Table 2.

Recommendations for test development

Psychometric properties
Validity (i.e. does the task measure what it is intended to measure?)	Content validity Construct validity Sensitivity and specificity (according to the examination issue) Predictive validity Cross-cultural validity (including multiple language versions for verbal tests)	Clinical relevance Definitions updating according to most recent neuroscientific evidence Highly sensitive to early phases of dementia; differential diagnosis among dementias Change monitoring over time Minimal effects of literacy
Reliability	Test/re-test stability; inter-rater reliability	Avoid practice effects; parallel forms; avoid floor- and ceiling effects
Feasibility	Balance between economic and human resources costs. Availability of normative data across European countries	Optimise administration time; minimise discomfort to the patient