify a value or define a condition, such as a service deadline, access distance, or rating cutoff. State the rule and units; a threshold serves different purposes in eligibility, geography, and performance ratings.

Rating Rule

Also Known as: Traffic-Light Rating / On Track / Watch / Off Track

The criteria that map a measure's value to a status, including direction, cutoffs, unit, and applicable period. Green, yellow, and red need text labels and context; prototype ratings use unapproved demo rules, not actual performance judgments.

Not Rated

No applicable rating rule has been applied to the available value. It does not mean poor performance; an approved target alone does not supply every part of a rating method.

Min/Max

Also Known as: Minimum / Maximum / Range

The smallest and largest available comparable values shown for a measure, excluding missing or suppressed values. This descriptive span is not a target, a forecast, an expected range, or a confidence interval.

Trend

A pattern across comparable observations over time. One value is not a trend, and changes in definitions, coverage, or the people measured can create misleading patterns.

Unit of Measurement

What a number represents, such as households, dollars, days, trips, percent, or percentage points. Values with different units should not be added or compared as if they were the same quantity.

Numerator

The measured amount meeting the condition in a fraction, such as applications completed on time. Its unit and population must match the denominator appropriately.

Denominator

The reference amount in a fraction, such as applications due for a decision. It may be people, projects, dollars, time, or another defined exposure, so it is not automatically a participant count.

Count

The number of defined units or events. Specify whether the count is unique people, applications, households, projects, or repeated interactions, and how duplicates are handled.

Rate / Share

Also Known as: Proportion

A quantity expressed relative to a defined denominator or exposure. A share may be expressed as a percentage; not every rate is a percent, and the population and time window must remain clear.

Percentage Points

Also Known as: pp / Percentage-Point Change

The difference between two percentage values. Moving from 60 percent to 65 percent is a gain of five percentage points, not a five-percent relative increase.

Percent Change

Also Known as: Relative Change

A change expressed relative to the starting value, rather than a subtraction of percentage values. It cannot be calculated in the usual way when the starting value is zero.

Cohort

A group defined by a shared event or time window, such as enrollment or project completion. Following the same cohort is different from comparing whichever people happen to be present in two reporting periods.

Follow-Up Window

The defined interval after an event when an outcome is assessed. A 12-month outcome should include the cohort with a full opportunity to reach that follow-up point, not just recent enrollments.

Sample Size

The number of valid units included in an analysis. A large sample alone does not establish representativeness or remove bias; a small sample can also limit privacy-safe reporting and precision.

Snapshot / Stock

An amount recorded at a point in time, such as active members or cumulative obligations at quarter end. Adding successive snapshots generally double-counts the same underlying stock.

Period Flow

Activity occurring during a defined interval, such as new awards or expenditures in a quarter. It is different from a cumulative balance or a count of active participants at period end.

Equity and Interpretation

Equity questions apply to reach, service experience, and benefits, not only to the Equity and Outcomes view.

Equity

Attention to fair access, treatment, opportunity, and outcomes while addressing barriers and unequal burdens. Equal activity totals or equal treatment alone do not demonstrate equitable results.

Disaggregation

Separating a measure into valid groups or geographies to examine differences hidden by the overall average. The grouping must be supported by the data and meet privacy and disclosure requirements.

Disparity / Outcome Gap

A difference between groups in access, experience, or outcomes. State the groups, subtraction order, units, and periods; a signed difference, an absolute magnitude, and a relative ratio are not interchangeable.

Priority Group

A group identified for focused attention using an explicit policy or analytical definition. The label does not mean every person in the group has the same needs, eligibility, or outcomes.

Comparison Group

Also Known as: Reference Group

The group used as a reference when examining a difference. It is not automatically an experimental control group, and differences between groups may reflect factors beyond the program.

Eligible Population

People, households, organizations, or projects meeting the applicable program rules for the period. It is not automatically the regional population, the surveyed sample, or everyone living in an EPC.

Reach

Who receives or uses a service relative to the intended or eligible group. Service counts alone do not establish what share of eligible people has been reached.

Equity Priority Community

Also Known as: EPC / Equity Priority Communities

A geographic designation under MTC's applicable methodology. Use the correct geography and vintage; residence in an EPC does not by itself establish an individual's program eligibility or lived experience.

Coverage

The extent to which relevant people, places, services, or records are represented. State which coverage is meant: service access, a geographic catchment, complete follow-up, and survey responses use different denominators.

Selection / Nonresponse Bias

A systematic difference between the people observed and those missing from the analysis. Outcomes among respondents or linked records may not represent all participants or the eligible population.

Causal Attribution

Also Known as: Causation / Attribution

A supported claim that an intervention caused a change, rather than merely occurring alongside it. Before-and-after differences or favorable trends alone do not establish this claim.

Contribution

The role a program or partner plays in a wider result affected by many factors. Contribution can be described and evaluated without claiming sole causal responsibility for a population outcome.

Uncertainty

Limits on how precisely or confidently a value or interpretation is known. Sampling, missing data, modeling assumptions, and measurement error can all matter; a synthetic example is not an estimate with an implied accuracy.

Evidence, Data Quality, and Privacy

These are measurement and governance conventions used by this prototype, not a separate set of official RBA framework definitions.

Reported Evidence

A value stated by an identified source or report. Retain the source's population, period, and definition; a reported program total does not automatically validate a different KPI interpretation.

Derived Evidence

A value calculated from source observations through a documented transformation, such as a count, ratio, or change. Its validity depends on the inputs, definitions, and calculation.

Modeled Evidence

An estimate produced using analytical assumptions or a model rather than a direct observation of the outcome. State the method, inputs, and uncertainty, and do not relabel it as observed performance.

Synthetic Demo Data

Also Known as: Synthetic / Illustrative Data / Scenario / Fixture

Invented values used to demonstrate and discuss the site. Even when a scenario references a disclosed public total, its invented splits and outcomes remain synthetic, not evidence of actual performance or approved commitments.

Data Source

The report, system, dataset, survey, or record collection supplying evidence. A proposed source is a discovery lead, not proof of authorized access or fitness for a KPI.

Provenance / Data Lineage

The traceable record of where a value came from, which inputs and transformations were used, and which versions and dates apply. It helps readers distinguish source evidence from estimates and invented examples.

Source Readiness

Also Known as: Source Verification / Analytical Availability / Availability Today

Whether the source, permissions, definitions, quality, and reporting controls have been demonstrated to support the intended measure. Provisional NO or PARTIAL catalog flags do not prove that the agency has no data.

Data Quality

How well data supports its intended use, including accuracy, completeness, consistency, timeliness, and valid definitions. Data being present does not establish that it is suitable for a decision.

Completeness

How much required information is present and valid under a stated rule. Field completeness is different from population coverage, representativeness, or proof that recorded values are correct.

Definition Version

The version of a measure's calculation, population, exclusions, or rules. Changes can break comparability even if the KPI ID remains the same.

Reporting Cadence

Also Known as: Reporting Frequency / Refresh Cadence

How often a measure can reasonably be updated after its inputs are available and validated. Monthly source records do not imply real-time reporting.

Last Source Refresh

The recorded date of the relevant source update, not the date a page was opened or a demonstration was rebuilt. A missing refresh date should remain explicit rather than imply current source access.

Paired Data

Also Known as: Linked Baseline and Follow-Up / Paired Survey

Comparable observations linked to the same units before and after the relevant event. Unmatched samples or incomplete links should not be silently treated as paired changes.

Follow-Up Coverage

The share of the cohort due for follow-up with valid required outcome records. A favorable outcome among those observed can coexist with low coverage and uncertainty about those missing.

Survey Coverage

Also Known as: Response Rate / Survey Response

The extent to which a specified survey population or follow-up cohort has valid responses. Identify who was eligible or invited, the response definition, and whether paired responses are required; these denominators are not interchangeable.

Aggregate Data

Summaries over groups, events, or periods rather than individual records. Aggregation alone does not guarantee anonymity, particularly for small groups or overlapping releases.

Suppression

Also Known as: Suppressed / Disclosure Rule / Small Cell

Withholding a value under a disclosure or privacy rule. A suppressed value is not zero and must remain withheld in charts, calculations, components, and exports where disclosure would occur.

Missing / Unavailable Data

Also Known as: No Data / Unavailable / Missing Values

A value that is absent or not established for the stated purpose. It is different from an observed zero, a suppressed value, or a measured failure.

Normalized Measure

A value adjusted to support comparison under stated assumptions, such as energy use adjusted for weather. It is not automatically an observed bill change, net savings, or a causal program effect.

Strategy, Learning, and Responsibility

RBA connects evidence to decisions, partners, action, and a continuing review of what is changing.

Story behind the Curve

Also Known as: Story behind the Measure / Drivers / Barriers

An evidence-informed explanation of the factors shaping the pattern in the data, including conditions, policies, delivery issues, and missing evidence. It is a basis for inquiry and action, not an unsupported causal story.

Turn the Curve

A structured learning process for changing an indicator or performance measure's likely path in a beneficial direction. Consider history and forecast, investigate contributing factors, involve partners, choose feasible strategies, and review their effects; improvement is not always an upward line.

Strategy

Also Known as: What Works / Low-Cost / No-Cost Action

A reasoned set of actions intended to address contributing factors and improve a result. Consider evidence, community experience, feasibility, cost, and low-cost alternatives before choosing what to implement.

Action Plan

Also Known as: Corrective Action

Specific work translating a strategy into steps with responsible owners, timing, and a way to review progress. A status color without an action or learning process is not an improvement plan.

Partner

An organization or community participant with a role in improving a result. State each party's actual contribution and authority rather than assuming one program controls every part of delivery or every population outcome.

Measure / Action Owner

Also Known as: Candidate Owner / Accountable Owner

The proposed or confirmed party responsible for the measure's definition, interpretation, or agreed action. A candidate role in a catalog is not a confirmed assignment or proof of source access.

Data Steward

The role responsible for agreed data definitions, quality, access, and handling within its authority. Stewardship and program decision-making can sit with different people.

Data Development Agenda

Also Known as: Evidence Gap / Data Gap

The prioritized work needed to obtain useful evidence that is currently missing or inadequate. Identify the question, source or collection need, approvals, responsible party, and next step rather than inventing a result.

Director Led Equity Steering Committee

The agency review group using this prototype to discuss selection, equity, process, and direction. Viewing a demonstration does not itself approve operational measures, targets, access, or publication.

Prototype and Scorecard Labels

These terms describe this site's presentation and review states, not additional RBA performance categories.

Scorecard

A decision-focused presentation connecting a result to measures, evidence, interpretation, and next actions. It is not necessarily a combined numerical score or ranking.

Dashboard

The program's collection of scorecard views and supporting explanations. It helps people navigate evidence; its visual polish does not establish data readiness or approval.

Scorecard View

Also Known as: Executive Scorecard / Manager Scorecard / Operations Scorecard / Equity and Outcomes Scorecard

A selection of measures for an audience and decision: leadership priorities, delivery management, process follow-up, or equity and participant outcomes. The four views describe the same program; equity remains relevant to every view.

Executive Dashboard

Understand the overall result and decide where to focus resources, policy, or further evidence development.

Manager Dashboard

Understand how delivery is working and decide what to change in staffing, outreach, coordination, or participant support.

Operations Dashboard

Identify the specific process issues that need follow-up and assign corrective work.

KPI Catalog

The draft library of candidate measure definitions, formulas, sources, owners, and reporting expectations. Its scope is broader than the measures currently displayed in scorecard views.

Catalog Only

The metric is documented in the catalog and indicator glossary but is not selected for any current prototype scorecard view. It is not a label for rejection, low importance, missing source data, or poor performance.

In Prototype Views

The metric is selected for at least one of its program's current scorecard views, not necessarily all four. This describes display scope, not approval of the metric, data, or target.

Demo Target

Also Known as: Illustrative Benchmark

An invented benchmark used to demonstrate target display and rating behavior. Its value, period, direction, and status thresholds are not approved commitments or funding recommendations.

Next Decision or Review

A proposed follow-up for the program and audience, such as validating a definition before interpreting a measure. Currently this is fixed setup guidance, not a recommendation calculated from the displayed numbers, a recorded decision, or an approved action assignment.

Reporting Scope

The program, audience, institution, and other boundaries governing which approved records a user may see. A view name or glossary entry does not grant live-data access.

Sources and Scope

RBA was developed by Mark Friedman. Core framework explanations are original plain-language summaries informed by Clear Impact's RBA Introduction and Turn the Curve Guidance. Equity and local reporting conventions should be read alongside the MTC Equity Platform and the project's draft definitions. This glossary is guidance for prototype review, not a substitute for source validation, program policy, or formal training, and does not imply endorsement.