As I introduce the
versatility of using SEM in research in my previous blog, at this time I will
further share the importance of Regression Model or Regression Analysis in
SEM. As Correlation and Regression
analysis is known to the measurement and testing of association, therefore
Regression Analysis is useful in SEM modeling in testing the associative
relationship of the model.
Least
Squares Regression
The fundamental
observation of variables by using regression analysis is by two variables or
between two variables. Usually the given
series of given data is represented by two groups frequently X and Y. As Y is the Independent variable and X is the
dependent variable, the movement of X is dependent to the movement of Y.
Any negative behavior movement of X is a variance movement towards Y. If the series of events of X is congruent to the
series events of Y then the correlation of the two variables are high which
means that the given X and Y is closely related to each other. But on the contrary, if the observed variable
is far enough to prove that its behavior is not the same with the latent
variable therefore the given two variables are not related to each other.
Further representation of Regression Modeling
is best presented through the use of line graphical presentation. By using line
graph presentation each given event is well observed through a series.
Least
Squares Multiple Regression
Multiple Regression
Analysis implies to the testing of several dependent variables basing from one
independent variable. Dependent variables are usually represented by X1, X2,
X3, & X4. As to the SEM modeling, sometimes the independent variable may
use more than one Independent variable depending upon the SEM model.
As to this case, to make it simpler we will
only discuss the common example of Multiple Regression model using one Y and
four Xs. With one independent variable
all dependent variables will be subject to a test with the independent
variable.
With the same nature of test in
the two variables testing, multiple regressions applies the same process to all
given dependent variables and with a series of events, the nearest correlation
behavior to the independent variable Y is the highest possible X variable to be
chosen as a predictor.
Variance
Variance usually
occurs from one event of a dependent variable basing it to the corresponding
independent variable movement. The
distance between the independent variable and the dependent variable are
commonly called the variance. The
variances of the series of events between the X and the Y will determine for
the whole correlation of the two groups of variables.
As for the SEM modeling,
regression model will get complicated as the structural framework of the study will
completely define the desired data analysis.
Comments
- RockyMountainLabs.com