Article Preview
Top1. Overview Of An Example From The Directory Of Dr. Sachs
The authors of this article reviewed an excerpt from the authoritative reference book on statistical evaluation by Professor Dr. Sachs (Germany). Just by looking at the amazing size of the bibliography, you will quickly understand how fundamental this work is. Nevertheless, the authors of the article discovered flaws of the directory:
- 1.
The error of the minimum permissible statistical reliability for this example is a «sight» error, which may lead the authors of new forms of treatment, manufacturers of new drugs, just new substances, to serious consequences;
- 2.
The understated value of the statistical reliability obtained in this example;
In addition to the above, we can point to two insignificant technical gaps in essence:
Consider this example (Sachs, 1976): «Experimental plan: there are three groups of 40 patients each ...1 the third group receives specific treatment with a double normal dose.»
Zero hypothesis: the results of the treatment of the three methods are the same.
Alternative hypothesis: the outcome of the treatment of the three methods is not the same.
Level of significance: 95%.
Dr. Sachs continues (Sachs, 1976): «The choice of criteria: a multicellular criterion is considered.» And here (Sachs, 1976): «The main field of application of this distribution is open, Helmert (1876) and Pearson (1900), – it is a test of hypotheses about contingency tables of signs 2x2 and larger size.»
The universal table 1 served as a template for the contingency table of size signs 3x3 (Table 2).
Table 1. Scheme for two-sided classification: one of two series of features can be considered as a series of samples
| 1-2-j-c | Sums in rows |
1 |
| |
2 | | |
- | ------ | |
I | | |
- | ------ | |
r | | |
Sums by columns | | |
Source: Sachs, L. Statistical estimators, Transl. from Germ. / L. Sachs – Moscow: Statistics, 1976. – 598; P. 433.