- Practice
- About 15 min
- 5 practice questions
Accountability for process and for results
Compare process accountability with outcome accountability, see how targets can be gamed, and design performance measures that hold officials to account without distorting their work.
If something looks wrong, please tell us.
Traditional administration held officials to account mainly for following procedure. Modern reform asks them to answer for results: children learning, patients cured, roads built. Both have strengths and both can go wrong. This lesson applies the distinction to cases.
Two kinds of accountability
| Feature | Process accountability | Outcome accountability |
|---|---|---|
| Question asked | Did you follow the rules? | Did you achieve the result? |
| Strength | Fairness, legality, protection against arbitrariness | Focus on what citizens actually receive |
| Risk | Box-ticking; rules followed while goals fail | Gaming numbers; cutting corners; ignoring hard cases |
| Example | Every file has the required signatures | Percentage of children reading at grade level |
Gaming of targets
The social scientist Donald Campbell observed, in substance, that the more a quantitative indicator is used for social decision-making, the more it is subject to corruption pressures and the more it distorts the very processes it is meant to monitor. The economist Charles Goodhart made a related point about economic targets. In administration, gaming takes familiar forms:
- Creaming: serving easy cases to meet targets and neglecting hard ones.
- Reclassification: recording cases in categories that improve the figures.
- Teaching to the test: improving the measured skill while neglecting real learning.
- Falsification: simply reporting false numbers.
Worked example 1: the crime statistics
A police district is ranked by its fall in recorded crime. Within a year, recorded crime falls sharply, but complaints from citizens that police refused to register cases rise. Analysis: outcome accountability based on a single indicator has created an incentive to avoid registering crime. Fix: measure several indicators (registration on complaint, victim satisfaction surveys, cases solved), use independent checks such as surveys, and make refusal to register a serious lapse. The aim is to make honest reporting the safest choice.
Worked example 2: the hospital that followed every rule
A district hospital passes every procedural inspection: registers maintained, signatures complete, stock books balanced. Yet patient deaths after surgery are high. Analysis: process accountability is satisfied while outcomes fail. Fix: add outcome measures such as infection rates and post-surgery mortality, with clinical audit by peers, and support for improvement rather than only punishment.
Designing good measures
- Use a small set of indicators that together capture the goal.
- Include at least one measure that is hard to game, such as an independent survey.
- Combine numbers with qualitative review.
- Adjust for difficulty, so that serving hard cases is not penalised.
- Keep process safeguards for fairness and legality.
Hold officials to account for results, but measure in ways that make honest work the easiest path, and never drop the rules that protect fairness.
Key takeaways
- Process accountability protects fairness; outcome accountability focuses on results.
- Single indicators invite gaming: creaming, reclassification, teaching to the test, falsification.
- Good measurement uses several indicators, independent checks and difficulty adjustment.
Check your understanding
-
1. 'Did you achieve the result?' is the question of:
- Process accountability
- Outcome accountability
- Judicial accountability
- Collective responsibility
Show answer
Answer: B. Outcome accountability judges by results rather than by procedure followed.
-
2. A placement agency serves only easily employable candidates to meet its targets. This is called:
- Creaming
- Escalation
- Severance
- Recusal
Show answer
Answer: A. Creaming means selecting easy cases to meet targets while neglecting harder ones.
-
3. Recorded crime falls but complaints of refusal to register cases rise. The most likely explanation is:
- Crime has really fallen
- Citizens are complaining more for no reason
- The recording system has improved
- Gaming of a single outcome indicator
Show answer
Answer: D. A single indicator tied to rankings creates incentives to avoid recording crime.
-
4. A hospital meets every procedural requirement but has high post-surgery deaths. What should be added?
- More registers and forms for staff
- Outcome measures with clinical audit
- Fewer inspections of the wards
- Stricter attendance only for doctors
Show answer
Answer: B. Outcome measures such as infection and mortality rates, with peer clinical audit, address the gap.
-
5. Which design feature most reduces gaming?
- Relying on a single indicator for every decision
- Removing all process rules once targets are set
- Including an independent measure such as a citizen survey
- Penalising staff who serve the most difficult cases
Show answer
Answer: C. Independent measures are harder to manipulate and reveal gaming of other figures.
Sign in to keep track of the lessons you finish.