The Procedures in TESTIMATE
After being developed, all algorithms were validated in detail and the results were compared with examples from the literature. After installing the program in the concrete user environment, you can use the build in validation function - this is a high-end validation standard - to validate again all procedures yourself by simply pressing a button.
Why Do I Need These Methods?
Nonparametric methods
- Appropriate for all data distributions arising in practical work
- Confirmatory trials should be robust - no pre-tests (see also ICH Biostatistics Guidelines E9, 1998)
Exact procedures
- Necessary for small, sparse and unbalanced data sets
- The procedures give correct, robust results - independent of the data situation. (These procedures are called the Gold-Standard of Statistics.)
Effect size measures and confidence intervals
- ICH Biostatistics Guideline E9, 1998
- CONSORT Publication Guidelines
- Publication Manual of the American Psychological Association [APA], 1994
Tests for equivalence / non-inferiority
- ICH Biostatistics Guideline E9, 1998 (for studies with active control)
- Recommendations for Biostatistics of Mutagenicity Studies, Drug. Inf. J. 31, 1997, 323-326 (risk assessment)
What about my 'old' studies?
All applications to authorities and publication manuscripts for scientific journals must now be submitted in accordance with the above requirements regardless of the procedures with which the study was planned!