Question

What is covariation between two variables?

How is "covariation" (or correlation) between two variables different from "relationship" between them?

Why do we need regression analysis to establish relationships between the dependent variable and one or more of the independent variables and correlation analysis is not enough?

Answer 1

1. Covariance is a proportion of the joint changeability of two irregular factors. On the off chance that the more prominent estimations of one variable, for the most part, relate with the more noteworthy estimations of the other variable, and similar holds for the lesser qualities, (i.e., the factors will in general show comparable conduct), the covariance is positive. On the contrary, case, when the more prominent estimations of one variable principally relate to the lesser estimations of the other, (i.e., the factors will in general show inverse conduct), the covariance is negative. The indication of the covariance in this manner shows the propensity in the direct connection between the factors. The extent of the covariance isn't anything but difficult to decipher because it isn't standardized and thus relies upon the sizes of the factors. The standardized rendition of the covariance, the correlation coefficient, be that as it may, it appears by its size the quality of the direct connection.

2. If the adjustment in one variable has all the earmarks of being joined by an adjustment in the other variable, the two variables are said to be correlated and this inter­dependence is called covariation. In short, the propensity of concurrent variety between two variables is called covariation. For instance, there may exist a relationship somewhere in the range of statures and loads of a gathering of understudies, the scores of understudies in two distinct subjects are relied upon to have an association or relationship between them.

To quantify the level of relationship between two variables is the topic of correlation investigation. Consequently, correlation implies the relationship or "going-harmony" or correspondence between two variables.

3. Correlation and Regression are the two examinations dependent on multivariate dissemination. A multivariate circulation is portrayed as a conveyance of various variables. Correlation is portrayed as the examination which tells us the affiliation or the nonattendance of the relationship between two variables 'x' and 'y'. On the opposite end, Regression examination, predicts the estimation of the dependent variable dependent on the known estimation of the independent variable, expecting that normal numerical relationship between two or more variables.

There is no difference between the dependent variable and the independent variable in Correlation. Whereas, in Regression both the variables are different. That's why we need regression analysis to establish relationships between the dependent variable and the independent variable.