Zjawisko współliniowości w statystycznych badaniach przyczynowo-skutkowych procesów ekonomicznych
Loading...
Date
1982
Authors
Advisor
Editor
Journal Title
Journal ISSN
Volume Title
Publisher
Wydział Prawa i Administracji UAM
Title alternative
Collinearity Phenomenon in Statistical Cause-Effect Studies of Economic Processes
Abstract
The elaboration is devoted to the problems of variables collinearity and its
repercussions for a study process of cause-effect conditions of socio-economic phenomena.
The study is carried out with the use of statistical methods of interdependence
analysis. It appears that collinearity is a crucial obstacle for correct
estimation of mutual relations between economic variables.
The solution of that problem needs several measures. Their synthesis has
been accomplished in the paper. There have been discussed the following problems:
1) measurement of collinearity (Glauber-Farrar test), 2) range of collinearity
on a set of explanatory variables (F-test, t-test), 3) search for ways of reduction
adverse effects of collinearity.
In the second part of the paper has been presented one of the methods of
reducing collinearity i.e. the ridge regression method. That method is generally
founded on such basis that it is better to use biassed estimators but with variance
smaller than variance of those estimators obtained with the help of the CLSM.
The method has been implemented for an analysis of a real statistical date. The
figures regarded costs and their determinants in state-owned foodstuffs retail
trading enterprises.
The study confirmed large utility of the ridge regression method in a process
of identification of cost-creating factors, but considering narrow scope of its
hitherto applications further verification studies are indispensable.
Description
Sponsor
Digitalizacja i deponowanie archiwalnych zeszytów RPEiS sfinansowane przez MNiSW w ramach realizacji umowy nr 541/P-DUN/2016
Keywords
Citation
Ruch Prawniczy, Ekonomiczny i Socjologiczny 44, 1982, z. 2, s. 135-145