Naslov
Algorithm and program for decomposition of one set of variables from another set in STATISTICA

Autori
Dizdar, Dražan ; Viskić-Štalec, Nataša

Vrsta, podvrsta i kategorija rada
Radovi u zbornicima skupova, cjeloviti rad (in extenso), znanstveni

Izvornik
Kinesiology: New Perspectives / Milanović, Dragan ; Prot, Franjo - Zagreb, 2002, 656-659

Skup
3rd International Scientific Conference

Mjesto i datum
Opatija, Hrvatska, 25.09.2002. - 29.09.2002

Vrsta sudjelovanja
Poster

Vrsta recenzije
Međunarodna recenzija

Ključne riječi
algorithm; program; decomposition of one set of variables

Sažetak
One of the real methodological problems in anthropological research studies is the partialization of impact of one set of variables from another set of variables which constitutes the issue of our analysis. Thus it is possible, for example, to identify the impact of motor variables on success in sport (criterion) assuming that the entities be equal in terms of morphological variables. In the process, the motor results of entities on the variables are decomposed into two parts: the first part being the one that depends and the second being the part that does not depend on morphological variables. After that the relations are determined between the part of the results of entities in motor variables, independent of body built on the one hand and success in a particular sport on the other. The partialisation of impact of morphological variables conducted in such a way enables the monitoring of success in sport exclusively depending on the motor component of an athlete. The second example may be found in kinesiological research studies of the situation-related space in sports in which two opponents (individual or team) confront one another directly. The functional dependence of the efficiency of a team/individual on the efficiency of an opponent may be determined by the relations between the variables that represent the play of the two opponents and the results of competition participants in the monitored situation-related variables may be decomposed into a prognosticated part (dependent) and a residual part (independent). An experimental design with two independent groups, out of which one did not undergo a treatment (K) and the other one did (E), is a design aimed at determining the effects of an experimental program of the training process. There exists a high probability that the K and E groups that had been randomly selected from the same population do not differ from each other statistically significantly in the initial measurement, so that the significance of changes occurring under a treatment in that case is determined by the difference between the vectors of arithmetic means of the final measurement. Otherwise, it would be necessary to neutralize the impact of the differences between the K and E group in the initial measurement and determine the significance of changes occurring under the treatment by the difference between the vectors of arithmetic means of the final measurement (the variables of the first measurement of the K and E group represent the predictor and the variables of the second measurement the criterion set). Thus the results of the entities of groups of the final measurement are decomposed into a part that is dependent on the variables of the initial measurement and the residual part that is independent of the differences between the E and K groups before the commencement of a treatment. The residual results may be used to determine by means of the analysis of covariance the significance of changes occurring under the treatment in such a way that the difference between the vectors of arithmetic means between E and K groups in the final state be used. Usually, the ANOVA/MANOVA or discriminant analysis (Momirović, Gredelj and Sirovicza, 1977) are applied. The program DEVRA may be brought up from any module of the program system STATISTICA that is active in that moment by activating the option STATISTICA BASIC Program and the submenu Open Other of the scrolling menu File. Before using the program it is necessary to select a data file formed in such a way that the vectors of entities are put in rows and the vectors of variables in columns. If a user wants to decompose the standardized variables, it is necessary that first the variables be standardized. The program is started by selecting the option Execute that may be found on the toolbar.

Izvorni jezik
Engleski

POVEZANOST RADA