Correlation or Causality: New Evidence from Cross-Sectional Statistical Analysis of Auto-Crash Data in Nigeria

by Olushina Olawale Awe.

Abstract: Correlation Analysis is a key statistical technique for detecting the extent of relationship that exists among variables of interest in any environment and in various fields of human endeavor ranging from science, social sciences arts and even engineering and technology. This study is aimed at investigating the extent of association among some cross sectional, categorical, auto-crash variables in Nigeria. Differences between causation and correlation are re-examined. In all analyses, we find that all four variables considered are highly correlated. However, we note that correlation is not tantamount to causality. Finally, we compute the statistic of the coefficients of determination which is further used to determine the proportion of variation of one variable that is attributable to the variance in a related variable.

Key Words: Correlation, Classical Regression Models, Coefficient of Determination, Causality

Author:
Olushina Olawale Awe, olawalewe@yahoo.co.uk

Editor: R.G. Graf, rgraf@sunstroke.sdsu.edu

READING THE ARTICLE: You can read the article in portable document (.pdf) format (332633 bytes.)

NOTE: The content of this article is the intellectual property of the authors, who retains all rights to future publication.

This page has been accessed 879 times since MAY 6, 2013.


Return to the InterStat Home Page.