<<
>>

Statistical methods of processing of the received data


Statistical processing of a material of dissertational work
It is made on personal computer IBM PC/AT - IV. in Windows environment 2000 with application of a package of computer programs and spreadsheets (Excel, Fox Pro, Open Access, Paradox), and also with use of statistical library of programs «Statistica».
The Design procedure stole up according to references of following authors: J. pollard, J. Glass, J. Stenli, with use of methods of descriptive statistics, the correlation analysis. The correlation analysis included methods Pirsona, Spirmena, and also the analysis of dichotomies Kertena, depending on scales of investigated sizes.
Fxy =

Reliability of the calculated factor of correlation estimated by means of an average error (shg) and criterion t under more low resulted formulas:

At check of hypotheses for compared vyborok were applied the variance analysis with methods of plural comparisons SHeffe and Tjuki, and also T - criterion Stjudenta to didymous comparisons.
The estimation of the importance of risk factors was defined by means of definition of an indicator of relative risk and borders of its possible fluctuations - OP (With і).
For acknowledgement of a nonrandomness of differences confidential borders (Ci) relative risk under the formula offered Haldane (1970) were defined

Where N - the sum of all observations of the basic and control groups.
The calculation method consisted in calculation of logarithms of confidential intervals so-called korrigirovannogo relative risk OPk then transition to a normal scale follows.
OPk (the relative risk, korrigirovannyj) was calculated by us under the formula:
(a+0,5) (d+0,5)
(Ъ+0,5) (с+0,5)
To tap degree of influence of the factor in whole and its separate components, we have taken advantage of a technique of non-uniform consecutive procedure of recognition - NPP which allows to define individual risk of occurrence of the pathology depending on all complex of risk factors taking into account the importance each. For this purpose we sectioned each selected factor into certain number of ranges (gradation) of a sign, counting up frequency of each of them in the groups working at factory "Тольяттиазот" and factory "Тольяттитрансформатор", defined diagnostic factor (recreation centre) for each range under the formula:
Pi
Dk = IOlg^i
Pl
Where R] and P2 - frequency of a range of a sign at women of compared groups, and informativnost (J) under the formula:
J = IOlg^i-0,5 {P\-Pi)
In some cases, dictated by type of distribution of investigated sizes, data processing was made by methods of nonparametric statistics, among them: Vilkinsona - the God-send-uitni - criterion of differences in average tendencies for independent vyborok; serial criterion Valda - Volfovitsa; Fisher's exact method for chetyrehpolnyh tables, a method of angle transformation of Fisher and method Z of statistics for an estimation in aggregate; chi-square criterion for the analysis of tables of an associativity of signs.
<< | >>
A source: KHUTORSKAYA NATALIA NIKOLAEVNA. FEATURES of FEMALE STERILITY At the WORKING WOMEN OCCUPIED In MANUFACTURE of KARBAMIDNO-FORMALDEHYDE PITCH And the WAY of PREVENTIVE MAINTENANCE of ITS HARMFUL INFLUENCE. Dissertatsijana competition of a scientific degree of the candidate of medical sciences. Samara - 2003. 2003

More on topic Statistical methods of processing of the received data:

  1. statistical processing of the received data
  2. statistical processing of the received data
  3. statistical processing of the received data.
  4. 2.3 Statistical processing of the received data
  5. STATISTICAL PROCESSING OF THE RECEIVED DATA
  6. methods of the scientific analysis and statistical processing of the received results
  7. Methods of statistical data processing
  8. 2.3. Methods of statistical processing of the results received in research
  9. Methods of statistical data processing.
  10. Statistical processing of the received results
  11. statistical processing of the received results