The characteristics of the three types of measurement in health care6
| Characteristic | Research | Accountability | Improvement |
| Aim/Intent | New generalizable knowledge | Comparison, choice, reassurance, spur change | Improvement of processes and outcomes |
| Observability of testing | Blinded or controlled | No test, observe current performance | Tests are observable |
| Bias | Design to eliminate bias | Measure and adjust to reduce bias | Accept consistent bias |
| Sample size | As many as possible for sufficient power | Obtain 100% of available data | ‘Just enough’ data with sequential samples |
| Flexibility of hypothesis | Fixed hypothesis | No hypothesis | Hypotheses are flexible and change as learning takes place |
| Testing strategy | One test or several large tests | No tests | Sequential tests |
| Determining if a change is an improvement | Hypothesis testing, statistics (eg, t-test, χ2, ANOVA) | No change focus | Statistical process control charts |