Structural Equation Modeling, Article Critique Example

Structural Equation Modeling is a common statistical technique utilized to examine the correlative and exploratory relationships between multiple qualitative variables.  While it is commonly utilized for its applications-based format, it can often be used to analyze relationships between two or more variables within a given environment that may not actually have any relative characteristics between one another.  It is this very flaw that should be analyzed carefully before implementing such research characteristics in order to draw inappropriate conclusions on a given topic.

In general, the use of structural equation modeling is to adequately determine the relationship between two or more variables.  “A final issue to consider is that there is some work that has redefined the decomposition of effects to conform to specific definitions of causality that involve the manipulations of variables in the model” (Bollen, 1998).  It is these very manipulative variables that create the need for structural equation modeling.  Through careful statistical analysis and evidence, multiple variables can be compared as seen throughout the research compiled by Bollen (1998, 2001).  Sociological variables within professional research can vary by relative factors, which is a main consideration for how structural equation modeling is able to be used as a correlative and predictive measure for relationship analysis.

One key sociological area of study for relationship analysis is that of studying a multitude of variables and their correlation toward achieving a model higher educational environment.  For instance, Hamzah et al. (2008) produced research specifically on this topic and concluded that teachers with a higher level of subject matter proficiency and expertise were more likely to translate into more proficient levels of teaching practice.  This shows how structural equation modeling can be used to combine two relative variables to determine predictive analytical results.  The next step in the process is to utilize the structural equation modeling to determine the relationship of these variables so that managers can then make appropriate decisions and implement practice behaviors that correlate with these research findings.  For instance, in the study with Hamzah et al. (2008), higher educational personnel should attempt to find professors with the greatest levels of expertise in their subject areas so that students are receiving the best possible teaching.

Bollen (1998) analyzed the Wright’s work in which the development of the path analysis and the path diagram were used to determine relationships between multiple variables.  Over the course of time, this technique has evolved into structural equation modeling and the same diagram techniques can be utilized to show a visual representation of variable relationships for managerial and functional analysis.  The use of this visual component can be much more useful for individuals not specially-trained or knowledgeable in statistics to see the inter-relationships of variables.

As previously stated, a common error associated with structural equation modeling is to utilize this technique to analyze the relationships between two unrelated variables that may be inappropriately believed to be related.  For instance, the quality of light within an environmental setting would have little relationship with a teacher’s topic knowledge and quality levels of expertise.  These variables are clearly unrelated; however, researchers may attempt to draw a relationship comparison of variables using structural equation modeling and path diagrams.  This does not take away from the importance or useful nature of structural equation modeling; it is simply to state that such analyses may be inappropriate and have little bearing on the true nature of variables and their inherent co-relationships.  Given the appropriate use, structural equation modeling can be utilized to benefit almost any sociological environment where variables must be analyzed.

References

Bollen, K.A. (1998). Encyclopedia of Biostatistics. Sussex, England: Wiley.

Bollen, K.A. (2001). Indicator: methodology. Religions of India, 7282-7287.

Hamzah, M.S.G., Mohamad, H., & Ghorbani, M.R. (2008). Excellent teachers’ thinking model: implications for effective teaching. Australian Journal of Teacher Education, 33(4), 11-27.