Some general principles of assessment

• Measurement error
  • Typically, a single observation is hard to interpret. It needs context
    • E.g. clinical presentation, complaints, other measures, repeated observations
  • Some cut-offs may be of clinical relevance (e.g. BMI)
  • 
• Calibration!
• Standardisation
• Normal values
  • Percent predicted or percentile à lower limit of normal
  • For the test as you conduct it