That is, between the average monthly air temperature and the number of deaths from acute intestinal infection there is a direct correlation.
It is extremely easy to understand calculation requires quite hard, though simple mathematical operation. The computational work is especially difficult when the members of the correlated series have large numerical values, especially if the options series are correlated in the form of grouped intervals and, therefore, not necessary to calculate the simple and weighted average.
The average error of the correlation coefficient. Since the correlation coefficient is calculated in clinical trials are usually a limited number of cases, there is often a question of the reliability coefficient obtained. To this end, the average error is determined correlation coefficient. With a sufficiently large number of observations (more than 100), the average error of the correlation coefficient (mr) is given by:
n – number of observations.
In that case, if the number of observations is less than 100, but more than 30, more accurately determine the average error of the correlation coefficient, using the formula:

With sufficient reliability for medical research of a more or less connection can be stated only when the value of the correlation coefficient greater than or equal to the value of his three errors (3mr). Typically, the ratio of the correlation coefficient () to its mean error (mr) designated by the letter t and

If 3, the correlation coefficient is valid.
Variance analysis
Analysis of variance allows us to give a generalized description of three or more averages or indexes, and allows you to:
- To measure the force of impact;
- To evaluate the difference between private secondary or performance;
- To determine the accuracy of the difference or private secondary indicators.
Analysis of variance indicates the degree of dispersion of variation (dispersion) of the measured properties of a typical level around the middle, so it makes it possible to study the effect on the end result of several factors including the role of each of them and compare the effect of individual factors among themselves.
Study of the factors by comparing the average values ​​observed trait resulting from the effects of each of these factors in different combination thereof.
The following types of analysis of variance: one-factor, two-factor and multi-factor.
The methodology of the analysis of variance is presented in numerous special editions of health statistics.
Non-parametric tests
Non-parametric tests are used mainly in cases where the phenomenon under study is different from the normal distribution. On the one hand, they allow you to evaluate the nature, the tendency of the phenomenon (increase, decrease, no change), although, on the other, most of them have fairly high statistical power (sensitivity). Especially effective is the use of nonparametric tests for small samples (n <30), in the study of qualitative features. The advantage of most of the non-parametric tests is the relative ease of calculation.